[V0_FILE]json:file=".claude/settings.local.json" { "permissions": { "allow": [ "Bash(tree:*)", "Bash(git config:*)", "Bash(git add:*)", "Bash(git commit:*)", "Bash(git push:*)", "Bash(git log:*)", "Bash(git pull:*)", "Bash(git mv:*)", "Bash(git rm:*)" ], "deny": [], "ask": [] } } [V0_FILE]plaintext:file=".gitignore" # See https://help.github.com/articles/ignoring-files/ for more about ignoring files. # dependencies /node_modules # next.js /.next/ /out/ # production /build # debug npm-debug.log* yarn-debug.log* yarn-error.log* .pnpm-debug.log* # env files .env* # vercel .vercel # typescript *.tsbuildinfo next-env.d.ts [V0_FILE]markdown:file="CHANGELOG.md" # Linda Health Companion - Change Log This document tracks all significant changes, new features, and updates to the Linda Health Companion application. ## [v0.2.0] - 2025-09-11 ### 🎯 Major Feature Addition: Plans Management System #### **New Tab Added: Plans** - **Date**: September 11, 2025 - **Commit**: `f97cb03` - feat: inline plan editing with modal and toast notifications - **Timestamp**: 2025-09-11 19:28:34 UTC - **Developer**: v0.dev (AI-assisted development platform) - **Documentation Updated**: 2025-09-11 15:40:55 EDT by nanaoosaki #### **What's New** ##### 📱 **Navigation Enhancement** - Added 4th navigation tab: **Plans** between Today and History - Updated navigation structure: Chat → Today → **Plans** → History - Navigation accessible via bottom navigation bar with Target icon ##### 🏥 **Comprehensive Treatment Plan Management** **Core Features:** - **Plan Dashboard** (`/plans`) - Visual management of all treatment plans - **Individual Plan View** (`/plans/[id]`) - Detailed plan tracking with analytics - **Plan Creation Wizard** - Multi-step guided plan setup - **Effect Logging System** - Track benefits, side effects, and outcomes - **Status Management** - Active, Paused, Ended plan states **Supported Plan Types:** - 💊 **Medication Plans** - Prescription tracking with dosage and timing - 🏃‍♀️ **Physical Therapy** - Exercise routines and progress tracking - 📍 **Acupuncture** - Session scheduling and effectiveness monitoring - 🔄 **Health Routines** - Daily habits and wellness practices ##### 📊 **Analytics & Tracking** - **Effect Timeline** - Visual representation of treatment outcomes over time - **Severity Scoring** - 0-10 scale for tracking symptom changes - **Signal Monitoring** - Custom health metrics per plan (pain, sleep, mood, etc.) - **Trend Visualization** - Charts showing plan effectiveness - **Quick Logging** - Streamlined data entry for rapid updates ##### 🛠 **Technical Implementation** **New API Endpoints:** \`\`\` GET /api/plans - Retrieve all plans with filtering POST /api/plans - Create new treatment plan GET /api/plans/[id] - Get individual plan details PATCH /api/plans/[id] - Update plan information GET /api/plans/[id]/effects - Retrieve effect logs POST /api/plans/[id]/effects - Log new effects/outcomes \`\`\` **New Components Created:** - `plan-card.tsx` - Rich plan display with status management - `new-plan-wizard.tsx` - Multi-step plan creation interface - `edit-plan-sheet.tsx` - Slide-in plan editing panel - `effect-log-modal.tsx` - Detailed effect logging modal - `quick-effect-log.tsx` - Streamlined logging interface - `effect-timeline.tsx` - Visual timeline for effects **Enhanced UI Components:** - `ui/badge.tsx` - Status and category indicators - `ui/sheet.tsx` - Slide-in panel component - `ui/tabs.tsx` - Tab navigation component - `ui/toast.tsx` - Real-time notification system **Data Models Added:** \`\`\`typescript interface Plan { id: string type: "medication" | "pt" | "acupuncture" | "routine" name: string description?: string status: "active" | "paused" | "ended" startDate: string endDate?: string targetDuration?: number cadence?: string clinician?: string notes?: string monitorSignals: string[] lastUpdated: string } interface EffectLog { id: string planId: string timestamp: string effectType: "benefit" | "side-effect" | "neutral" severity: number freeText?: string linkedSignals: string[] } \`\`\` **React Hooks Integration:** - `usePlans.ts` - SWR-powered data fetching and caching - Comprehensive CRUD operations with optimistic updates - Real-time plan status synchronization ##### 👥 **User Experience Enhancements** - **Floating Action Button** - Quick plan creation when active plans exist - **Status Badges** - Visual indicators for plan states - **Provider Tracking** - Clinician assignment and notes - **Outcome Strip** - Quick view of latest plan results - **Toast Notifications** - Real-time feedback for all actions ##### 🗂 **Mock Data System Enhancement** - Enhanced `MockStore` class with plan management capabilities - Mock plan data with realistic scenarios - Effect logging simulation with trend generation - Comprehensive test data for development #### **Files Added/Modified** \`\`\` 📁 New Files: app/plans/page.tsx - Plans dashboard app/plans/[id]/page.tsx - Individual plan view app/api/plans/route.ts - Plans CRUD API app/api/plans/[id]/route.ts - Individual plan API app/api/plans/[id]/effects/route.ts - Effect logging API components/plan-card.tsx - Plan display component components/new-plan-wizard.tsx - Plan creation wizard components/edit-plan-sheet.tsx - Plan editing interface components/effect-log-modal.tsx - Effect logging modal components/quick-effect-log.tsx - Quick logging interface components/effect-timeline.tsx - Effect timeline visualization components/ui/badge.tsx - Badge component components/ui/sheet.tsx - Sheet component components/ui/tabs.tsx - Tabs component components/ui/toast.tsx - Toast notification system hooks/use-toast.ts - Toast hook lib/api/usePlans.ts - Plans data hook 📝 Modified Files: components/bottom-nav.tsx - Added Plans tab lib/api/types.ts - Added Plan/Effect interfaces lib/mocks/index.ts - Enhanced with plan data package.json - Updated dependencies \`\`\` #### **Dependencies Added** - Enhanced Radix UI components for tabs and toast notifications - Additional utility components for plan management interface ### 🔄 **Previous Updates** #### **AI-Powered Health Data Parsing** - 2025-09-11 - **Developer**: nanaoosaki + AI assistance - **Timestamp**: 2025-09-11 15:36:21 EDT **Features Added:** - GPT-4o integration for intelligent goal extraction - AI-powered medication, routine, and symptom parsing - Multimodal input support (text, voice, image) - Structured output validation with Zod schemas - Enhanced onboarding flow with AI assistance **Components Added:** - `goal-input.tsx` - AI-powered goal setting - `medication-input.tsx` - Smart medication parsing - `routine-input.tsx` - Routine configuration - `symptom-input.tsx` - Symptom tracking - `profile-coach.tsx` - AI coaching system - `first-run-nudge.tsx` - Onboarding enhancement **API Endpoints Added:** - `/api/parse-goals` - Intelligent goal extraction - `/api/parse-medications` - Medication parsing - `/api/parse-routines` - Routine extraction - `/api/parse-symptoms` - Symptom analysis #### **Initial Codebase Analysis** - 2025-09-11 - **Developer**: nanaoosaki - **Timestamp**: 2025-09-11 10:01:32 EDT - **Documentation**: Comprehensive CLAUDE.md created with architecture analysis --- ## Development Attribution ### Core Development Team - **v0.dev Platform**: AI-assisted rapid prototyping and feature development - **nanaoosaki**: Project maintenance, documentation, and integration ### Development Process - **Primary Development**: v0.dev platform for rapid feature implementation - **Code Review & Integration**: Manual review and git management by nanaoosaki - **Documentation**: Comprehensive technical documentation by nanaoosaki - **Quality Assurance**: Iterative testing and refinement ### Repository Management - **Main Branch**: Production-ready code with comprehensive testing - **Commit Strategy**: Feature-based commits with detailed descriptions - **Documentation**: Keep CLAUDE.md and CHANGES.md updated with all major changes - **Deployment**: Automatic deployment to Vercel from main branch --- *This changelog follows [Semantic Versioning](https://semver.org/) and [Keep a Changelog](https://keepachangelog.com/) principles.* [V0_FILE]markdown:file="CLAUDE.md" # Linda Health Companion - Comprehensive Codebase Analysis ## 1. Project Overview **Project Type**: Progressive Web Application (PWA) for health tracking and monitoring **Primary Purpose**: Personal health companion app focused on tracking migraine, asthma, reflux, sleep, and mood through quick voice notes **Tech Stack**: Next.js 15.2.4 + React 19 + TypeScript + Tailwind CSS + AI SDK **Architecture Pattern**: Next.js App Router with API routes and modular component structure **Target Platform**: Web-based with mobile-responsive design **Development Status**: Built with v0.dev, deployed on Vercel **AI Integration**: OpenAI GPT-4o for intelligent health data parsing and goal extraction ## 2. Detailed Directory Structure Analysis ### `/app` - Next.js App Router - **Purpose**: Main application routes and API endpoints using Next.js 13+ App Router - **Key Files**: - `layout.tsx` - Root layout with theme provider setup - `page.tsx` - Home/chat page - main interaction interface - `globals.css` - Global styles and Tailwind CSS imports - **Subfolders**: - `/api/*` - Server-side API routes for health data processing - `/history` - Health data history view (accessible via bottom navigation) - `/onboarding` - AI-powered user setup flow with intelligent parsing - `/plans` - Treatment and routine plan management (accessible via bottom navigation) - `/profile` - User profile management and health goal configuration - `/settings` - App configuration and preferences - `/today` - Today's health summary and insights (accessible via bottom navigation) **Current Navigation Structure**: - **Chat** (`/`) - Main interaction interface with multimodal composer - **Today** (`/today`) - Daily health summary and metrics - **Plans** (`/plans`) - Treatment plan management and effect tracking - **History** (`/history`) - Historical health data and patterns ### `/components` - Reusable UI Components - **Purpose**: Shared React components following atomic design principles - **Structure**: - `/ui` - Base UI components (shadcn/ui based) - Root-level components for app-specific functionality - **Key Components**: - `layout.tsx` - App-level layout wrapper - `header.tsx`, `bottom-nav.tsx` - Navigation components - `multimodal-composer-v2.tsx` - Voice/text/file input interface - UI components for health data visualization ### `/lib` - Core Business Logic - **Purpose**: Shared utilities, API clients, and business logic - **Structure**: - `/api` - API client and data source abstraction - `/logic` - Business logic and processing - `/mocks` - Mock data and testing fixtures - **Key Files**: - `api/client.ts` - Main API client with mock/real API switching - `api/types.ts` - TypeScript interfaces for data models - `mocks/index.ts` - Comprehensive mock store for development ### `/hooks` - Custom React Hooks - **Purpose**: Shared React hooks for state management and data fetching - **Focus**: Data fetching, profile management, and health data operations ### `/public` - Static Assets - **Purpose**: Static files served by Next.js - **Contents**: Images, icons, manifests for PWA functionality ### `/styles` - Styling Configuration - **Purpose**: Additional CSS files and style configurations - **Integration**: Works with Tailwind CSS for comprehensive styling ## 3. File-by-File Breakdown ### **Core Application Files** #### Main Entry Points - **`app/layout.tsx`** - Root application layout with theme provider integration - **`app/page.tsx`** - Primary chat interface for health data input via voice/text/files - **`app/plans/page.tsx`** - Treatment plan management dashboard with active/ended plans - **`app/plans/[id]/page.tsx`** - Individual plan detail view with effect tracking and analytics - **`components/layout.tsx`** - Reusable layout component with header and bottom navigation #### API Routes (Server-side) - **`app/api/intake/route.ts`** - Processes multimodal health data input (text, voice, files) - **`app/api/calendar/route.ts`** - Calendar-related health data operations - **`app/api/history/route.ts`** - Historical health data retrieval - **`app/api/insights/route.ts`** - Health insights and analytics - **`app/api/plans/route.ts`** - Plan management CRUD operations (GET, POST) - **`app/api/plans/[id]/route.ts`** - Individual plan operations (GET, PATCH, DELETE) - **`app/api/plans/[id]/effects/route.ts`** - Effect logging and retrieval for specific plans - **`app/api/parse-goals/route.ts`** - AI-powered goal extraction from user input using GPT-4o - **`app/api/parse-medications/route.ts`** - Intelligent medication parsing and structuring - **`app/api/parse-routines/route.ts`** - Daily routine and habit extraction - **`app/api/parse-symptoms/route.ts`** - Symptom pattern analysis and categorization ### **Configuration Files** #### Build & Development - **`package.json`** - Dependencies: React 19, Next.js 15, Radix UI, Tailwind, AI SDK - **`next.config.mjs`** - Next.js configuration with build optimizations - **`tsconfig.json`** - TypeScript configuration with path aliases - **`tailwind.config.js`** - Tailwind CSS configuration - **`postcss.config.mjs`** - PostCSS configuration for CSS processing - **`components.json`** - shadcn/ui component library configuration #### Development Workflow - **`pnpm-lock.yaml`** - Package manager lock file (using pnpm) - **`.gitignore`** - Git ignore patterns for Node.js/Next.js projects ### **Data Layer** #### API Client Architecture - **`lib/api/client.ts`** - Centralized API client with environment-based switching - **`lib/api/dataSource.ts`** - Data source configuration (mock vs real API) - **`lib/api/types.ts`** - TypeScript interfaces for Profile, Suggestion, Plan, EffectLog, and API responses - **`lib/api/usePlans.ts`** - React hook for plan management with SWR caching #### Mock Data System - **`lib/mocks/index.ts`** - Comprehensive mock store with: - Profile management simulation - Health data parsing algorithms - Suggestion generation system - Onboarding flow simulation - **Plan management system** - CRUD operations for treatment plans - **Effect logging system** - Track outcomes and side effects - **Trend data generation** - Mock analytics for plan effectiveness ### **Frontend/UI Components** #### Core UI Framework - **shadcn/ui integration** - Modern, accessible component library - **Radix UI primitives** - Headless UI components for complex interactions - **Tailwind CSS** - Utility-first styling approach - **next-themes** - Dark/light theme switching #### Specialized Components - **`MultimodalComposerV2`** - Advanced input component for voice, text, and file upload - **`goal-input.tsx`** - AI-powered goal setting component with multimodal input - **`medication-input.tsx`** - Smart medication parsing and management interface - **`routine-input.tsx`** - Daily routine configuration and tracking - **`symptom-input.tsx`** - Symptom tracking and pattern recognition interface - **`first-run-nudge.tsx`** - Onboarding experience component - **`profile-coach.tsx`** - AI coaching and guidance system #### Plan Management Components - **`plan-card.tsx`** - Plan display component with status management and quick actions - **`new-plan-wizard.tsx`** - Multi-step plan creation wizard with validation - **`edit-plan-sheet.tsx`** - Plan editing interface with form validation - **`effect-log-modal.tsx`** - Effect logging modal with severity tracking - **`quick-effect-log.tsx`** - Streamlined effect logging interface - **`effect-timeline.tsx`** - Visual timeline for plan effects and outcomes #### Additional UI Components - **`ui/badge.tsx`** - Status and category badges - **`ui/sheet.tsx`** - Slide-in panel component - **`ui/tabs.tsx`** - Tab navigation component - **`ui/toast.tsx`** - Toast notification system - **Health data visualization components** - Charts and displays for health metrics - **Navigation components** - Header and bottom navigation for mobile-first design ### **Documentation** #### Project Information - **`README.md`** - Basic project overview with deployment information - **`CLAUDE.md`** - This comprehensive analysis document ## 4. API Endpoints Analysis ### Health Data Processing Endpoints #### `/api/intake` (POST) - **Purpose**: Primary endpoint for health data ingestion - **Input**: Multipart form data or JSON with text and files - **Processing**: Parses multimodal input (voice transcripts, images, documents) - **Output**: Structured health entries with generated tips - **Example Response**: \`\`\`json { "message": "Thanks for sharing! I've logged your health data...", "entries": [ {"category": "Sleep", "value": "7", "unit": "hours"}, {"category": "Hydration", "value": "650", "unit": "ml"} ], "tip": "Great job tracking your health!", "filesProcessed": 2 } \`\`\` #### AI-Powered Parser Endpoints - **`/api/parse-goals`** - Intelligent goal extraction using GPT-4o with Zod schema validation - Supports multimodal input (text, voice, image) - Extracts condition-specific goals with measurable targets - Returns structured goal objects with confidence scoring - **`/api/parse-medications`** - Smart medication parsing and structuring - Identifies medication names, dosages, frequencies, and timings - Categorizes by type (daily, PRN, supplement) - **`/api/parse-routines`** - Daily routine and habit extraction - Processes sleep schedules, hydration goals, exercise plans - Extracts quantified wellness targets - **`/api/parse-symptoms`** - Advanced symptom pattern analysis - Categorizes symptoms by severity and frequency - Identifies potential triggers and patterns #### Plan Management Endpoints - **`/api/plans` (GET)** - Retrieve all plans with optional status filtering - Query parameters: `?status=active|paused|ended` for filtering - Returns array of Plan objects with metadata - **`/api/plans` (POST)** - Create new treatment/routine plan - Accepts plan data: type, name, description, schedule, monitoring signals - Returns created plan with generated ID and timestamps - **`/api/plans/[id]` (GET)** - Retrieve specific plan with full details - Includes plan metadata, effect logs, and trend data - **`/api/plans/[id]` (PATCH)** - Update plan details or status - Supports status changes (active, paused, ended) - Updates metadata fields like notes, cadence, clinician - **`/api/plans/[id]/effects` (GET)** - Retrieve effect logs for specific plan - **`/api/plans/[id]/effects` (POST)** - Log new effect/outcome for plan - Records benefits, side effects, or neutral observations - Includes severity scoring (0-10) and linked health signals #### Data Retrieval Endpoints - **`/api/history`** - Retrieves historical health data with filtering options - **`/api/insights`** - Provides analytics and patterns from health data - **`/api/calendar`** - Calendar-integrated health event management ### Authentication & Authorization - **Current State**: No authentication system implemented - **Data Storage**: Uses localStorage for client-side persistence in mock mode - **Future Consideration**: Ready for backend integration with user authentication ## 5. Architecture Deep Dive ### Overall Application Architecture \`\`\` ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐ │ Next.js App │────▶│ API Routes │────▶│ Data Layer │ │ (Frontend) │ │ (Server-side) │ │ (Mock/Real API) │ │ │ │ │ │ │ │ • React 19 │ │ • Health Parser │ │ • MockStore │ │ • TypeScript │ │ • File Upload │ │ • API Client │ │ • Tailwind CSS │ │ • Data Proc. │ │ • localStorage │ │ • shadcn/ui │ │ • Response Gen. │ │ • Future DB │ └─────────────────┘ └─────────────────┘ └─────────────────┘ \`\`\` ### Data Flow and Request Lifecycle 1. **User Input** → Multimodal composer captures voice, text, files 2. **Client Processing** → Format data for API consumption 3. **API Route** → `/api/intake` processes multimodal input 4. **Data Extraction** → Parse health metrics from various input types 5. **Storage** → Save to mock store (or future database) 6. **Response Generation** → Create structured health entries + tips 7. **UI Update** → Display processed data in chat interface ### Key Design Patterns #### Environment-Based Architecture - **Dual Mode Operation**: Mock mode for development, API mode for production - **Configuration**: `NEXT_PUBLIC_DATA_SOURCE` environment variable switches modes - **Fallback Strategy**: Graceful degradation from API to mock mode on errors #### Component Architecture - **Atomic Design**: UI components organized in hierarchical structure - **Composition Pattern**: Layout components compose smaller UI elements - **Props Interface**: Strict TypeScript interfaces for component contracts #### Data Management - **SWR Integration**: Client-side caching and data fetching - **Local Storage**: Persistence layer for offline-first experience - **Type Safety**: Full TypeScript coverage for data models ## 6. Environment & Setup Analysis ### Required Environment Variables #### API Configuration \`\`\`bash # Backend API base URL (when using real API) NEXT_PUBLIC_API_BASE="" # Data source selection: "mock" or "api" NEXT_PUBLIC_DATA_SOURCE="mock" \`\`\` ### Installation and Setup Process #### Prerequisites - Node.js 18+ - pnpm (preferred package manager) #### Setup Steps \`\`\`bash # 1. Clone repository git clone # 2. Install dependencies pnpm install # 3. Configure environment (optional for mock mode) cp .env.example .env.local # Edit NEXT_PUBLIC_DATA_SOURCE and NEXT_PUBLIC_API_BASE # 4. Start development server pnpm dev # 5. Build for production pnpm build # 6. Start production server pnpm start \`\`\` ### Development Workflow #### Available Scripts - **`pnpm dev`** - Development server with hot reload - **`pnpm build`** - Production build - **`pnpm start`** - Production server - **`pnpm lint`** - Code linting with Next.js ESLint #### Mock Development Mode - **Default State**: Runs in mock mode without backend requirements - **Data Persistence**: Uses localStorage for client-side data persistence - **Full Functionality**: All features work with mock data for development ### Backend Integration Points #### API Client Setup (`lib/api/client.ts:3-4`) \`\`\`typescript const API_BASE = process.env.NEXT_PUBLIC_API_BASE || "" const DATA_SOURCE = (process.env.NEXT_PUBLIC_DATA_SOURCE || "mock") as "mock" | "api" \`\`\` #### Database Connection Strategy - **Current**: No direct database connections (stateless API approach) - **Architecture**: API routes would connect to external database/services - **Data Models**: Fully typed interfaces ready for database integration - **Future Integration**: Profile and health data persistence through API endpoints ## 7. Technology Stack Breakdown ### Runtime Environment - **Node.js**: 18+ for Next.js 15 compatibility - **Next.js 15.2.4**: React framework with App Router - **React 19**: Latest React with concurrent features ### Frontend Frameworks & Libraries - **UI Framework**: shadcn/ui + Radix UI primitives - **Styling**: Tailwind CSS 3.4+ with custom configuration - **Icons**: Lucide React for consistent iconography - **Theming**: next-themes for dark/light mode - **Forms**: React Hook Form + Zod validation - **Charts**: Recharts for data visualization ### Development Tools - **TypeScript 5**: Full type safety across the application - **PostCSS**: CSS processing pipeline - **Autoprefixer**: CSS vendor prefixing - **ESLint**: Code linting with Next.js configuration ### Data & State Management - **SWR**: Data fetching and client-side caching - **React Hook Form**: Form state management - **localStorage**: Client-side persistence - **Zod**: Runtime type validation ### AI & Processing - **@ai-sdk/openai**: OpenAI integration for health data processing - **ai**: Vercel AI SDK for model interactions and structured output generation - **zod**: Runtime schema validation for AI-generated structured data - **OpenAI GPT-4o**: Primary model for intelligent parsing and goal extraction ### Mobile & PWA - **Responsive Design**: Mobile-first Tailwind CSS approach - **Touch Interactions**: Optimized for mobile input methods - **PWA Ready**: Configuration for progressive web app features ## 8. Visual Architecture Diagram \`\`\` LINDA HEALTH COMPANION ARCHITECTURE ┌─────────────────────────────────────────────────────────────────────────────┐ │ FRONTEND LAYER │ │ ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐ │ │ │ Next.js App │ │ React Components│ │ UI Library │ │ │ │ Router │ │ │ │ │ │ │ │ • / │ │ • Layout │ │ • shadcn/ui │ │ │ │ • /profile │ │ • Multimodal │ │ • Radix UI │ │ │ │ • /history │ │ • Charts │ │ • Tailwind CSS │ │ │ │ • /settings │ │ • Navigation │ │ • Lucide Icons │ │ │ └─────────────────┘ └─────────────────┘ └─────────────────┘ │ └─────────────────────────────────────────────────────────────────────────────┘ │ ▼ ┌─────────────────────────────────────────────────────────────────────────────┐ │ API LAYER │ │ ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐ │ │ │ Health Data │ │ Specialized │ │ File & Media │ │ │ │ Processing │ │ Parsers │ │ Processing │ │ │ │ │ │ │ │ │ │ │ │ • /api/intake │ │ • parse-goals │ │ • Image Upload │ │ │ │ • /api/history │ │ • parse-meds │ │ • Voice Proc. │ │ │ │ • /api/insights │ │ • parse-symptoms│ │ • File Analysis │ │ │ │ • /api/calendar │ │ • parse-routines│ │ • Multipart │ │ │ └─────────────────┘ └─────────────────┘ └─────────────────┘ │ └─────────────────────────────────────────────────────────────────────────────┘ │ ▼ ┌─────────────────────────────────────────────────────────────────────────────┐ │ DATA LAYER │ │ ┌─────────────────┐ │ ┌─────────────────┐ │ │ │ MOCK MODE │ │ │ API MODE │ │ │ │ (Development) │ │ │ (Production) │ │ │ │ │ │ │ │ │ │ │ • MockStore │ │ │ • External API │ │ │ │ • localStorage │ │ │ • Database │ │ │ │ • Test Data │◄─────────────┼─────────────►│ • Real Backend │ │ │ │ • Simulation │ │ │ • Authentication│ │ │ └─────────────────┘ │ └─────────────────┘ │ │ │ │ │ Environment Variable: │ │ NEXT_PUBLIC_DATA_SOURCE │ └─────────────────────────────────────────────────────────────────────────────┘ DATA FLOW DIAGRAM User Input Processing Storage & Response │ │ │ ▼ ▼ ▼ ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐ │ Multimodal │ │ API Routes │ │ Data Store │ │ Composer │───▶│ • Parse Text │───▶│ • Profile │ │ • Voice │ │ • Process Files │ │ • Health Data │ │ • Text │ │ • Extract │ │ • Suggestions │ │ • Images │ │ Metrics │ │ • History │ │ • Documents │ │ • Generate Tips │ │ │ └─────────────────┘ └─────────────────┘ └─────────────────┘ │ │ │ │ │ │ ▼ ▼ ▼ ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐ │ User Interface │◄───│ Response Format │◄───│ Data Retrieval │ │ • Chat Messages │ │ • Health Entry │ │ • Query Store │ │ • Health Cards │ │ • Tips & Advice │ │ • Format Data │ │ • Visual Data │ │ • Structured │ │ • Apply Logic │ │ • Suggestions │ │ Response │ │ │ └─────────────────┘ └─────────────────┘ └─────────────────┘ \`\`\` ## 9. Key Insights & Recommendations ### Code Quality Assessment #### Strengths - **Modern Stack**: Uses latest React 19 and Next.js 15 features - **Type Safety**: Comprehensive TypeScript implementation - **Component Architecture**: Well-structured, reusable components - **Responsive Design**: Mobile-first approach with excellent UX - **Environment Flexibility**: Sophisticated mock/API switching system #### Areas for Improvement - **Error Handling**: Limited error boundaries and user feedback - **Testing**: No test suite identified in current codebase - **Accessibility**: Could benefit from ARIA labels and screen reader support - **Performance**: Image optimization and lazy loading opportunities ### Security Considerations #### Current Security Posture - **No Authentication**: Currently operates without user authentication - **Client-Side Storage**: Sensitive health data stored in localStorage - **File Upload**: Basic file handling without security validation - **Environment Variables**: Proper separation of config from code #### Security Recommendations - **Implement Authentication**: Add user authentication system - **Data Encryption**: Encrypt sensitive health data at rest - **Input Validation**: Add comprehensive server-side validation - **File Security**: Implement file type validation and virus scanning - **HTTPS Enforcement**: Ensure all production traffic uses HTTPS ### Performance Optimization Opportunities #### Frontend Performance - **Code Splitting**: Implement route-based code splitting - **Image Optimization**: Add Next.js Image component usage - **Bundle Analysis**: Monitor and optimize bundle size - **Lazy Loading**: Implement lazy loading for components and routes #### Data Management - **Caching Strategy**: Enhance SWR caching configuration - **Offline Support**: Implement service worker for offline functionality - **Background Sync**: Add background data synchronization ### Maintainability Suggestions #### Code Organization - **API Layer**: Create dedicated API service layer - **Constants**: Centralize configuration and constants - **Utilities**: Expand utility functions library - **Documentation**: Add inline code documentation #### Development Workflow - **Testing Suite**: Implement Jest + React Testing Library - **CI/CD Pipeline**: Add automated testing and deployment - **Code Standards**: Establish consistent coding standards - **Type Coverage**: Achieve 100% TypeScript coverage ### Scalability Considerations #### Backend Integration - **Database Design**: Plan schema for user profiles and health data - **API Architecture**: Design RESTful or GraphQL API structure - **Microservices**: Consider service-oriented architecture for scale - **Real-time Features**: Plan WebSocket integration for live features #### Feature Extensions - **Multi-user Support**: Design for family/caregiver access - **Healthcare Integration**: Plan for EHR and device integrations - **AI Enhancement**: Expand AI-powered health insights - **Notification System**: Implement push notifications ## 10. Recent Updates (Updated: 2025-01-11) ### Major Features Added #### Comprehensive Plans Management System - **Plans Dashboard** (`app/plans/page.tsx`) - Visual plan management with active/paused/ended status tracking - Quick effect logging and plan editing capabilities - Floating action button for new plan creation - Separate sections for active plans and recently ended plans - **Individual Plan Details** (`app/plans/[id]/page.tsx`) - Detailed plan view with effect timeline and analytics - Comprehensive effect logging with severity tracking - Plan modification and status management - Visual trend data for monitored health signals - **Plan Creation & Management** - Multi-step plan wizard with validation (`new-plan-wizard.tsx`) - Support for multiple plan types: medication, physical therapy, acupuncture, routines - Clinician assignment and note management - Customizable monitoring signals and cadence settings #### Effect Logging & Analytics - **Effect Tracking System** - Benefit, side-effect, and neutral observation logging - Severity scoring (0-10 scale) with linked health signals - Quick logging interface for streamlined data entry - Timeline visualization of effects over time - **Plan Analytics** - Trend data visualization for monitored signals - Effect correlation analysis - Plan effectiveness tracking and outcomes display #### AI-Powered Onboarding System - **Enhanced Onboarding Flow** (`app/onboarding/page.tsx`) - Comprehensive health profile setup with step-by-step guidance - AI-powered parsing of user input for automated profile completion - Support for voice, text, and image input modalities #### Intelligent Health Data Parsing - **Goal Extraction System** (`app/api/parse-goals/route.ts`) - GPT-4o integration for intelligent goal parsing from natural language - Zod schema validation for structured output consistency - Support for condition-specific goal setting with measurable targets - Multimodal input processing (text, voice, image descriptions) - **Medication Management** (`app/api/parse-medications/route.ts`) - Smart medication parsing with dosage and timing extraction - Automatic categorization (daily, PRN, supplements) - Structured medication profiles for tracking adherence - **Routine & Symptom Analysis** (`app/api/parse-routines/route.ts`, `app/api/parse-symptoms/route.ts`) - Daily routine optimization with quantified wellness targets - Advanced symptom pattern recognition and trigger identification - Personalized health coaching recommendations #### Enhanced UI Components - **Plan Management Components** - `plan-card.tsx` - Rich plan display with status management and quick actions - `new-plan-wizard.tsx` - Multi-step plan creation with comprehensive validation - `edit-plan-sheet.tsx` - Slide-in plan editing interface - `effect-log-modal.tsx` - Detailed effect logging with severity tracking - `quick-effect-log.tsx` - Streamlined effect logging for rapid data entry - `effect-timeline.tsx` - Visual timeline for tracking plan effects over time - **Specialized Input Components** - `goal-input.tsx` - AI-powered goal setting with real-time parsing - `medication-input.tsx` - Smart medication entry and management - `routine-input.tsx` - Daily routine configuration interface - `symptom-input.tsx` - Comprehensive symptom tracking - **User Experience Enhancements** - `first-run-nudge.tsx` - Improved onboarding experience - `profile-coach.tsx` - AI coaching and personalized guidance - Enhanced UI components: `ui/badge.tsx`, `ui/sheet.tsx`, `ui/tabs.tsx`, `ui/toast.tsx` - Enhanced form validation and error handling across components #### Architecture Improvements - **Plan Management Architecture**: Complete CRUD system for treatment plans and effect logging - **Enhanced Data Models**: New TypeScript interfaces for Plan, EffectLog, and PlanTrendData - **React Hook Integration**: SWR-powered data fetching with `usePlans` hook - **Structured AI Output**: Integration of Zod schemas for reliable AI response parsing - **Enhanced Error Handling**: Comprehensive error boundaries and user feedback - **Type Safety**: Extended TypeScript coverage for AI-generated and plan-related data structures - **Performance Optimizations**: Improved component rendering and data caching - **Toast Notification System**: Real-time feedback for user actions and plan updates ### Breaking Changes - Enhanced onboarding flow may require existing users to complete profile setup - New AI parsing endpoints require OpenAI API key configuration - Updated TypeScript interfaces for Profile and health data structures ### Integration Points - **New Environment Variables Required**: \`\`\`bash OPENAI_API_KEY="your-openai-api-key" # Required for AI parsing features \`\`\` - **Updated Data Models**: Enhanced Profile interface with detailed health goal structures - **New API Dependencies**: @ai-sdk/openai and zod packages for AI functionality --- ## Frontend Architecture Entry Points ### Main Application Entry Point - **File**: `app/layout.tsx:14-33` - **Purpose**: Root layout wrapper with theme provider - **Key Features**: Theme switching, font loading, HTML structure ### Chat Interface (Primary User Interaction) - **File**: `app/page.tsx:36-239` - **Purpose**: Main health data input interface - **Key Features**: Multimodal input, chat UI, health entry editing ### Plans Management Interface - **File**: `app/plans/page.tsx:16-220` - **Purpose**: Treatment plan management dashboard - **Key Features**: Plan creation, status management, effect logging, analytics - **Navigation**: Accessible via bottom navigation bar (Plans tab) ### Individual Plan Details - **File**: `app/plans/[id]/page.tsx` - **Purpose**: Detailed plan view with comprehensive tracking - **Key Features**: Effect timeline, trend visualization, plan modifications ### API Connection Configuration - **File**: `lib/api/client.ts:3-4` - **Environment Variables**: \`\`\`bash NEXT_PUBLIC_API_BASE="" # Backend API URL NEXT_PUBLIC_DATA_SOURCE="mock" # "mock" or "api" \`\`\` ### Component Layout System - **File**: `components/layout.tsx:15-43` - **Purpose**: Consistent app layout with navigation - **Features**: Responsive design, action bar integration ## Database Connection Entry Points ### Current State: Mock Data System - **File**: `lib/mocks/index.ts:31-285` - **Purpose**: Comprehensive mock data store for development - **Features**: Profile simulation, health data parsing, suggestion generation ### API Client Architecture - **File**: `lib/api/client.ts:32-200` - **Purpose**: Environment-based API client with fallback to mock data - **Integration Points**: - Profile management (`getProfile`, `patchProfile`) - Health data processing (`parseOnboarding`) - Suggestion system (`listSuggestions`, `actOnSuggestion`) ### Plans Management System - **File**: `lib/api/usePlans.ts` - **Purpose**: React hook for plan management with SWR caching - **Features**: CRUD operations, status updates, effect logging - **Mock Store**: `lib/mocks/index.ts` - Enhanced with plan and effect management ### Future Database Integration Strategy 1. **Replace Mock Store**: Implement real API endpoints 2. **Environment Configuration**: Set `NEXT_PUBLIC_DATA_SOURCE="api"` 3. **API Base URL**: Configure `NEXT_PUBLIC_API_BASE` to backend service 4. **Data Models**: Use existing TypeScript interfaces in `lib/api/types.ts` 5. **Database Schema**: Design tables for Profile, HealthData, Suggestions entities The application is architecturally prepared for seamless backend integration while maintaining full functionality in mock mode for development. [V0_FILE]markdown:file="PLANS.md" # Master Health Companion Agent - Architecture & Implementation Plan ## Executive Summary This document outlines the integration strategy for building a master health companion agent using the AGNO framework to power the existing Linda Health Companion frontend application. The plan focuses on creating a sophisticated AI agent that provides intelligent health coaching, data analysis, and personalized recommendations while maintaining clear architectural boundaries and scalability. ## 1. Architecture Overview ### Current State Analysis - **Frontend**: Next.js 15.2.4 + React 19 with comprehensive health tracking UI - **Data Layer**: Mock-based development with API abstraction layer ready - **Features**: Plans management, multimodal input, AI parsing, effect tracking - **Gap**: No intelligent backend agent for health coaching and analysis ### **Refined Approach (Based on GPT Analysis)** **Key Principle**: **Preserve existing frontend contracts** while adding intelligent backend capabilities - **No Frontend Changes**: Existing `/api/*` routes stay identical - **Single Master Agent**: One "Linda Master Agent" with comprehensive health tools - **Stable Tool Catalog**: Well-defined I/O contracts for consistent behavior - **Provenance-First**: Every response includes data sources and uncertainty levels - **Mock/Agent Switch**: `NEXT_PUBLIC_DATA_SOURCE` toggles between mock and agent backend ### Target Architecture \`\`\` LINDA HEALTH COMPANION SYSTEM $ FRONTEND LAYER Next.js App React UI API Client (Port 3000) Components Layer " Chat Tab " Multimodal " AGNO Client " Today Tab " Plans UI " Health API " Plans Tab " History UI " Data Sync " History Tab " Analytics � HTTP/WebSocket API BACKEND LAYER AGNO AgentOS Master Health Custom Tools (Port 7777) Agent Ecosystem " Agent Runtime " Health Coach " Plan Manager " Memory System " Data Analyst " Effect Logger " Session Mgmt " Goal Tracker " Trend Analyzer " Control Plane " Recommender " Calendar Tool � DATA LAYER PostgreSQL Health Data Agent Memory Database Schema System " User Profiles " Health Metrics " Chat History " Plans & Goals " Effect Logs " User Context " Health History " Trends Data " Preferences " Agent Sessions " Calendar Data " Learning Data \`\`\` ## 2. Monorepo Structure Strategy ### Recommended Directory Structure \`\`\` linda_healthcompanion/ frontend/ # Next.js Application app/ # Next.js App Router components/ # React Components lib/ # Frontend utilities public/ # Static assets backend/ # AGNO Agent System agents/ # Agent Definitions health_master_agent.py # Main health companion agent plan_manager_agent.py # Plans management specialist data_analyst_agent.py # Health data analysis agent tools/ # Custom AGNO Tools health_data_tool.py # Health data CRUD operations plan_management_tool.py # Plan tracking and updates effect_logger_tool.py # Effect logging and analysis calendar_tool.py # Calendar integration notification_tool.py # User notifications models/ # Data Models & Schemas health_models.py # Health data Pydantic models plan_models.py # Plan and effect models user_models.py # User profile models config/ # Configuration agent_config.py # Agent settings database_config.py # DB connection config tool_config.py # Tool configurations agentos_app.py # Main AGNO application requirements.txt # Python dependencies shared/ # Shared Resources database/ # Database schemas and migrations migrations/ # Database migration files schemas/ # SQL schema definitions seed_data.sql # Initial data setup api_contracts/ # API interface definitions health_api.ts # TypeScript API interfaces agent_api.py # Python API interfaces types/ # Shared type definitions health_types.ts # Health data types (TS) health_types.py # Health data types (Python) docs/ # Documentation ARCHITECTURE.md # Detailed architecture docs API_REFERENCE.md # API documentation DEPLOYMENT.md # Deployment instructions DEVELOPMENT.md # Development setup guide scripts/ # Development & Deployment Scripts setup_dev.sh # Development environment setup migrate_db.py # Database migration runner deploy_backend.sh # Backend deployment deploy_frontend.sh # Frontend deployment docker-compose.yml # Multi-service Docker setup docker-compose.dev.yml # Development Docker setup README.md # Project overview \`\`\` ### Benefits of This Structure - **Clear Separation**: Frontend and backend are distinct but coordinated - **Shared Resources**: Common types and database schemas prevent duplication - **Scalability**: Each service can be deployed and scaled independently - **Development Workflow**: Maintains clear boundaries for team collaboration - **Monorepo Advantages**: Single repository for coordinated development ## 2.1 Stable Tool Catalog (Core Innovation) ### **Linda Master Agent Tool Interface** Based on GPT analysis, the agent exposes a **stable, well-defined tool catalog** with consistent I/O: \`\`\`python # Core Health Tools with Stable Contracts class LindaHealthTools: # Data Intake & Processing def intake_parse_and_extract(self, text: str, files: List[File]) -> StructuredHealthData: """Extract structured health metrics from multimodal input""" return { "entries": [...], # Structured health entries "confidence": 0.95, "provenance": {"source": "voice_note", "timestamp": "..."}, "uncertainties": ["medication_timing_unclear"] } # Database Operations def db_store(self, user_id: str, data: HealthEntry) -> StorageResult: """Store health data with provenance tracking""" def db_get(self, user_id: str, filters: Dict) -> List[HealthEntry]: """Retrieve health data with full provenance""" # Analytics & Trends def analytics_trend(self, metric: str, window: int) -> TrendAnalysis: """Compute trends with uncertainty bounds and contributing data""" return { "trend": "improving", "confidence": 0.85, "window": "14_days", "contributing_entry_ids": ["entry_123", "entry_124"], "uncertainty_notes": ["limited_data_points"] } def analytics_association(self, target: str, candidates: List[str], window: int): """Find correlations (NOT causation) between health metrics""" return { "associations": [{"metric": "sleep", "correlation": 0.7, "note": "association_only"}], "provenance": {...} } # Plan Evaluation (New Core Capability) def plans_evaluate(self, plan_id: str, window: int) -> PlanEvaluation: """Evaluate plan effectiveness with provenance""" return { "verdict": "working_well", # working_well|unclear|needs_adjustment "confidence": 0.82, "adherence_rate": 0.85, "recommendations": [ "Continue current approach", "Consider tracking sleep quality for better insights" ], "provenance": { "window": "28_days", "entry_ids": ["entry_234", "entry_235"], "baseline_period": "14_days_before_start" }, "safety_flags": [] } # Coaching & Recommendations def coaching_generate(self, context: UserContext) -> CoachingResponse: """Generate personalized, educational health coaching""" return { "recommendations": [ { "action": "Log sleep quality for 1 week", "rationale": "Better data for trend analysis", "safety_level": "low_risk" } ], "educational_note": "This is guidance, not medical advice", "red_flags": [] # Only populated when explicit criteria met } \`\`\` ### **Response Envelope (All Agent Outputs)** Every agent response follows this provenance-first structure: \`\`\`python { "data": { "type": "plan_evaluation", "summary": "Plan showing positive trends with high adherence", "structured": {...}, # Tool-specific data "recommendations": [...], }, "provenance": { "window": "28_days", "entry_ids": ["entry_123", "entry_124"], "data_quality": "sufficient" }, "uncertainties": [ "Limited weekend data", "One measurement seems anomalous" ], "safety": { "level": "educational_only", "red_flags": [], "disclaimer": "This is informational guidance, consult healthcare providers for medical decisions" }, "meta": { "timestamp": "2025-01-11T...", "agent_version": "1.0", "confidence": 0.85 } } ## 3. Tab Integration Strategy ### 3.1 Chat Tab Integration **Purpose**: Primary interface for user-agent interaction with intelligent health coaching **Architecture**: \`\`\`typescript // Frontend: Real-time chat with AGNO agent const chatWithAgent = async (message: string, files?: File[]) => { const response = await fetch(`${AGNO_BACKEND_URL}/chat`, { method: 'POST', headers: { 'Content-Type': 'application/json' }, body: JSON.stringify({ message, user_id: getUserId(), session_id: getCurrentSessionId(), multimodal_data: files ? await processFiles(files) : null }) }); return response.json(); }; \`\`\` **AGNO Agent Capabilities**: - **Health Coaching**: Personalized advice based on user history and current symptoms - **Data Interpretation**: Analyze uploaded health data (blood pressure logs, photos, documents) - **Goal Setting**: Interactive goal setting and plan creation - **Symptom Analysis**: Intelligent symptom tracking and pattern recognition - **Medication Management**: Reminders and interaction checking ### 3.2 Today Tab Integration **Purpose**: Display AI-processed daily health insights and extracted data **Data Flow**: \`\`\` Chat Interaction � AGNO Agent Processing � Structured Data Extraction � Today Tab Display \`\`\` **AGNO Processing Pipeline**: 1. **Data Extraction**: Extract health metrics from chat conversations 2. **Trend Analysis**: Compare with historical data for insights 3. **Daily Summary**: Generate personalized daily health summary 4. **Action Items**: Suggest specific actions based on today's data 5. **Progress Updates**: Track goal progress and plan adherence **Frontend Integration**: - Real-time updates from agent analysis - Visual dashboards powered by agent insights - Actionable recommendations from agent ### 3.3 Plans Tab Integration **Purpose**: AI-powered plan management with intelligent tracking and recommendations **AGNO Agent Enhancements**: - **Plan Optimization**: Analyze plan effectiveness and suggest modifications - **Effect Correlation**: Identify patterns between treatments and outcomes - **Adherence Monitoring**: Track plan compliance and provide coaching - **Plan Interaction Analysis**: Detect potential conflicts between multiple plans - **Personalized Adjustments**: Recommend plan modifications based on user responses **Integration Points**: \`\`\`python # AGNO Tool: Intelligent Plan Management class IntelligentPlanTool: def analyze_plan_effectiveness(self, plan_id: str, user_id: str): # Analyze effects, adherence, and outcomes return plan_analysis def recommend_plan_adjustments(self, plan_id: str, recent_effects: List[Effect]): # AI-powered plan optimization return recommendations def detect_plan_conflicts(self, user_plans: List[Plan]): # Identify potential interactions or conflicts return conflict_analysis \`\`\` ### 3.4 History Tab Integration **Purpose**: Intelligent historical data analysis and pattern recognition **AGNO Agent Features**: - **Pattern Recognition**: Identify long-term trends and correlations - **Predictive Analytics**: Forecast potential health events or flare-ups - **Historical Insights**: Generate insights from long-term data patterns - **Comparative Analysis**: Compare current status with historical baselines - **Data Storytelling**: Create narrative summaries of health journey ## 4. Database Architecture ### 4.1 Shared Database Strategy **Database Choice**: PostgreSQL for robust health data storage with AGNO integration **Core Tables**: \`\`\`sql -- User management users (id, email, profile_data, created_at, updated_at) -- Health data storage health_entries (id, user_id, entry_date, data_json, source, agent_processed) plans (id, user_id, plan_type, config_json, status, created_by_agent) plan_effects (id, plan_id, effect_data, severity, logged_at, agent_analyzed) -- Agent-specific tables agent_sessions (id, user_id, session_data, created_at, updated_at) agent_memories (id, user_id, memory_content, context, created_at) agent_insights (id, user_id, insight_data, confidence_score, generated_at) -- Analytics and trends health_trends (id, user_id, metric_type, trend_data, period, computed_at) \`\`\` ### 4.2 Data Synchronization Strategy **Bidirectional Sync**: Frontend � Database � AGNO Agent \`\`\`python # AGNO Agent: Health data tool with database integration class HealthDataTool: def __init__(self, db_connection): self.db = db_connection def log_health_data(self, user_id: str, data: HealthEntry): # Store in database and update agent context return self.db.store_health_entry(user_id, data) def get_user_health_history(self, user_id: str, days: int = 30): # Retrieve and format for agent processing return self.db.get_health_history(user_id, days) def update_agent_memories(self, user_id: str, insights: List[str]): # Update agent's memory with new insights return self.db.store_agent_memories(user_id, insights) \`\`\` ## 5. Implementation Phases ### Phase 1: Foundation Setup (Weeks 1-2) - **Monorepo Structure**: Implement recommended directory structure - **Database Setup**: PostgreSQL schema creation and migration system - **Basic AGNO Agent**: Simple health companion agent with database connection - **API Gateway**: Basic communication layer between frontend and AGNO - **Development Environment**: Docker Compose setup for local development ### Phase 2: Linda Master Agent (Weeks 3-5) - **Single Agent Architecture**: One comprehensive Linda Master Agent with all health capabilities - **Stable Tool Catalog**: Implement the 5 core tools (intake, db, analytics, plans, coaching) - **Provenance System**: Every response includes data sources, confidence, and uncertainty - **API Forwarding**: Frontend routes forward to agent backend when `DATA_SOURCE="api"` - **Memory & Context**: User-scoped health history and session persistence ### Phase 3: Tab Integration (Weeks 6-8) - **Today Tab**: AI-powered daily insights and extracted data display - **Plans Tab**: Intelligent plan management with agent recommendations - **History Tab**: Pattern recognition and historical analysis integration - **Real-time Updates**: WebSocket integration for live agent interactions - **Data Synchronization**: Bidirectional data flow optimization ### Phase 4: Advanced Features (Weeks 9-12) - **Predictive Analytics**: Health trend forecasting and early warning systems - **Multi-Agent System**: Specialized agents for different health domains - **Advanced Coaching**: Personalized health coaching with behavioral insights - **Integration APIs**: External health device and service integrations - **Performance Optimization**: Caching, indexing, and query optimization ### Phase 5: Production Readiness (Weeks 13-16) - **Security Implementation**: Authentication, authorization, data encryption - **Monitoring & Logging**: Comprehensive system monitoring and alerting - **Deployment Pipeline**: CI/CD for both frontend and backend services - **Performance Testing**: Load testing and optimization - **Documentation**: Complete API documentation and user guides ## 6. Technical Considerations ### 6.1 Communication Architecture - **Primary Protocol**: RESTful API with JSON payloads - **Real-time Updates**: WebSocket for live agent interactions - **Authentication**: JWT tokens for secure communication - **Rate Limiting**: Protect agent resources from excessive requests ### 6.2 Agent Memory & Context Management - **Session Persistence**: AGNO's built-in session management - **User Context**: Comprehensive user health profiles in agent memory - **Historical Context**: Access to complete health history for informed responses - **Privacy Controls**: User control over what data the agent remembers ### 6.3 Scalability Planning - **Horizontal Scaling**: Multiple AGNO agent instances behind load balancer - **Database Optimization**: Indexed queries and caching for performance - **Message Queuing**: Asynchronous processing for complex health analyses - **Microservices Ready**: Architecture supports future service decomposition ### 6.4 Data Privacy & Security - **Data Encryption**: All health data encrypted at rest and in transit - **Access Controls**: Role-based access to different types of health data - **Audit Logging**: Complete audit trail of all data access and modifications - **HIPAA Compliance**: Architecture designed for healthcare data compliance ## 7. Success Metrics ### Technical Metrics - **Response Time**: < 2 seconds for agent responses - **Uptime**: 99.9% system availability - **Data Accuracy**: > 95% accuracy in health data extraction - **User Engagement**: Daily active user growth ### Health Outcomes - **Goal Achievement**: % of users meeting their health goals - **Plan Adherence**: Treatment plan compliance rates - **User Satisfaction**: Net Promoter Score (NPS) tracking - **Health Insights**: Quality and relevance of agent recommendations ## 8. Risk Mitigation ### Technical Risks - **Agent Reliability**: Comprehensive testing and fallback mechanisms - **Database Performance**: Monitoring and optimization strategies - **Integration Complexity**: Phased rollout with extensive testing - **Data Migration**: Safe migration strategies for existing user data ### Business Risks - **User Adoption**: Gradual feature rollout with user feedback integration - **Data Privacy**: Proactive compliance and security measures - **Regulatory Compliance**: Healthcare regulation adherence - **Scalability Planning**: Infrastructure planning for user growth ## 8.1 Refined Success Criteria (GPT-Aligned) Based on the GPT analysis, the implementation should deliver these **specific, testable outcomes**: ### **Historical Analysis Capability** \`\`\`typescript // From Chat: "What changed this week?" const response = await chatWithAgent("What changed this week?"); // Expected response structure: { data: { type: "historical_analysis", summary: "Sleep improved 15%, pain decreased 20% vs last week", trends: [{metric: "sleep", change: "+15%", confidence: 0.85}], recommendations: ["Continue current sleep routine"] }, provenance: { window: "7_days", entry_ids: ["entry_123", "entry_124"], baseline: "previous_7_days" }, uncertainties: ["Weekend data incomplete"] } \`\`\` ### **Plan Evaluation Capability** \`\`\`typescript // From Plans: Request evaluation for last 28 days const evaluation = await evaluatePlan(planId, 28); // Expected response: { data: { type: "plan_evaluation", verdict: "working_well", adherence_rate: 0.85, recommendations: [ "Continue current approach", "Consider adding sleep quality tracking" ] }, provenance: { window: "28_days", entry_ids: ["entry_234", "entry_235"], baseline_period: "14_days_before_start" }, confidence: 0.82 } \`\`\` ### **Contract Preservation** - **✅ Zero Frontend Changes**: All existing Chat/Today/Plans/History tabs work identically - **✅ Mock Mode Preserved**: Development continues with mock data - **✅ API Routes Unchanged**: `/api/intake`, `/api/insights`, `/api/history`, `/api/plans/*` maintain contracts - **✅ Environment Switch**: `NEXT_PUBLIC_DATA_SOURCE="api"` enables agent backend ## 9. Next Steps 1. **Immediate Actions (Refined Approach)**: - Set up simplified monorepo structure with `backend/` folder - Create Linda Master Agent with stable tool catalog - Implement API forwarding layer (mock → agent switch) - Establish basic provenance tracking system 2. **Week 1 Deliverables**: - Working Linda Master Agent with 5 core tools - API forwarding that preserves frontend contracts - Basic historical analysis ("What changed this week?") - Plan evaluation prototype with provenance 3. **Success Validation**: - Chat tab can ask "What changed this week?" and get insights with provenance - Plans tab can request evaluation and get verdict + recommendations - All existing functionality works identically in both mock and agent modes - Response envelope includes data sources, confidence, and uncertainty This architecture provides a robust foundation for building an intelligent health companion that enhances the existing Linda Health Companion with powerful AI capabilities while maintaining clear architectural boundaries and scalability for future growth. [V0_FILE]markdown:file="README.md" # Linda Health Companion *Automatically synced with your [v0.dev](https://v0.dev) deployments* [![Deployed on Vercel](https://img.shields.io/badge/Deployed%20on-Vercel-black?style=for-the-badge&logo=vercel)](https://vercel.com/wangnanahf-7914s-projects/v0-linda-health-companion) [![Built with v0](https://img.shields.io/badge/Built%20with-v0.dev-black?style=for-the-badge)](https://v0.dev/chat/projects/o4GmFYDC6Vu) ## Overview This repository will stay in sync with your deployed chats on [v0.dev](https://v0.dev). Any changes you make to your deployed app will be automatically pushed to this repository from [v0.dev](https://v0.dev). ## Deployment Your project is live at: **[https://vercel.com/wangnanahf-7914s-projects/v0-linda-health-companion](https://vercel.com/wangnanahf-7914s-projects/v0-linda-health-companion)** ## Build your app Continue building your app on: **[https://v0.dev/chat/projects/o4GmFYDC6Vu](https://v0.dev/chat/projects/o4GmFYDC6Vu)** ## How It Works 1. Create and modify your project using [v0.dev](https://v0.dev) 2. Deploy your chats from the v0 interface 3. Changes are automatically pushed to this repository 4. Vercel deploys the latest version from this repository [V0_FILE]markdown:file="agno/advanced/access_reasoning_content_from_the_agent_s_run_response_after_streaming.md" --- title: Access reasoning_content from the agent's run_response after streaming category: advanced source_lines: 21757-21802 line_count: 45 --- # Access reasoning_content from the agent's run_response after streaming print("\n--- reasoning_content from agent.run_response after streaming ---") if ( hasattr(streaming_agent_with_model, "run_response") and streaming_agent_with_model.run_response and hasattr(streaming_agent_with_model.run_response, "reasoning_content") and streaming_agent_with_model.run_response.reasoning_content ): print(streaming_agent_with_model.run_response.reasoning_content) else: print("No reasoning_content found in agent.run_response after streaming") \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno \`\`\` \`\`\`bash Mac python cookbook/reasoning/agents/capture_reasoning_content_default_COT.py \`\`\` \`\`\`bash Windows python cookbook/reasoning/agents/capture_reasoning_content_default_COT.py \`\`\` [V0_FILE]markdown:file="agno/advanced/add_context.md" --- title: Add Context category: advanced source_lines: 11439-11483 line_count: 44 --- # Add Context Source: https://docs.agno.com/examples/concepts/context/01-add_context This example demonstrates how to create a context-aware agent that can access real-time data from HackerNews. ## Code \`\`\`python cookbook/agent_concepts/context/01-add_context.py import json from textwrap import dedent import httpx from agno.agent import Agent from agno.models.openai import OpenAIChat def get_top_hackernews_stories(num_stories: int = 5) -> str: """Fetch and return the top stories from HackerNews. Args: num_stories: Number of top stories to retrieve (default: 5) Returns: JSON string containing story details (title, url, score, etc.) """ # Get top stories stories = [ { k: v for k, v in httpx.get( f"https://hacker-news.firebaseio.com/v0/item/{id}.json" ) .json() .items() if k != "kids" # Exclude discussion threads } for id in httpx.get( "https://hacker-news.firebaseio.com/v0/topstories.json" ).json()[:num_stories] ] return json.dumps(stories, indent=4) [V0_FILE]markdown:file="agno/advanced/agent_context.md" --- title: Agent Context category: advanced source_lines: 33838-33889 line_count: 51 --- # Agent Context Source: https://docs.agno.com/examples/getting-started/agent-context This example shows how to inject external dependencies into an agent. The context is evaluated when the agent is run, acting like dependency injection for Agents. Example prompts to try: * "Summarize the top stories on HackerNews" * "What are the trending tech discussions right now?" * "Analyze the current top stories and identify trends" * "What's the most upvoted story today?" ## Code \`\`\`python agent_context.py import json from textwrap import dedent import httpx from agno.agent import Agent from agno.models.openai import OpenAIChat def get_top_hackernews_stories(num_stories: int = 5) -> str: """Fetch and return the top stories from HackerNews. Args: num_stories: Number of top stories to retrieve (default: 5) Returns: JSON string containing story details (title, url, score, etc.) """ # Get top stories stories = [ { k: v for k, v in httpx.get( f"https://hacker-news.firebaseio.com/v0/item/{id}.json" ) .json() .items() if k != "kids" # Exclude discussion threads } for id in httpx.get( "https://hacker-news.firebaseio.com/v0/topstories.json" ).json()[:num_stories] ] return json.dumps(stories, indent=4) [V0_FILE]markdown:file="agno/advanced/agent_that_uses_structured_outputs_with_streaming.md" --- title: Agent that uses structured outputs with streaming category: advanced source_lines: 2462-2481 line_count: 19 --- # Agent that uses structured outputs with streaming structured_output_agent = Agent( model=OpenAIChat(id="gpt-4o"), description="You write movie scripts.", response_model=MovieScript, ) structured_output_agent.print_response( "New York", stream=True, stream_intermediate_steps=True ) \`\`\` ## Developer Resources * View [Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/async/structured_output.py) * View [Parser Model Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/other/parse_model.py) * View [Streaming Structured Output](https://github.com/agno-agi/agno/blob/main/cookbook/models/openai/chat/structured_output_stream.py) [V0_FILE]markdown:file="agno/advanced/agent_with_streaming.md" --- title: Agent with Streaming category: advanced source_lines: 48407-48455 line_count: 48 --- # Agent with Streaming Source: https://docs.agno.com/examples/models/vllm/basic_stream ## Code \`\`\`python cookbook/models/vllm/basic_stream.py from agno.agent import Agent from agno.models.vllm import vLLM agent = Agent( model=vLLM(id="Qwen/Qwen2.5-7B-Instruct", top_k=20, enable_thinking=False), markdown=True, ) agent.print_response("Share a 2 sentence horror story", stream=True) \`\`\` ## Usage \`\`\`bash pip install -U agno openai vllm \`\`\` \`\`\`bash vllm serve Qwen/Qwen2.5-7B-Instruct \ --enable-auto-tool-choice \ --tool-call-parser hermes \ --dtype float16 \ --max-model-len 8192 \ --gpu-memory-utilization 0.9 \`\`\` \`\`\`bash python cookbook/models/vllm/basic_stream.py \`\`\` [V0_FILE]markdown:file="agno/advanced/async_agent_with_streaming.md" --- title: Async Agent with Streaming category: advanced source_lines: 48305-48352 line_count: 47 --- # Async Agent with Streaming Source: https://docs.agno.com/examples/models/vllm/async_basic_stream ## Code \`\`\`python cookbook/models/vllm/async_basic_stream.py import asyncio from agno.agent import Agent from agno.models.vllm import vLLM agent = Agent(model=vLLM(id="Qwen/Qwen2.5-7B-Instruct"), markdown=True) asyncio.run(agent.aprint_response("Share a 2 sentence horror story", stream=True)) \`\`\` ## Usage \`\`\`bash pip install -U agno openai vllm \`\`\` \`\`\`bash vllm serve Qwen/Qwen2.5-7B-Instruct \ --enable-auto-tool-choice \ --tool-call-parser hermes \ --dtype float16 \ --max-model-len 8192 \ --gpu-memory-utilization 0.9 \`\`\` \`\`\`bash python cookbook/models/vllm/async_basic_stream.py \`\`\` [V0_FILE]markdown:file="agno/advanced/async_streaming_agent.md" --- title: Async Streaming Agent category: advanced source_lines: 48935-48951 line_count: 16 --- # Async Streaming Agent Source: https://docs.agno.com/examples/models/xai/basic_async_stream ## Code \`\`\`python cookbook/models/xai/basic_async_stream.py import asyncio from typing import Iterator from agno.agent import Agent, RunResponseEvent from agno.models.xai import xAI agent = Agent(model=xAI(id="grok-3"), markdown=True) [V0_FILE]markdown:file="agno/advanced/asynchronous_streaming.md" --- title: Asynchronous streaming category: advanced source_lines: 70372-70384 line_count: 12 --- # Asynchronous streaming async for chunk in await team.arun("What is the weather in Tokyo?", stream=True): print(chunk.content, end="", flush=True) \`\`\` ## Streaming Intermediate Steps Throughout the execution of a team, multiple events take place, and we provide these events in real-time for enhanced team transparency. You can enable streaming of intermediate steps by setting `stream_intermediate_steps=True`. \`\`\`python [V0_FILE]markdown:file="agno/advanced/audio_streaming_agent.md" --- title: Audio Streaming Agent category: advanced source_lines: 18553-18567 line_count: 14 --- # Audio Streaming Agent Source: https://docs.agno.com/examples/concepts/multimodal/audio-streaming ## Code \`\`\`python import base64 import wave from agno.agent import Agent from agno.models.openai import OpenAIChat from typing import Iterator [V0_FILE]markdown:file="agno/advanced/basic_streaming.md" --- title: Basic Streaming category: advanced source_lines: 39084-39099 line_count: 15 --- # Basic Streaming Source: https://docs.agno.com/examples/models/azure/openai/basic_stream ## Code \`\`\`python cookbook/models/azure/openai/basic_stream.py from typing import Iterator # noqa from agno.agent import Agent, RunResponse # noqa from agno.models.azure import AzureOpenAI agent = Agent(model=AzureOpenAI(id="gpt-4o"), markdown=True) [V0_FILE]markdown:file="agno/advanced/context_in_instructions.md" --- title: Context In Instructions category: advanced source_lines: 11618-11641 line_count: 23 --- # Context In Instructions Source: https://docs.agno.com/examples/concepts/context/03-context_in_instructions ## Code \`\`\`python cookbook/agent_concepts/context/03-context_in_instructions.py import json from textwrap import dedent import httpx from agno.agent import Agent from agno.models.openai import OpenAIChat def get_upcoming_spacex_launches(num_launches: int = 5) -> str: url = "https://api.spacexdata.com/v5/launches/upcoming" launches = httpx.get(url).json() launches = sorted(launches, key=lambda x: x["date_unix"])[:num_launches] return json.dumps(launches, indent=4) [V0_FILE]markdown:file="agno/advanced/create_a_context-aware_agent_that_can_access_real-time_hackernews_data.md" --- title: Create a Context-Aware Agent that can access real-time HackerNews data category: advanced source_lines: 33889-33907 line_count: 18 --- # Create a Context-Aware Agent that can access real-time HackerNews data agent = Agent( model=OpenAIChat(id="gpt-4o"), # Each function in the context is evaluated when the agent is run, # think of it as dependency injection for Agents context={"top_hackernews_stories": get_top_hackernews_stories}, # add_context will automatically add the context to the user message # add_context=True, # Alternatively, you can manually add the context to the instructions instructions=dedent("""\ You are an insightful tech trend observer! 📰 Here are the top stories on HackerNews: {top_hackernews_stories}\ """), markdown=True, ) [V0_FILE]markdown:file="agno/advanced/create_a_fresh_agent_for_streaming.md" --- title: Create a fresh agent for streaming category: advanced source_lines: 21712-21719 line_count: 7 --- # Create a fresh agent for streaming streaming_agent = Agent( model=OpenAIChat(id="gpt-4o"), reasoning=True, markdown=True, ) [V0_FILE]markdown:file="agno/advanced/create_a_fresh_agent_with_reasoning_model_for_streaming.md" --- title: Create a fresh agent with reasoning_model for streaming category: advanced source_lines: 21742-21749 line_count: 7 --- # Create a fresh agent with reasoning_model for streaming streaming_agent_with_model = Agent( model=OpenAIChat(id="gpt-4o"), reasoning_model=OpenAIChat(id="gpt-4o"), markdown=True, ) [V0_FILE]markdown:file="agno/advanced/create_streaming_team.md" --- title: Create streaming team category: advanced source_lines: 70867-70878 line_count: 11 --- # Create streaming team market_research_team = Team( name="Market Research Team", mode="coordinate", model=OpenAIChat("gpt-4o"), members=[trend_analyst, risk_assessor], response_model=MarketAnalysis, markdown=True, show_members_responses=True, ) [V0_FILE]markdown:file="agno/advanced/example_prompts.md" --- title: Example prompts: category: advanced source_lines: 27587-27614 line_count: 27 --- # Example prompts: """ Customer Management: - "Create a customer. Name: ACME Corp, Email: billing@acme.example.com" - "List my customers." - "Find customer by email 'jane.doe@example.com'" # Note: Requires 'customers.retrieve' or search capability Product and Price Management: - "Create a new product called 'Basic Plan'." - "Create a recurring monthly price of $10 USD for product 'Basic Plan'." - "Create a product 'Ebook Download' and a one-time price of $19.95 USD." - "List all products." # Note: Requires 'products.list' capability - "List all prices." # Note: Requires 'prices.list' capability Payment Links: - "Create a payment link for the $10 USD monthly 'Basic Plan' price." - "Generate a payment link for the '$19.95 Ebook Download'." Combined Tasks: - "Create a product 'Pro Service', add a price $150 USD (one-time), and give me the payment link." - "Register a new customer 'support@example.com' named 'Support Team'." """ \`\`\` [V0_FILE]markdown:file="agno/advanced/example_prompts_to_explore.md" --- title: Example prompts to explore: category: advanced source_lines: 9125-9183 line_count: 58 --- # Example prompts to explore: """ Issue queries: 1. "Find issues needing attention" 2. "Show me issues by label" 3. "What issues are being actively discussed?" 4. "Find related issues" 5. "Analyze issue resolution patterns" Pull request queries: 1. "What PRs need review?" 2. "Show me recent merged PRs" 3. "Find PRs with conflicts" 4. "What features are being developed?" 5. "Analyze PR review patterns" Repository queries: 1. "Show repository health metrics" 2. "What are the contribution guidelines?" 3. "Find documentation gaps" 4. "Analyze code quality trends" 5. "Show repository activity patterns" """ \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx export GITHUB_TOKEN=xxx \`\`\` \`\`\`bash pip install -U agno "uvicorn[standard]" openai nest-asyncio mcp \`\`\` \`\`\`bash Mac python cookbook/apps/playground/mcp_demo.py \`\`\` \`\`\`bash Windows python cookbook/apps/playground/mcp_demo.py \`\`\` [V0_FILE]markdown:file="agno/advanced/example_prompts_to_try.md" --- title: Example prompts to try: category: advanced source_lines: 27696-27709 line_count: 13 --- # Example prompts to try: """ A SaaS tool that helps businesses automate document processing using AI. Users can upload invoices, contracts, or PDFs and get structured data, smart summaries, and red flag alerts for compliance or anomalies. Ideal for legal teams, accountants, and enterprise back offices. An AI-enhanced SaaS platform for streamlining the recruitment process. Features include automated candidate screening using NLP, AI interview scheduling, bias detection in job descriptions, and pipeline analytics. Designed for fast-growing startups and mid-sized HR teams. An internal SaaS tool for HR departments to monitor employee wellbeing. Combines weekly mood check-ins, anonymous feedback, and AI-driven burnout detection models. Integrates with Slack and HR systems to support a healthier workplace culture. """ \`\`\` [V0_FILE]markdown:file="agno/advanced/live_search_streaming_agent.md" --- title: Live Search Streaming Agent category: advanced source_lines: 49375-49427 line_count: 52 --- # Live Search Streaming Agent Source: https://docs.agno.com/examples/models/xai/live_search_agent_stream ## Code \`\`\`python cookbook/models/xai/live_search_agent_stream.py from agno.agent import Agent from agno.models.xai import xAI agent = Agent( model=xAI( id="grok-3", search_parameters={ "mode": "on", "max_search_results": 20, "return_citations": True, }, ), markdown=True, ) agent.print_response( "Provide me a digest of world news in the last 24 hours.", stream=True ) \`\`\` ## Usage \`\`\`bash export XAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno \`\`\` \`\`\`bash python cookbook/models/xai/live_search_agent_stream.py \`\`\` [V0_FILE]markdown:file="agno/advanced/model_context_protocol_mcp.md" --- title: Model Context Protocol (MCP) category: advanced source_lines: 71881-71992 line_count: 111 --- # Model Context Protocol (MCP) Source: https://docs.agno.com/tools/mcp/mcp Learn how to use MCP with Agno to enable your agents to interact with external systems through a standardized interface. The [Model Context Protocol (MCP)](https://modelcontextprotocol.io) enables Agents to interact with external systems through a standardized interface. You can connect your Agents to any MCP server, using Agno's MCP integration. ## Usage You can use any working MCP server. To see some examples, you can check [this GitHub repository](https://github.com/modelcontextprotocol/servers), by the maintainers of the MCP themselves. Intialize the `MCPTools` class and connect to the MCP server. This needs to be done inside an async function. The recommended way to define the MCP server, is to use the `command` or `url` parameters. With `command`, you can pass the command used to run the MCP server you want. With `url`, you can pass the URL of the running MCP server you want to use. For example, to use the "[mcp-server-git](https://github.com/modelcontextprotocol/servers/tree/main/src/git)" server, you can do the following: \`\`\`python from agno.tools.mcp import MCPTools async def run_mcp_agent(): # Initialize the MCP tools mcp_tools = MCPTools(command=f"uvx mcp-server-git") # Connect to the MCP server await mcp_tools.connect() ... \`\`\` When initializing the Agent, pass the `MCPTools` class in the `tools` parameter. The agent will now be ready to use the MCP server: \`\`\`python from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.mcp import MCPTools async def run_mcp_agent(): # Initialize the MCP tools mcp_tools = MCPTools(command=f"uvx mcp-server-git") # Connect to the MCP server await mcp_tools.connect() # Setup and run the agent agent = Agent(model=OpenAIChat(id="gpt-4o"), tools=[mcp_tools]) await agent.aprint_response("What is the license for this project?", stream=True) \`\`\` ### Basic example: Filesystem Agent Here's a filesystem agent that uses the [Filesystem MCP server](https://github.com/modelcontextprotocol/servers/tree/main/src/filesystem) to explore and analyze files: \`\`\`python filesystem_agent.py import asyncio from pathlib import Path from textwrap import dedent from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.mcp import MCPTools from mcp import StdioServerParameters async def run_mcp_agent(message: str) -> None: """Run the filesystem agent with the given message.""" file_path = str(Path(__file__).parent.parent.parent.parent) # Initialize the MCP tools mcp_tools = MCPTools(f"npx -y @modelcontextprotocol/server-filesystem {file_path}") # Connect to the MCP server await mcp_tools.connect() # Use the MCP tools with an Agent agent = Agent( model=OpenAIChat(id="gpt-4o"), tools=[mcp_tools], instructions=dedent("""\ You are a filesystem assistant. Help users explore files and directories. - Navigate the filesystem to answer questions - Use the list_allowed_directories tool to find directories that you can access - Provide clear context about files you examine - Use headings to organize your responses - Be concise and focus on relevant information\ """), markdown=True, show_tool_calls=True, ) # Run the agent await agent.aprint_response(message, stream=True) # Close the MCP connection await mcp_tools.close() [V0_FILE]markdown:file="agno/advanced/more_example_prompts_to_explore.md" --- title: More example prompts to explore: category: advanced source_lines: 27038-27064 line_count: 26 --- # More example prompts to explore: """ Issue queries: 1. "Find issues needing attention" 2. "Show me issues by label" 3. "What issues are being actively discussed?" 4. "Find related issues" 5. "Analyze issue resolution patterns" Pull request queries: 1. "What PRs need review?" 2. "Show me recent merged PRs" 3. "Find PRs with conflicts" 4. "What features are being developed?" 5. "Analyze PR review patterns" Repository queries: 1. "Show repository health metrics" 2. "What are the contribution guidelines?" 3. "Find documentation gaps" 4. "Analyze code quality trends" 5. "Show repository activity patterns" """ \`\`\` [V0_FILE]markdown:file="agno/advanced/more_example_prompts_to_try.md" --- title: More example prompts to try: category: advanced source_lines: 36530-36566 line_count: 36 --- # More example prompts to try: """ Try these creative prompts: 1. "Create a video of autumn leaves falling in a peaceful forest" 2. "Generate a video of a cat playing with a ball" 3. "Create a video of a peaceful koi pond with rippling water" 4. "Generate a video of a cozy fireplace with dancing flames" 5. "Create a video of a mystical portal opening in a magical realm" """ \`\`\` ## Usage \`\`\`bash pip install openai agno \`\`\` \`\`\`bash export MODELS_LAB_API_KEY=**** \`\`\` \`\`\`bash python video_generation.py \`\`\` [V0_FILE]markdown:file="agno/advanced/print_response_which_includes_processing_streaming_responses.md" --- title: Print response (which includes processing streaming responses) category: advanced source_lines: 21749-21757 line_count: 8 --- # Print response (which includes processing streaming responses) print("Running with reasoning_model specified (streaming)...") streaming_agent_with_model.print_response( "What is the value of 5! (factorial)?", stream=True, show_full_reasoning=True, ) [V0_FILE]markdown:file="agno/advanced/print_the_response_in_the_terminal_streaming.md" --- title: Print the response in the terminal (streaming) category: advanced source_lines: 39866-39901 line_count: 35 --- # Print the response in the terminal (streaming) agent.print_response("write a two sentence horror story", stream=True) \`\`\` ## Usage \`\`\`bash export CEREBRAS_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno \`\`\` \`\`\`bash Mac python cookbook/models/cerebras_openai/basic_stream.py \`\`\` \`\`\`bash Windows python cookbook/models/cerebras_openai/basic_stream.py \`\`\` [V0_FILE]markdown:file="agno/advanced/prompts.md" --- title: Prompts category: advanced source_lines: 1388-1513 line_count: 125 --- # Prompts Source: https://docs.agno.com/agents/prompts We prompt Agents using `description` and `instructions` and a number of other settings. These settings are used to build the **system** message that is sent to the language model. Understanding how these prompts are created will help you build better Agents. The 2 key parameters are: 1. **Description**: A description that guides the overall behaviour of the agent. 2. **Instructions**: A list of precise, task-specific instructions on how to achieve its goal. Description and instructions only provide a formatting benefit, we do not alter or abstract any information and you can always set the `system_message` to provide your own system prompt. ## System message The system message is created using `description`, `instructions` and a number of other settings. The `description` is added to the start of the system message and `instructions` are added as a list after `Instructions`. For example: \`\`\`python instructions.py from agno.agent import Agent agent = Agent( description="You are a famous short story writer asked to write for a magazine", instructions=["You are a pilot on a plane flying from Hawaii to Japan."], markdown=True, debug_mode=True, ) agent.print_response("Tell me a 2 sentence horror story.", stream=True) \`\`\` Will translate to (set `debug_mode=True` to view the logs): \`\`\`js DEBUG ============== system ============== DEBUG You are a famous short story writer asked to write for a magazine ## Instructions - You are a pilot on a plane flying from Hawaii to Japan. - Use markdown to format your answers. DEBUG ============== user ============== DEBUG Tell me a 2 sentence horror story. DEBUG ============== assistant ============== DEBUG As the autopilot disengaged inexplicably mid-flight over the Pacific, the pilot glanced at the copilot's seat only to find it empty despite his every recall of a full crew boarding. Hands trembling, he looked into the cockpit's rearview mirror and found his own reflection grinning back with blood-red eyes, whispering, "There's no escape, not at 30,000 feet." DEBUG **************** METRICS START **************** DEBUG * Time to first token: 0.4518s DEBUG * Time to generate response: 1.2594s DEBUG * Tokens per second: 63.5243 tokens/s DEBUG * Input tokens: 59 DEBUG * Output tokens: 80 DEBUG * Total tokens: 139 DEBUG * Prompt tokens details: {'cached_tokens': 0} DEBUG * Completion tokens details: {'reasoning_tokens': 0} DEBUG **************** METRICS END ****************** \`\`\` ## Set the system message directly You can manually set the system message using the `system_message` parameter. \`\`\`python from agno.agent import Agent agent = Agent(system_message="Share a 2 sentence story about") agent.print_response("Love in the year 12000.") \`\`\` Some models via some model providers, like `llama-3.2-11b-vision-preview` on Groq, require no system message with other messages. To remove the system message, set `create_default_system_message=False` and `system_message=None`. Additionally, if `markdown=True` is set, it will add a system message, so either remove it or explicitly disable the system message. ## User message The input `message` sent to the `Agent.run()` or `Agent.print_response()` functions is used as the user message. ## Default system message The Agent creates a default system message that can be customized using the following parameters: | Parameter | Type | Default | Description | | ------------------------------- | ----------- | -------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `description` | `str` | `None` | A description of the Agent that is added to the start of the system message. | | `goal` | `str` | `None` | Describe the task the agent should achieve. | | `instructions` | `List[str]` | `None` | List of instructions added to the system prompt in `` tags. Default instructions are also created depending on values for `markdown`, `output_model` etc. | | `additional_context` | `str` | `None` | Additional context added to the end of the system message. | | `expected_output` | `str` | `None` | Provide the expected output from the Agent. This is added to the end of the system message. | | `markdown` | `bool` | `False` | Add an instruction to format the output using markdown. | | `add_datetime_to_instructions` | `bool` | `False` | If True, add the current datetime to the prompt to give the agent a sense of time. This allows for relative times like "tomorrow" to be used in the prompt | | `system_message` | `str` | `None` | System prompt: provide the system prompt as a string | | `system_message_role` | `str` | `system` | Role for the system message. | | `create_default_system_message` | `bool` | `True` | If True, build a default system prompt using agent settings and use that. | Disable the default system message by setting `create_default_system_message=False`. ## Default user message The Agent creates a default user message, which is either the input message or a message with the `context` if `enable_rag=True`. The default user message can be customized using: | Parameter | Type | Default | Description | | ----------------------------- | ------------------------- | -------- | ---------------------------------------------------------------------------------------------------------------------------- | | `context` | `str` | `None` | Additional context added to the end of the user message. | | `add_context` | `bool` | `False` | If True, add the context to the user prompt. | | `resolve_context` | `bool` | `True` | If True, resolve the context (i.e. call any functions in the context) before adding it to the user prompt. | | `add_references` | `bool` | `False` | Enable RAG by adding references from the knowledge base to the prompt. | | `retriever` | `Callable` | `None` | Function to get references to add to the user\_message. This function, if provided, is called when `add_references` is True. | | `references_format` | `Literal["json", "yaml"]` | `"json"` | Format of the references. | | `add_history_to_messages` | `bool` | `False` | If true, adds the chat history to the messages sent to the Model. | | `num_history_responses` | `int` | `3` | Number of historical responses to add to the messages. | | `user_message` | `Union[List, Dict, str]` | `None` | Provide the user prompt as a string. Note: this will ignore the message sent to the run function. | | `user_message_role` | `str` | `user` | Role for the user message. | | `create_default_user_message` | `bool` | `True` | If True, build a default user prompt using references and chat history. | Disable the default user message by setting `create_default_user_message=False`. [V0_FILE]markdown:file="agno/advanced/question_with_streaming.md" --- title: Question with streaming category: advanced source_lines: 18585-18640 line_count: 55 --- # Question with streaming output_stream: Iterator[RunResponse] = agent.run( "Is a golden retriever a good family dog?", stream=True ) with wave.open("tmp/answer_1.wav", "wb") as wav_file: wav_file.setnchannels(CHANNELS) wav_file.setsampwidth(SAMPLE_WIDTH) wav_file.setframerate(SAMPLE_RATE) for response in output_stream: if response.response_audio: if response.response_audio.transcript: print(response.response_audio.transcript, end="", flush=True) if response.response_audio.content: try: pcm_bytes = base64.b64decode(response.response_audio.content) wav_file.writeframes(pcm_bytes) except Exception as e: print(f"Error decoding audio: {e}") print() \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/multimodal/audio_streaming.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/multimodal/audio_streaming.py \`\`\` [V0_FILE]markdown:file="agno/advanced/refresh_the_context.md" --- title: Refresh the context category: advanced source_lines: 73629-73632 line_count: 3 --- # Refresh the context agent.context["memory"] = zep_tools.get_zep_memory(memory_type="context") [V0_FILE]markdown:file="agno/advanced/run_the_agent_non-streaming.md" --- title: Run the agent (non-streaming) category: advanced source_lines: 21696-21702 line_count: 6 --- # Run the agent (non-streaming) print("Running with reasoning_model specified (non-streaming)...") response = agent_with_reasoning_model.run( "What is the sum of the first 10 natural numbers?" ) [V0_FILE]markdown:file="agno/advanced/streaming_agent.md" --- title: Streaming Agent category: advanced source_lines: 48985-48999 line_count: 14 --- # Streaming Agent Source: https://docs.agno.com/examples/models/xai/basic_stream ## Code \`\`\`python cookbook/models/xai/basic_stream.py from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.xai import xAI agent = Agent(model=xAI(id="grok-3"), markdown=True) [V0_FILE]markdown:file="agno/advanced/streaming_basic_agent.md" --- title: Streaming Basic Agent category: advanced source_lines: 42969-42983 line_count: 14 --- # Streaming Basic Agent Source: https://docs.agno.com/examples/models/ibm/basic_stream ## Code \`\`\`python cookbook/models/ibm/watsonx/basic_stream.py from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.ibm import WatsonX agent = Agent(model=WatsonX(id="ibm/granite-20b-code-instruct"), markdown=True) [V0_FILE]markdown:file="agno/advanced/streaming_external_tool_execution.md" --- title: Streaming External Tool Execution category: advanced source_lines: 32397-32412 line_count: 15 --- # Streaming External Tool Execution Source: https://docs.agno.com/examples/concepts/user-control-flows/10-external-tool-execution-stream This example demonstrates how to implement streaming external tool execution, allowing for real-time interaction and response streaming while executing tools outside of the agent's control. ## Code \`\`\`python cookbook/agent_concepts/user_control_flows/external_tool_execution_stream.py import subprocess from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools import tool from agno.utils import pprint [V0_FILE]markdown:file="agno/advanced/streaming_responses.md" --- title: Streaming responses category: advanced source_lines: 70368-70372 line_count: 4 --- # Streaming responses for chunk in team.run("What is the weather in Tokyo?", stream=True): print(chunk.content, end="", flush=True) [V0_FILE]markdown:file="agno/advanced/streaming_user_confirmation.md" --- title: Streaming User Confirmation category: advanced source_lines: 31644-31768 line_count: 124 --- # Streaming User Confirmation Source: https://docs.agno.com/examples/concepts/user-control-flows/03-confirmation-required-stream This example demonstrates how to implement streaming user confirmation flows, allowing for real-time interaction and response streaming. ## Code \`\`\`python cookbook/agent_concepts/user_control_flows/confirmation_required_stream.py import json import httpx from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools import tool from agno.utils import pprint from rich.console import Console from rich.prompt import Prompt console = Console() @tool(requires_confirmation=True) def get_top_hackernews_stories(num_stories: int) -> str: """Fetch top stories from Hacker News. Args: num_stories (int): Number of stories to retrieve Returns: str: JSON string containing story details """ # Fetch top story IDs response = httpx.get("https://hacker-news.firebaseio.com/v0/topstories.json") story_ids = response.json() # Yield story details all_stories = [] for story_id in story_ids[:num_stories]: story_response = httpx.get( f"https://hacker-news.firebaseio.com/v0/item/{story_id}.json" ) story = story_response.json() if "text" in story: story.pop("text", None) all_stories.append(story) return json.dumps(all_stories) agent = Agent( model=OpenAIChat(id="gpt-4o-mini"), tools=[get_top_hackernews_stories], markdown=True, ) for run_response in agent.run("Fetch the top 2 hackernews stories", stream=True): if run_response.is_paused: for tool in run_response.tools_requiring_confirmation: # Ask for confirmation console.print( f"Tool name [bold blue]{tool.tool_name}({tool.tool_args})[/] requires confirmation." ) message = ( Prompt.ask("Do you want to continue?", choices=["y", "n"], default="y") .strip() .lower() ) if message == "n": tool.confirmed = False else: # We update the tools in place tool.confirmed = True run_response = agent.continue_run( run_id=agent.run_response.run_id, updated_tools=agent.run_response.tools, stream=True, ) pprint.pprint_run_response(run_response) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U agno httpx rich openai \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/user_control_flows/confirmation_required_stream.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/user_control_flows/confirmation_required_stream.py \`\`\` ## Key Features * Uses `agent.run(stream=True)` for streaming responses * Implements streaming continuation with `agent.continue_run(stream=True)` * Maintains real-time interaction with user confirmation * Demonstrates how to handle streaming responses with user input ## Use Cases * Real-time user interaction * Streaming applications requiring user input * Interactive chat interfaces * Progressive response generation [V0_FILE]markdown:file="agno/advanced/streaming_user_input.md" --- title: Streaming User Input category: advanced source_lines: 31974-32078 line_count: 104 --- # Streaming User Input Source: https://docs.agno.com/examples/concepts/user-control-flows/06-user-input-required-stream This example demonstrates how to implement streaming user input collection, allowing for real-time interaction and response streaming while gathering user information. ## Code \`\`\`python cookbook/agent_concepts/user_control_flows/user_input_required_stream.py from typing import List from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools import tool from agno.tools.function import UserInputField from agno.utils import pprint @tool(requires_user_input=True, user_input_fields=["to_address"]) def send_email(subject: str, body: str, to_address: str) -> str: """ Send an email. Args: subject (str): The subject of the email. body (str): The body of the email. to_address (str): The address to send the email to. """ return f"Sent email to {to_address} with subject {subject} and body {body}" agent = Agent( model=OpenAIChat(id="gpt-4o-mini"), tools=[send_email], markdown=True, ) for run_response in agent.run( "Send an email with the subject 'Hello' and the body 'Hello, world!'", stream=True ): if run_response.is_paused: for tool in run_response.tools_requiring_user_input: input_schema: List[UserInputField] = tool.user_input_schema for field in input_schema: # Display field information to the user print(f"\nField: {field.name}") print(f"Description: {field.description}") print(f"Type: {field.field_type}") # Get user input if field.value is None: user_value = input(f"Please enter a value for {field.name}: ") # Update the field value field.value = user_value else: print(f"Value: {field.value}") run_response = agent.continue_run(stream=True) pprint.pprint_run_response(run_response) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U agno openai \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/user_control_flows/user_input_required_stream.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/user_control_flows/user_input_required_stream.py \`\`\` ## Key Features * Uses `agent.run(stream=True)` for streaming responses * Implements streaming continuation with `agent.continue_run(stream=True)` * Maintains real-time interaction with user input collection * Demonstrates how to handle streaming responses with user input ## Use Cases * Real-time user interaction * Streaming applications requiring user input * Interactive form-like interfaces * Progressive response generation with user input [V0_FILE]markdown:file="agno/core/agent_api.md" --- title: Agent API category: core source_lines: 1-27 line_count: 26 --- # Agent API Source: https://docs.agno.com/agent-api/introduction A robust, production-ready application for serving Agents as an API. Welcome to the Simple Agent API: a robust, production-ready application for serving Agents as an API. It includes: * A FastAPI server for handling API requests. * A PostgreSQL database for storing Agent sessions, knowledge, and memories. * A set of pre-built Agents to use as a starting point. ## Additional Information Congratulations on running your Agent API. * Read how to [use workspaces with your Agent API](/workspaces/introduction) [V0_FILE]markdown:file="agno/core/agent_api_fastapi_and_postgres.md" --- title: Agent API: FastAPI and Postgres category: core source_lines: 85246-85277 line_count: 31 --- # Agent API: FastAPI and Postgres Source: https://docs.agno.com/workspaces/agent-api/local The Agent API workspace provides a simple RestAPI + database for serving agents. It contains: * A FastAPI server for serving Agents, Teams and Workflows. * A postgres database for session and vector storage. ## Next Congratulations on running your Agent API locally. Next Steps: * [Run your Agent API on AWS](/workspaces/agent-api/aws) * Read how to [update workspace settings](/workspaces/workspace-management/workspace-settings) * Read how to [create a git repository for your workspace](/workspaces/workspace-management/git-repo) * Read how to [manage the development application](/workspaces/workspace-management/development-app) * Read how to [format and validate your code](/workspaces/workspace-management/format-and-validate) * Read how to [add python libraries](/workspaces/workspace-management/install) * Chat with us on [discord](https://agno.link/discord) [V0_FILE]markdown:file="agno/core/running_the_agent_api_on_aws.md" --- title: Running the Agent API on AWS category: core source_lines: 85213-85246 line_count: 33 --- # Running the Agent API on AWS Source: https://docs.agno.com/workspaces/agent-api/aws Let's run the **Agent API** in production on AWS. ## Next Congratulations on running your Agent API on AWS. Next Steps: * Read how to [update workspace settings](/workspaces/workspace-management/workspace-settings) * Read how to [create a git repository for your workspace](/workspaces/workspace-management/git-repo) * Read how to [manage the production application](/workspaces/workspace-management/production-app) * Read how to [format and validate your code](/workspaces/workspace-management/format-and-validate) * Read how to [add python libraries](/workspaces/workspace-management/install) * Read how to [add a custom domain and HTTPS](/workspaces/workspace-management/domain-https) * Read how to [implement CI/CD](/workspaces/workspace-management/ci-cd) * Chat with us on [discord](https://agno.link/discord) [V0_FILE]markdown:file="agno/core/running_your_agent.md" --- title: Running your Agent category: core source_lines: 1513-1534 line_count: 21 --- # Running your Agent Source: https://docs.agno.com/agents/run Learn how to run an agent and get the response. The `Agent.run()` function runs the agent and generates a response, either as a `RunResponse` object or a stream of `RunResponse` objects. Many of our examples use `agent.print_response()` which is a helper utility to print the response in the terminal. It uses `agent.run()` under the hood. ## Running your Agent Here's how to run your agent. The response is captured in the `response`. \`\`\`python from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.openai import OpenAIChat from agno.utils.pprint import pprint_run_response agent = Agent(model=OpenAIChat(id="gpt-4o-mini")) [V0_FILE]markdown:file="agno/core/sessions.md" --- title: Sessions category: core source_lines: 1723-1751 line_count: 28 --- # Sessions Source: https://docs.agno.com/agents/sessions When we call `Agent.run()`, it creates a stateless, singular Agent run. But what if we want to continue this run i.e. have a multi-turn conversation? That's where `sessions` come in. A session is collection of consecutive runs. In practice, a session is a multi-turn conversation between a user and an Agent. Using a `session_id`, we can connect the conversation history and state across multiple runs. Let's outline some key concepts: * **User:** A user represents an individual that interacts with the Agent. Each user has associated memories, sessions, and conversation history separate from other users. * **Session:** A session is collection of consecutive runs like a multi-turn conversation between a user and an Agent. Sessions are identified by a `session_id` and each turn is a **run**. * **Run:** Every interaction (i.e. chat or turn) with an Agent is called a **run**. Runs are identified by a `run_id` and `Agent.run()` creates a new `run_id` when called. * **Messages:** are the individual messages sent between the model and the Agent. Messages are the communication protocol between the Agent and model. Let's start with an example where a single run is created with an Agent. A `run_id` is automatically generated, as well as a `session_id` (because we didn't provide one to continue the conversation). This run is not yet associated with a user. \`\`\`python from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.openai import OpenAIChat from agno.utils.pprint import pprint_run_response agent = Agent(model=OpenAIChat(id="gpt-4o-mini")) [V0_FILE]markdown:file="agno/core/store_agent_sessions_in_a_sqlite_database.md" --- title: Store agent sessions in a SQLite database category: core source_lines: 60051-60168 line_count: 117 --- # Store agent sessions in a SQLite database storage = SqliteStorage(table_name="agent_sessions", db_file="tmp/agent.db") agent = Agent( name="Agno Assist", model=Claude(id="claude-sonnet-4-20250514"), instructions=[ "Search your knowledge before answering the question.", "Only include the output in your response. No other text.", ], knowledge=knowledge, storage=storage, add_datetime_to_instructions=True, # Add the chat history to the messages add_history_to_messages=True, # Number of history runs num_history_runs=3, markdown=True, ) if __name__ == "__main__": # Load the knowledge base, comment out after first run # Set recreate to True to recreate the knowledge base if needed agent.knowledge.load(recreate=False) agent.print_response("What is Agno?", stream=True) \`\`\` Install dependencies, export your `OPENAI_API_KEY` and run the Agent \`\`\`bash Mac uv pip install -U lancedb tantivy openai sqlalchemy \`\`\` \`\`\`bash Windows uv pip install -U lancedb tantivy openai sqlalchemy \`\`\` \`\`\`shell python level_2_agent.py \`\`\` ## Level 3: Agents with memory and reasoning * **Reasoning:** enables Agents to **"think" & "analyze"**, improving reliability and quality. `ReasoningTools` is one of the best approaches to improve an Agent's response quality. * **Memory:** enables Agents to classify, store and recall user preferences, personalizing their responses. Memory helps the Agent build personas and learn from previous interactions. \`\`\`python level_3_agent.py from agno.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory from agno.models.anthropic import Claude from agno.tools.reasoning import ReasoningTools from agno.tools.yfinance import YFinanceTools memory = Memory( # Use any model for creating and managing memories model=Claude(id="claude-sonnet-4-20250514"), # Store memories in a SQLite database db=SqliteMemoryDb(table_name="user_memories", db_file="tmp/agent.db"), # We disable deletion by default, enable it if needed delete_memories=True, clear_memories=True, ) agent = Agent( model=Claude(id="claude-sonnet-4-20250514"), tools=[ ReasoningTools(add_instructions=True), YFinanceTools(stock_price=True, analyst_recommendations=True, company_info=True, company_news=True), ], # User ID for storing memories, `default` if not provided user_id="ava", instructions=[ "Use tables to display data.", "Include sources in your response.", "Only include the report in your response. No other text.", ], memory=memory, # Let the Agent manage its memories enable_agentic_memory=True, markdown=True, ) if __name__ == "__main__": # This will create a memory that "ava's" favorite stocks are NVIDIA and TSLA agent.print_response( "My favorite stocks are NVIDIA and TSLA", stream=True, show_full_reasoning=True, stream_intermediate_steps=True, ) # This will use the memory to answer the question agent.print_response( "Can you compare my favorite stocks?", stream=True, show_full_reasoning=True, stream_intermediate_steps=True, ) \`\`\` Run the Agent \`\`\`shell python level_3_agent.py \`\`\` You can use the `Memory` and `Reasoning` separately, you don't need to use them together. [V0_FILE]markdown:file="agno/core/what_are_agents.md" --- title: What are Agents? category: core source_lines: 59929-60038 line_count: 109 --- # What are Agents? Source: https://docs.agno.com/introduction/agents **Agents are AI programs that operate autonomously.** Traditional software follows a pre-programmed sequence of steps. Agents dynamically determine their course of action using a machine learning **model**, its core components are: * **Model:** controls the flow of execution. It decides whether to reason, act or respond. * **Tools:** enable an Agent to take actions and interact with external systems. * **Instructions:** are how we program the Agent, teaching it how to use tools and respond. Agents also have **memory**, **knowledge**, **storage** and the ability to **reason**: * **Reasoning:** enables Agents to "think" before responding and "analyze" the results of their actions (i.e. tool calls), this improves reliability and quality of responses. * **Knowledge:** is domain-specific information that the Agent can **search at runtime** to make better decisions and provide accurate responses (RAG). Knowledge is stored in a vector database and this **search at runtime** pattern is known as Agentic RAG/Agentic Search. * **Storage:** is used by Agents to save session history and state in a database. Model APIs are stateless and storage enables us to continue conversations from where they left off. This makes Agents stateful, enabling multi-turn, long-term conversations. * **Memory:** gives Agents the ability to store and recall information from previous interactions, allowing them to learn user preferences and personalize their responses. Let's build a few Agents to see how they work. ## Level 1: Agents with tools and instructions The simplest Agent has a model, a tool and instructions. Let's build an Agent that can fetch data using the `yfinance` library, along with instructions to display the results in a table. \`\`\`python level_1_agent.py from agno.agent import Agent from agno.models.anthropic import Claude from agno.tools.yfinance import YFinanceTools agent = Agent( model=Claude(id="claude-sonnet-4-20250514"), tools=[YFinanceTools(stock_price=True)], instructions="Use tables to display data. Don't include any other text.", markdown=True, ) agent.print_response("What is the stock price of Apple?", stream=True) \`\`\` Create a virtual environment, install dependencies, export your API key and run the Agent. \`\`\`bash Mac uv venv --python 3.12 source .venv/bin/activate \`\`\` \`\`\`bash Windows uv venv --python 3.12 .venv/Scripts/activate \`\`\` \`\`\`bash Mac uv pip install -U agno anthropic yfinance \`\`\` \`\`\`bash Windows uv pip install -U agno anthropic yfinance \`\`\` \`\`\`bash Mac export ANTHROPIC_API_KEY=sk-*** \`\`\` \`\`\`bash Windows setx ANTHROPIC_API_KEY sk-*** \`\`\` \`\`\`shell python level_1_agent.py \`\`\` Set `debug_mode=True` or `export AGNO_DEBUG=true` to see the system prompt and user messages. ## Level 2: Agents with knowledge and storage **Knowledge:** While models have a large amount of training data, we almost always need to give them domain-specific information to make better decisions and provide accurate responses (RAG). We store this information in a vector database and let the Agent **search** it at runtime. **Storage:** Model APIs are stateless and `Storage` drivers save chat history and state to a database. When the Agent runs, it reads the chat history and state from the database and add it to the messages list, resuming the conversation and making the Agent stateful. In this example, we'll use: * `UrlKnowledge` to load Agno documentation to LanceDB, using OpenAI for embeddings. * `SqliteStorage` to save the Agent's session history and state in a database. \`\`\`python level_2_agent.py from agno.agent import Agent from agno.embedder.openai import OpenAIEmbedder from agno.knowledge.url import UrlKnowledge from agno.models.anthropic import Claude from agno.storage.sqlite import SqliteStorage from agno.vectordb.lancedb import LanceDb, SearchType [V0_FILE]plaintext:file="agno/examples/__pycache__/openai.cpython-313.pyc" � V�h��SSKrS\l\RR5r\RHr\"\R5 M g)�N�YOUR_API_KEY)�openai�api_key�models�list�data�model�print�id��'d:\AI\AI_agents_agno\examples\openai.py�rs:�� [�[�E� �%�(�(�O�r [V0_FILE]python:file="agno/examples/agent_state.py" from agno.agent import Agent from agno.models.openai import OpenAIChat # Defrom agno.agent import Agent from agno.models.openai import OpenAIChat # load environment variables from dotenv import load_dotenv load_dotenv() import os # Define a tool that adds an item to the shopping list def add_item(agent: Agent, item: str) -> str: """Add an item to the shopping list.""" agent.session_state["shopping_list"].append(item) return f"The shopping list is now {agent.session_state['shopping_list']}" # Create an Agent that maintains state agent = Agent( model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY")), # Initialize the session state with a counter starting at 0 session_state={"shopping_list": []}, tools=[add_item], # You can use variables from the session state in the instructions instructions="Current state (shopping list) is: {shopping_list}", # Important: Add the state to the messages add_state_in_messages=True, markdown=True, ) # Example usage agent.print_response("Add milk, eggs, and bread to the shopping list", stream=True) print(f"Final session state: {agent.session_state}") #define a tool that adds an item to the shopping list def add_item(agent: Agent, item: str) -> str: """Add an item to the shopping list.""" agent.session_state["shopping_list"].append(item) return f"The shopping list is now {agent.session_state['shopping_list']}" # Create an Agent that maintains state agent = Agent( model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY")), # Initialize the session state with a counter starting at 0 session_state={"shopping_list": []}, tools=[add_item], # You can use variables from the session state in the instructions instructions="Current state (shopping list) is: {shopping_list}", # Important: Add the state to the messages add_state_in_messages=True, markdown=True, ) # Example usage agent.print_response("Add milk, eggs, and bread to the shopping list", stream=True) print(f"Final session state: {agent.session_state}") [V0_FILE]plaintext:file="agno/examples/agno_notebook.ipynb" { "cells": [ { "cell_type": "code", "execution_count": 8, "id": "ba618961", "metadata": {}, "outputs": [], "source": [ "from agno.agent import Agent\n", "from agno.models.anthropic import Claude\n", "from agno.tools.reasoning import ReasoningTools\n", "from agno.tools.yfinance import YFinanceTools\n", "\n", "# Claude with reasoning tools for complex analysis\n", "reasoning_agent = Agent(\n", " model=Claude(id=\"claude-3-7-sonnet-latest\"),\n", " tools=[\n", " ReasoningTools(\n", " think=True,\n", " analyze=True,\n", " add_instructions=True,\n", " ),\n", " YFinanceTools(\n", " stock_price=True,\n", " analyst_recommendations=True,\n", " company_info=True,\n", " ),\n", " ],\n", " instructions=\"Use structured reasoning for complex problems\",\n", " stream_intermediate_steps=True,\n", " show_tool_calls=True,\n", " markdown=True,\n", ")" ] }, { "cell_type": "code", "execution_count": 9, "id": "3b2e0e7a", "metadata": {}, "outputs": [], "source": [ "# This question will trigger multiple intermediate steps\n", "response_stream = reasoning_agent.run(\n", " \"What is the stock price of NVDA and should I invest in it? \"\n", " \"Consider market trends, financial health, and risk factors.\",\n", " stream=True,\n", " stream_intermediate_steps=True\n", ")\n" ] }, { "cell_type": "code", "execution_count": 11, "id": "608478ef", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Content: I'll help you analyze\n", "Content: NVDA stock and provide insights\n", "Content: on whether it might be a good investment.\n", "Content: Let me gather the relevant information for\n", "Content: you.\n", "Content: \n", "Tool call started: ToolExecution(tool_call_id='toolu_01C1VbMpQQYMXSs3c1tcdegX', tool_name='think', tool_args={'title': 'Planning my approach to analyze NVDA stock', 'thought': 'To provide a comprehensive analysis of NVDA (NVIDIA) as a potential investment, I need to gather several pieces of information:\\n1. Current stock price of NVDA using get_current_stock_price\\n2. Company information and overview using get_company_info\\n3. Analyst recommendations using get_analyst_recommendations\\n\\nOnce I have this data, I can analyze market trends, financial health, and risk factors to provide an informed perspective on whether NVDA might be a good investment.', 'action': 'Make tool calls to gather the necessary NVDA stock information', 'confidence': 0.95}, tool_call_error=None, result=None, metrics=None, stop_after_tool_call=False, created_at=1755337693, requires_confirmation=None, confirmed=None, confirmation_note=None, requires_user_input=None, user_input_schema=None, answered=None, external_execution_required=None)\n", "Reasoning step: title='Planning my approach to analyze NVDA stock' action='Make tool calls to gather the necessary NVDA stock information' result=None reasoning='To provide a comprehensive analysis of NVDA (NVIDIA) as a potential investment, I need to gather several pieces of information:\\n1. Current stock price of NVDA using get_current_stock_price\\n2. Company information and overview using get_company_info\\n3. Analyst recommendations using get_analyst_recommendations\\n\\nOnce I have this data, I can analyze market trends, financial health, and risk factors to provide an informed perspective on whether NVDA might be a good investment.' next_action= confidence=0.95\n", "Content: \n", "Tool call started: ToolExecution(tool_call_id='toolu_01VKf94t2Bybv1cXsW1TRtqv', tool_name='get_current_stock_price', tool_args={'symbol': 'NVDA'}, tool_call_error=None, result=None, metrics=None, stop_after_tool_call=False, created_at=1755337693, requires_confirmation=None, confirmed=None, confirmation_note=None, requires_user_input=None, user_input_schema=None, answered=None, external_execution_required=None)\n", "Content: \n", "Tool call started: ToolExecution(tool_call_id='toolu_01TasBH6hTQWGb2PsJq9jFWg', tool_name='get_company_info', tool_args={'symbol': 'NVDA'}, tool_call_error=None, result=None, metrics=None, stop_after_tool_call=False, created_at=1755337693, requires_confirmation=None, confirmed=None, confirmation_note=None, requires_user_input=None, user_input_schema=None, answered=None, external_execution_required=None)\n", "Content: \n", "Tool call started: ToolExecution(tool_call_id='toolu_01UgY5ptUbebGYA2Wk4MruTB', tool_name='get_analyst_recommendations', tool_args={'symbol': 'NVDA'}, tool_call_error=None, result=None, metrics=None, stop_after_tool_call=False, created_at=1755337693, requires_confirmation=None, confirmed=None, confirmation_note=None, requires_user_input=None, user_input_schema=None, answered=None, external_execution_required=None)\n", "Content: \n", "Tool call started: ToolExecution(tool_call_id='toolu_01JUZqJvQfXECSVSp6BhXqHg', tool_name='analyze', tool_args={'title': 'Analysis of NVDA stock information', 'result': 'Gathered current stock price ($180.45), company information, and analyst recommendations for NVDA', 'analysis': 'I now have comprehensive data on NVIDIA (NVDA) to evaluate it as a potential investment:\\n\\n1. Current Price and Valuation:\\n- Current price: $180.45\\n- P/E Ratio: 58.21 (relatively high)\\n- Market Cap: $4.4 trillion (extremely large)\\n\\n2. Financial Health:\\n- Strong financial position with $53.69B in cash\\n- Impressive free cash flow of $55.44B\\n- Operating cash flow of $76.16B\\n- EBITDA of $88.25B\\n- Excellent gross margins of 70.1%\\n- EBITDA margins of 59.42%\\n- Revenue growth of 69.2% (exceptional)\\n\\n3. Market Trends:\\n- Trading near its 52-week high of $184.48\\n- Well above both 50-day ($163.50) and 200-day ($136.87) moving averages\\n- Semiconductor industry is booming, especially with AI demand\\n\\n4. Analyst Sentiment:\\n- Overwhelmingly positive with 11 Strong Buy and 47 Buy ratings\\n- Only 6 Hold and 1 Sell ratings\\n- Consistent analyst support over past 3 months\\n\\n5. Risk Factors:\\n- High P/E ratio suggests high expectations built into the price\\n- Trading near all-time highs\\n- Semiconductor industry can be cyclical\\n- Intense competition in AI space\\n- Regulatory risks around AI\\n\\nThis information appears complete and I can now provide a comprehensive analysis to the user.', 'next_action': 'final_answer', 'confidence': 0.9}, tool_call_error=None, result=None, metrics=None, stop_after_tool_call=False, created_at=1755337693, requires_confirmation=None, confirmed=None, confirmation_note=None, requires_user_input=None, user_input_schema=None, answered=None, external_execution_required=None)\n", "Reasoning step: title='Analysis of NVDA stock information' action=None result='Gathered current stock price ($180.45), company information, and analyst recommendations for NVDA' reasoning='I now have comprehensive data on NVIDIA (NVDA) to evaluate it as a potential investment:\\n\\n1. Current Price and Valuation:\\n- Current price: $180.45\\n- P/E Ratio: 58.21 (relatively high)\\n- Market Cap: $4.4 trillion (extremely large)\\n\\n2. Financial Health:\\n- Strong financial position with $53.69B in cash\\n- Impressive free cash flow of $55.44B\\n- Operating cash flow of $76.16B\\n- EBITDA of $88.25B\\n- Excellent gross margins of 70.1%\\n- EBITDA margins of 59.42%\\n- Revenue growth of 69.2% (exceptional)\\n\\n3. Market Trends:\\n- Trading near its 52-week high of $184.48\\n- Well above both 50-day ($163.50) and 200-day ($136.87) moving averages\\n- Semiconductor industry is booming, especially with AI demand\\n\\n4. Analyst Sentiment:\\n- Overwhelmingly positive with 11 Strong Buy and 47 Buy ratings\\n- Only 6 Hold and 1 Sell ratings\\n- Consistent analyst support over past 3 months\\n\\n5. Risk Factors:\\n- High P/E ratio suggests high expectations built into the price\\n- Trading near all-time highs\\n- Semiconductor industry can be cyclical\\n- Intense competition in AI space\\n- Regulatory risks around AI\\n\\nThis information appears complete and I can now provide a comprehensive analysis to the user.' next_action= confidence=0.9\n", "Content: #\n", "Content: NVIDIA (NVDA\n", "Content: ) Stock Analysis\n", "\n", "## Current\n", "Content: Stock Information\n", "- **Current Price**: $180\n", "Content: .45\n", "- **52\n", "Content: -Week Range**: $86\n", "Content: .62 - $184.\n", "Content: 48\n", "- **Market Cap**: $4.\n", "Content: 4 trillion\n", "\n", "## Investment\n", "Content: Analysis\n", "\n", "### Strengths\n", "- **\n", "Content: Exceptional Financial Performance**: \n", "Content: \n", " - 69.2% revenue\n", "Content: growth\n", " - $53\n", "Content: .69B in cash reserves\n", "Content: \n", " - Strong free cash flow of $55\n", "Content: .44B\n", " -\n", "Content: Excellent gross margins of \n", "Content: 70.1%\n", "\n", "- **Market Position**\n", "Content: :\n", " - Leading position\n", "Content: in AI and graphics processing\n", " -\n", "Content: Dominant player in data center GPU market\n", " \n", "Content: - Well-positioned in emerging\n", "Content: technologies (autonomous vehicles, robotics)\n", "Content: \n", "\n", "- **Analyst Confidence\n", "Content: **:\n", " -\n", "Content: Overwhelming analyst support with 58\n", "Content: positive recommendations (11 Strong Buy,\n", "Content: 47 Buy)\n", " - Only \n", "Content: 7 neutral or negative ratings (6 \n", "Content: Hold, 1 \n", "Content: Sell)\n", "\n", "### Risk\n", "Content: Factors\n", "- **Valuation\n", "Content: Concerns**:\n", " \n", "Content: - P/E ratio of \n", "Content: 58.21 is relatively high\n", " -\n", "Content: Trading near its 52-week high\n", "Content: \n", " - High expectations built into the current\n", "Content: price\n", "\n", "- **Industry Risks**:\n", "Content: \n", " - Semiconductor industry can be cyclical\n", "Content: \n", " - Intense competition in the\n", "Content: AI chip space\n", " - Potential\n", "Content: regulatory challenges related to AI technology\n", "Content: \n", "\n", "## Investment Recommendation\n", "Content: \n", "\n", "NVIDIA presents a compelling investment case\n", "Content: based on its financial strength, market position,\n", "Content: and analyst sentiment. The\n", "Content: company is at the forefront of AI\n", "Content: and computational technologies with exceptional\n", "Content: growth and profitability metrics\n", "Content: .\n", "\n", "However, the current\n", "Content: valuation requires careful consideration. With\n", "Content: a high P/E ratio an\n", "Content: d trading near all-time highs, much\n", "Content: of the anticipated growth may already be priced in\n", "Content: .\n", "\n", "**Recommendation approach\n", "Content: **:\n", "- **\n", "Content: Long-term investors**: Consider establishing\n", "Content: or maintaining a position,\n", "Content: potentially through dollar-cost averaging\n", "-\n", "Content: **Short-term investors**: Exercise\n", "Content: caution due to val\n", "Content: uation and potential market volatility\n", "Content: \n", "- **Risk-averse investors**:\n", "Content: Consider waiting for a more attractive entry point or\n", "Content: taking a smaller position\n", "Content: \n", "\n", "As with any investment, consider how\n", "Content: NVIDIA fits within your overall portfolio\n", "Content: diversification, investment timeline, and risk tolerance\n", "Content: . Consulting with a financial advisor for\n", "Content: personalized advice is recommended.\n", "Content: \n" ] } ], "source": [ "\n", "for event in response_stream:\n", " if event.event == \"RunResponseContent\":\n", " print(f\"Content: {event.content}\")\n", " elif event.event == \"ToolCallStarted\":\n", " print(f\"Tool call started: {event.tool}\")\n", " elif event.event == \"ReasoningStep\":\n", " print(f\"Reasoning step: {event.content}\")\n", " ..." ] }, { "cell_type": "code", "execution_count": null, "id": "92b1bdf8", "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "uv", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.13.0" } }, "nbformat": 4, "nbformat_minor": 5 } [V0_FILE]python:file="agno/examples/changing_state_on_run.py" from agno.agent import Agent from agno.models.openai import OpenAIChat # load environment variables from dotenv import load_dotenv load_dotenv() import os agent = Agent( model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY")), add_state_in_messages=True, instructions="Users name is {user_name} and age is {age}", ) # Sets the session state for the session with the id "user_1_session_1" agent.print_response("What is my name?", session_id="user_1_session_1", user_id="user_1", session_state={"user_name": "John", "age": 30}) # Will load the session state from the session with the id "user_1_session_1" agent.print_response("How old am I?", session_id="user_1_session_1", user_id="user_1") # Sets the session state for the session with the id "user_2_session_1" agent.print_response("What is my name?", session_id="user_2_session_1", user_id="user_2", session_state={"user_name": "Jane", "age": 25}) # Will load the session state from the session with the id "user_2_session_1" agent.print_response("How old am I?", session_id="user_2_session_1", user_id="user_2") [V0_FILE]python:file="agno/examples/deep_research_agent.py" import os from agno.agent import Agent from agno.models.nebius import Nebius from dotenv import load_dotenv from typing import Iterator from agno.utils.log import logger from agno.utils.pprint import pprint_run_response from agno.tools.scrapegraph import ScrapeGraphTools from agno.workflow import RunEvent, RunResponse, Workflow from pydantic import BaseModel, Field # load environment variables import os from dotenv import load_dotenv load_dotenv() # get api keys openai_api_key = os.getenv("OPENAI_API_KEY") exa_api_key = os.getenv("EXA_API_KEY") nebius_api_key = os.getenv("NEBIUS_API_KEY") class DeepResearcherAgent(Workflow): """ A multi-stage research workflow that: 1. Gathers information from the web using advanced scraping tools. 2. Analyzes and synthesizes the findings. 3. Produces a clear, well-structured report. """ # Searcher: Finds and extracts relevant information from the web searcher: Agent = Agent( tools=[ScrapeGraphTools(api_key=os.getenv("SGAI_API_KEY"))], model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), show_tool_calls=True, markdown=True, description=( "You are ResearchBot-X, an expert at finding and extracting high-quality, " "up-to-date information from the web. Your job is to gather comprehensive, " "reliable, and diverse sources on the given topic." ), instructions=( "1. Search for the most recent and authoritative and up-to-date sources (news, blogs, official docs, research papers, forums, etc.) on the topic.\n" "2. Extract key facts, statistics, and expert opinions.\n" "3. Cover multiple perspectives and highlight any disagreements or controversies.\n" "4. Include relevant statistics, data, and expert opinions where possible.\n" "5. Organize your findings in a clear, structured format (e.g., markdown table or sections by source type).\n" "6. If the topic is ambiguous, clarify with the user before proceeding.\n" "7. Be as comprehensive and verbose as possible—err on the side of including more detail.\n" "8. Mention the References & Sources of the Content. (It's Must)" ), ) # Analyst: Synthesizes and interprets the research findings analyst: Agent = Agent( model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), markdown=True, description=( "You are AnalystBot-X, a critical thinker who synthesizes research findings " "into actionable insights. Your job is to analyze, compare, and interpret the " "information provided by the researcher." ), instructions=( "1. Identify key themes, trends, and contradictions in the research.\n" "2. Highlight the most important findings and their implications.\n" "3. Suggest areas for further investigation if gaps are found.\n" "4. Present your analysis in a structured, easy-to-read format.\n" "5. Extract and list ONLY the reference links or sources that were ACTUALLY found and provided by the researcher in their findings. Do NOT create, invent, or hallucinate any links.\n" "6. If no links were provided by the researcher, do not include a References section.\n" "7. Don't add hallucinations or make up information. Use ONLY the links that were explicitly passed to you by the researcher.\n" "8. Verify that each link you include was actually present in the researcher's findings before listing it.\n" "9. If there's no Link found from the previous agent then just say, No reference Found." ), ) # Writer: Produces a final, polished report writer: Agent = Agent( model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), markdown=True, description=( "You are WriterBot-X, a professional technical writer. Your job is to craft " "a clear, engaging, and well-structured report based on the analyst's summary." ), instructions=( "1. Write an engaging introduction that sets the context.\n" "2. Organize the main findings into logical sections with headings.\n" "3. Use bullet points, tables, or lists for clarity where appropriate.\n" "4. Conclude with a summary and actionable recommendations.\n" "5. Include a References & Sources section ONLY if the analyst provided actual links from their analysis.\n" "6. Use ONLY the reference links that were explicitly provided by the analyst in their analysis. Do NOT create, invent, or hallucinate any links.\n" "7. If the analyst provided links, format them as clickable markdown links in the References section.\n" "8. If no links were provided by the analyst, do not include a References section at all.\n" "9. Never add fake or made-up links - only use links that were actually found and passed through the research chain." ), ) def run(self, topic: str) -> Iterator[RunResponse]: """ Orchestrates the research, analysis, and report writing process for a given topic. """ logger.info(f"Running deep researcher agent for topic: {topic}") # Step 1: Research research_content = self.searcher.run(topic) # logger.info(f"Searcher content: {research_content.content}") logger.info("Analysis started") # Step 2: Analysis analysis = self.analyst.run(research_content.content) # logger.info(f"Analyst analysis: {analysis.content}") logger.info("Report Writing Started") # Step 3: Report Writing report = self.writer.run(analysis.content, stream=True) yield from report def run_research(query: str) -> str: agent = DeepResearcherAgent() final_report_iterator = agent.run( topic=query, ) logger.info("Report Generated") # Collect all streaming content into a single string full_report = "" for chunk in final_report_iterator: if chunk.content: full_report += chunk.content return full_report if __name__ == "__main__": topic = "Extract information about Nebius AI Studio, including its features, capabilities, and applications from available sources." response = run_research(topic) print(response) [V0_FILE]python:file="agno/examples/deep_research_agent_gemini.py" import os from agno.agent import Agent from agno.models.nebius import Nebius from dotenv import load_dotenv from typing import Iterator from agno.utils.log import logger from agno.utils.pprint import pprint_run_response from agno.tools.scrapegraph import ScrapeGraphTools from agno.workflow import RunEvent, RunResponse, Workflow from pydantic import BaseModel, Field # load environment variables import os from dotenv import load_dotenv load_dotenv() # get api keys openai_api_key = os.getenv("OPENAI_API_KEY") exa_api_key = os.getenv("EXA_API_KEY") nebius_api_key = os.getenv("NEBIUS_API_KEY") class DeepResearcherAgent(Workflow): """ A multi-stage research workflow that: 1. Gathers information from the web using advanced scraping tools. 2. Analyzes and synthesizes the findings. 3. Produces a clear, well-structured report. """ # Searcher: Finds and extracts relevant information from the web searcher: Agent = Agent( tools=[ScrapeGraphTools(api_key=os.getenv("SGAI_API_KEY"))], model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), show_tool_calls=True, markdown=True, description=( "You are ResearchBot-X, an expert at finding and extracting high-quality, " "up-to-date information from the web. Your job is to gather comprehensive, " "reliable, and diverse sources on the given topic." ), instructions=( "1. Search for the most recent and authoritative and up-to-date sources (news, blogs, official docs, research papers, forums, etc.) on the topic.\n" "2. Extract key facts, statistics, and expert opinions.\n" "3. Cover multiple perspectives and highlight any disagreements or controversies.\n" "4. Include relevant statistics, data, and expert opinions where possible.\n" "5. Organize your findings in a clear, structured format (e.g., markdown table or sections by source type).\n" "6. If the topic is ambiguous, clarify with the user before proceeding.\n" "7. Be as comprehensive and verbose as possible—err on the side of including more detail.\n" "8. Mention the References & Sources of the Content. (It's Must)" ), ) # Analyst: Synthesizes and interprets the research findings analyst: Agent = Agent( model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), markdown=True, description=( "You are AnalystBot-X, a critical thinker who synthesizes research findings " "into actionable insights. Your job is to analyze, compare, and interpret the " "information provided by the researcher." ), instructions=( "1. Identify key themes, trends, and contradictions in the research.\n" "2. Highlight the most important findings and their implications.\n" "3. Suggest areas for further investigation if gaps are found.\n" "4. Present your analysis in a structured, easy-to-read format.\n" "5. Extract and list ONLY the reference links or sources that were ACTUALLY found and provided by the researcher in their findings. Do NOT create, invent, or hallucinate any links.\n" "6. If no links were provided by the researcher, do not include a References section.\n" "7. Don't add hallucinations or make up information. Use ONLY the links that were explicitly passed to you by the researcher.\n" "8. Verify that each link you include was actually present in the researcher's findings before listing it.\n" "9. If there's no Link found from the previous agent then just say, No reference Found." ), ) # Writer: Produces a final, polished report writer: Agent = Agent( model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), markdown=True, description=( "You are WriterBot-X, a professional technical writer. Your job is to craft " "a clear, engaging, and well-structured report based on the analyst's summary." ), instructions=( "1. Write an engaging introduction that sets the context.\n" "2. Organize the main findings into logical sections with headings.\n" "3. Use bullet points, tables, or lists for clarity where appropriate.\n" "4. Conclude with a summary and actionable recommendations.\n" "5. Include a References & Sources section ONLY if the analyst provided actual links from their analysis.\n" "6. Use ONLY the reference links that were explicitly provided by the analyst in their analysis. Do NOT create, invent, or hallucinate any links.\n" "7. If the analyst provided links, format them as clickable markdown links in the References section.\n" "8. If no links were provided by the analyst, do not include a References section at all.\n" "9. Never add fake or made-up links - only use links that were actually found and passed through the research chain." ), ) def run(self, topic: str) -> Iterator[RunResponse]: """ Orchestrates the research, analysis, and report writing process for a given topic. """ logger.info(f"Running deep researcher agent for topic: {topic}") # Step 1: Research research_content = self.searcher.run(topic) # logger.info(f"Searcher content: {research_content.content}") logger.info("Analysis started") # Step 2: Analysis analysis = self.analyst.run(research_content.content) # logger.info(f"Analyst analysis: {analysis.content}") logger.info("Report Writing Started") # Step 3: Report Writing report = self.writer.run(analysis.content, stream=True) yield from report def run_research(query: str) -> str: agent = DeepResearcherAgent() final_report_iterator = agent.run( topic=query, ) logger.info("Report Generated") # Collect all streaming content into a single string full_report = "" for chunk in final_report_iterator: if chunk.content: full_report += chunk.content return full_report if __name__ == "__main__": topic = "Extract information about Nebius AI Studio, including its features, capabilities, and applications from available sources." response = run_research(topic) print(response) [V0_FILE]python:file="agno/examples/fetch_mesages_from_lastNsessions.py" # Remove the tmp db file before running the script import os from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.storage.sqlite import SqliteStorage # load enviroinment variables from dotenv import load_dotenv load_dotenv() import os os.remove("tmp/data.db") agent = Agent( model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY")), user_id="user_1", storage=SqliteStorage(table_name="agent_sessions_new", db_file="tmp/data.db"), search_previous_sessions_history=True, # allow searching previous sessions num_history_sessions=2, # only include the last 2 sessions in the search to avoid context length issues show_tool_calls=True, ) session_1_id = "session_1_id" session_2_id = "session_2_id" session_3_id = "session_3_id" session_4_id = "session_4_id" session_5_id = "session_5_id" agent.print_response("What is the capital of South Africa?", session_id=session_1_id) agent.print_response("What is the capital of China?", session_id=session_2_id) agent.print_response("What is the capital of France?", session_id=session_3_id) agent.print_response("What is the capital of Japan?", session_id=session_4_id) agent.print_response( "What did I discuss in my previous conversations?", session_id=session_5_id ) # It should only include the last 2 sessions # To enable fetching messages from the last N sessions, you need to use the following flags: # search_previous_sessions_history: Set this to True to allow searching through previous sessions. # num_history_sessions: Specify the number of past sessions to include in the search. In this example, it is set to 2 to include only the last 2 sessions. It’s advisable to keep this number to 2 or 3 for now, as a larger number might fill up the context length of the model, potentially leading to performance issues. # These flags help manage the context length and ensure that only relevant session history is included in the conversation. [V0_FILE]python:file="agno/examples/hangdlingevents.py" from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.openai import OpenAIChat from agno.utils.pprint import pprint_run_response agent = Agent(model=OpenAIChat(id="gpt-4o-mini")) # Run agent and return the response as a variabl response_stream = agent.run("Your prompt", stream=True, stream_intermediate_steps=True) for event in response_stream: if event.event == "RunResponseContent": print(f"Content: {event.content}") elif event.event == "ToolCallStarted": print(f"Tool call started: {event.tool}") elif event.event == "ReasoningStep": print(f"Reasoning step: {event.content}") ... [V0_FILE]python:file="agno/examples/level_1_agent.py" from agno.agent import Agent from agno.models.anthropic import Claude from agno.tools.yfinance import YFinanceTools agent = Agent( model=Claude(id="claude-sonnet-4-20250514"), tools=[YFinanceTools(stock_price=True)], instructions="Use tables to display data. Don't include any other text.", markdown=True, debug_mode=True ) agent.print_response("What is the stock price of Google?", stream=True) [V0_FILE]python:file="agno/examples/level_2_agent.py" from agno.agent import Agent from agno.embedder.openai import OpenAIEmbedder from agno.knowledge.url import UrlKnowledge from agno.models.anthropic import Claude from agno.storage.sqlite import SqliteStorage from agno.vectordb.lancedb import LanceDb, SearchType from agno.vectordb.chroma import ChromaDb # ChromaDB for WINDOWS friendly storage from dotenv import load_dotenv import os # Load environment variables from .env file load_dotenv() # Now you can access the OPENAI_API_KEY api_key = os.getenv("OPENAI_API_KEY") # Replace the LanceDb section with: # Load Agno documentation in a knowledge base # You can also use `https://docs.agno.com/llms-full.txt` for the full documentation knowledge = UrlKnowledge( # urls=["https://docs.agno.com/introduction.md"], # urls=["https://nbviewer.org/format/script/github/openai/openai-cookbook/blob/main/examples/gpt-5/gpt-5_prompting_guide.ipynb"], # urls=["https://img1.wsimg.com/blobby/go/06564960-dbf7-455b-9ea8-137ed61235b0/Migraine%20Headache%20Megahandout.pdf"], urls=["https://cookbook.openai.com/examples/gpt-5/gpt-5_prompting_guide"], # vector_db=LanceDb( # uri="./lancedb", # change to tmp/lancedb for local storage # table_name="agno_docs", # search_type=SearchType.hybrid, # # Use OpenAI for embeddings # embedder=OpenAIEmbedder(id="text-embedding-ada-002", dimensions=1536, api_key=api_key), # ), vector_db=ChromaDb( collection="agno_docs", embedder=OpenAIEmbedder(id="text-embedding-ada-002", dimensions=1536, api_key=api_key), ) ) # changed from text-embedding-3-small to text-embedding-ada-002 due to profile tiering issuess # Store agent sessions in a SQLite database storage = SqliteStorage(table_name="agent_sessions", db_file="tmp/agent.db") agent = Agent( name="Agno Assist", model=Claude(id="claude-sonnet-4-20250514"), instructions=[ "Search your knowledge before answering the question.", "Only include the output in your response. No other text.", ], knowledge=knowledge, storage=storage, add_datetime_to_instructions=True, # Add the chat history to the messages add_history_to_messages=True, # Number of history runs num_history_runs=3, markdown=True, ) if __name__ == "__main__": # Load the knowledge base, comment out after first run # Set recreate to True to recreate the knowledge base if needed agent.knowledge.load(recreate=True) # change to False after first run agent.print_response("GPT-5 is more like a model improvement or a product improvement?", stream=True) [V0_FILE]python:file="agno/examples/level_3_agent.py" from agno.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory from agno.models.anthropic import Claude from agno.tools.reasoning import ReasoningTools from agno.tools.yfinance import YFinanceTools memory = Memory( # Use any model for creating and managing memories model=Claude(id="claude-sonnet-4-20250514"), # Store memories in a SQLite database db=SqliteMemoryDb(table_name="user_memories", db_file="tmp/agent.db"), # We disable deletion by default, enable it if needed delete_memories=True, clear_memories=True, ) agent = Agent( model=Claude(id="claude-sonnet-4-20250514"), tools=[ ReasoningTools(add_instructions=True), YFinanceTools(stock_price=True, analyst_recommendations=True, company_info=True, company_news=True), ], # User ID for storing memories, `default` if not provided user_id="ava", instructions=[ "Use tables to display data.", "Include sources in your response.", "Only include the report in your response. No other text.", ], memory=memory, # Let the Agent manage its memories enable_agentic_memory=True, markdown=True, ) if __name__ == "__main__": # This will create a memory that "ava's" favorite stocks are NVIDIA and TSLA agent.print_response( "My favorite stocks are NVIDIA and TSLA", stream=True, show_full_reasoning=True, stream_intermediate_steps=True, ) # This will use the memory to answer the question agent.print_response( "Can you compare my favorite stocks?", stream=True, show_full_reasoning=True, stream_intermediate_steps=True, ) [V0_FILE]python:file="agno/examples/level_4_agent.py" from agno.agent import Agent from agno.models.google import Gemini from agno.models.openai import OpenAIChat from agno.models.anthropic import Claude from agno.team.team import Team from agno.tools.duckduckgo import DuckDuckGoTools from agno.tools.reasoning import ReasoningTools from agno.tools.yfinance import YFinanceTools from dotenv import load_dotenv import os # Load environment variables from .env file load_dotenv() # Get the API keys gemini_api_key = os.getenv("GEMINI_API_KEY") openai_api_key = os.getenv("OPENAI_API_KEY") anthropic_api_key = os.getenv("ANTHROPIC_API_KEY") web_agent = Agent( name="Web Search Agent", role="Handle web search requests and general research", model=Gemini(id="gemini-2.0-flash", api_key=gemini_api_key), tools=[DuckDuckGoTools()], instructions="Always include sources", add_datetime_to_instructions=True, ) finance_agent = Agent( name="Finance Agent", role="Handle financial data requests and market analysis", model=OpenAIChat(id="gpt-4o", api_key=openai_api_key), tools=[YFinanceTools(stock_price=True, stock_fundamentals=True,analyst_recommendations=True, company_info=True)], instructions=[ "Use tables to display stock prices, fundamentals (P/E, Market Cap), and recommendations.", "Clearly state the company name and ticker symbol.", "Focus on delivering actionable financial insights.", ], add_datetime_to_instructions=True, ) reasoning_finance_team = Team( name="Reasoning Finance Team", mode="coordinate", model=Claude(id="claude-sonnet-4-20250514", api_key=anthropic_api_key), members=[web_agent, finance_agent], tools=[ReasoningTools(add_instructions=True)], instructions=[ "Collaborate to provide comprehensive financial and investment insights", "Consider both fundamental analysis and market sentiment", "Use tables and charts to display data clearly and professionally", "Present findings in a structured, easy-to-follow format", "Only output the final consolidated analysis, not individual agent responses", ], markdown=True, show_members_responses=True, enable_agentic_context=True, add_datetime_to_instructions=True, success_criteria="The team has provided a complete financial analysis with data, visualizations, risk assessment, and actionable investment recommendations supported by quantitative analysis and market research.", ) if __name__ == "__main__": reasoning_finance_team.print_response("""Compare the tech sector giants (AAPL, GOOGL, MSFT) performance: 1. Get financial data for all three companies 2. Analyze recent news affecting the tech sector 3. Calculate comparative metrics and correlations 4. Recommend portfolio allocation weights""", stream=True, show_full_reasoning=True, stream_intermediate_steps=True, ) [V0_FILE]python:file="agno/examples/level_5_workflow.py" from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.openai import OpenAIChat from agno.utils.log import logger from agno.utils.pprint import pprint_run_response from agno.workflow import Workflow from dotenv import load_dotenv import os # Load environment variables from .env file load_dotenv() # Get the API keys api_key = os.getenv("OPENAI_API_KEY") class CacheWorkflow(Workflow): # Add agents or teams as attributes on the workflow agent = Agent(model=OpenAIChat(id="gpt-4o", api_key=api_key)) # Write the logic in the `run()` method def run(self, message: str) -> Iterator[RunResponse]: logger.info(f"Checking cache for '{message}'") # Check if the output is already cached if self.session_state.get(message): logger.info(f"Cache hit for '{message}'") yield RunResponse( run_id=self.run_id, content=self.session_state.get(message) ) return logger.info(f"Cache miss for '{message}'") # Run the agent and yield the response yield from self.agent.run(message, stream=True) # Cache the output after response is yielded self.session_state[message] = self.agent.run_response.content if __name__ == "__main__": workflow = CacheWorkflow() # Run workflow (this is takes ~1s) response: Iterator[RunResponse] = workflow.run(message="Tell me a joke.") # Print the response pprint_run_response(response, markdown=True, show_time=True) # Run workflow again (this is immediate because of caching) response: Iterator[RunResponse] = workflow.run(message="Tell me a joke.") # Print the response pprint_run_response(response, markdown=True, show_time=True) [V0_FILE]python:file="agno/examples/memo_demo_gemini.py" from agno.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory from agno.models.google import Gemini from agno.storage.sqlite import SqliteStorage from rich.pretty import pprint from dotenv import load_dotenv load_dotenv() import os # Get API keys gemini_api_key = os.getenv("GEMINI_API_KEY") # UserId for the memories user_id = "ava" # Database file for memory and storage db_file = "tmp/agent.db" # Initialize memory.v2 memory = Memory( # Use Gemini model for creating memories model=Gemini(id="gemini-2.0-flash", api_key=gemini_api_key), db=SqliteMemoryDb(table_name="user_memories", db_file=db_file), ) # Initialize storage storage = SqliteStorage(table_name="agent_sessions", db_file=db_file) # Initialize Agent memory_agent = Agent( model=Gemini(id="gemini-2.0-flash", api_key=gemini_api_key), # Store memories in a database memory=memory, # Give the Agent the ability to update memories enable_agentic_memory=True, # OR - Run the MemoryManager after each response enable_user_memories=True, # Store the chat history in the database storage=storage, # Add the chat history to the messages add_history_to_messages=True, # Number of history runs num_history_runs=3, markdown=True, ) memory.clear() memory_agent.print_response( "My name is Ava and I like to ski.", user_id=user_id, stream=True, stream_intermediate_steps=True, ) print("Memories about Ava:") pprint(memory.get_user_memories(user_id=user_id)) memory_agent.print_response( "I live in san francisco, where should i move within a 4 hour drive?", user_id=user_id, stream=True, stream_intermediate_steps=True, ) print("Memories about Ava:") pprint(memory.get_user_memories(user_id=user_id)) [V0_FILE]python:file="agno/examples/metrics.py" from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.google import Gemini from agno.tools.yfinance import YFinanceTools from rich.pretty import pprint # load environment variables from dotenv import load_dotenv load_dotenv() import os agent = Agent( model=Gemini(id="gemini-2.0-flash-001", api_key=os.getenv("GEMINI_API_KEY")), tools=[YFinanceTools(stock_price=True)], markdown=True, show_tool_calls=True, ) agent.print_response( "What is the stock price of NVDA", stream=True ) # Print metrics per message if agent.run_response.messages: for message in agent.run_response.messages: if message.role == "assistant": if message.content: print(f"Message: {message.content}") elif message.tool_calls: print(f"Tool calls: {message.tool_calls}") print("---" * 5, "Metrics", "---" * 5) pprint(message.metrics) print("---" * 20) # Print the aggregated metrics for the whole run print("---" * 5, "Collected Metrics", "---" * 5) pprint(agent.run_response.metrics) # Print the aggregated metrics for the whole session print("---" * 5, "Session Metrics", "---" * 5) pprint(agent.session_metrics) [V0_FILE]python:file="agno/examples/research_agent.py" from datetime import datetime from pathlib import Path from textwrap import dedent from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.exa import ExaTools from dotenv import load_dotenv import os # Load environment variables from .env file load_dotenv() # Get the API keys openai_api_key = os.getenv("OPENAI_API_KEY") # load exa api key exa_api_key = os.getenv("EXA_API_KEY") today = datetime.now().strftime("%Y-%m-%d") agent = Agent( model=OpenAIChat(id="gpt-4o", api_key=openai_api_key), tools=[ExaTools(start_published_date=today, type="keyword", api_key=exa_api_key)], description=dedent("""\ You are Professor X-1000, a distinguished AI research scientist with expertise in analyzing and synthesizing complex information. Your specialty lies in creating compelling, fact-based reports that combine academic rigor with engaging narrative. Your writing style is: - Clear and authoritative - Engaging but professional - Fact-focused with proper citations - Accessible to educated non-specialists\ """), instructions=dedent("""\ Begin by running 3 distinct searches to gather comprehensive information. Analyze and cross-reference sources for accuracy and relevance. Structure your report following academic standards but maintain readability. Include only verifiable facts with proper citations. Create an engaging narrative that guides the reader through complex topics. End with actionable takeaways and future implications.\ """), expected_output=dedent("""\ A professional research report in markdown format: # {Compelling Title That Captures the Topic's Essence} ## Executive Summary {Brief overview of key findings and significance} ## Introduction {Context and importance of the topic} {Current state of research/discussion} ## Key Findings {Major discoveries or developments} {Supporting evidence and analysis} ## Implications {Impact on field/society} {Future directions} ## Key Takeaways - {Bullet point 1} - {Bullet point 2} - {Bullet point 3} ## References - [Source 1](link) - Key finding/quote - [Source 2](link) - Key finding/quote - [Source 3](link) - Key finding/quote --- Report generated by Professor X-1000 Advanced Research Systems Division Date: {current_date}\ """), markdown=True, show_tool_calls=True, add_datetime_to_instructions=True, ) # Example usage if __name__ == "__main__": # Generate a research report on a cutting-edge topic agent.print_response( "Research the latest stock prices changes in FIGMA since it's IPO", stream=True ) # More example prompts to try: """ Try these research topics: 1. "Analyze the current state of solid-state batteries" 2. "Research recent breakthroughs in CRISPR gene editing" 3. "Investigate the development of autonomous vehicles" 4. "Explore advances in quantum machine learning" 5. "Study the impact of artificial intelligence on healthcare" """ [V0_FILE]python:file="agno/examples/running_agent.py" from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.openai import OpenAIChat from agno.utils.pprint import pprint_run_response from dotenv import load_dotenv import os # Load environment variables from .env file load_dotenv() # Get the API keys openai_api_key = os.getenv("OPENAI_API_KEY") agent = Agent(model=OpenAIChat(id="gpt-4o-mini", api_key=openai_api_key )) # Run agent and return the response as a variable response: RunResponse = agent.run("Tell me a 5 second short story about a robot") # Print the response in markdown format pprint_run_response(response, markdown=True) [V0_FILE]python:file="agno/examples/session_state_storage.py" """Run `pip install agno openai sqlalchemy` to install dependencies.""" from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.storage.sqlite import SqliteStorage from dotenv import load_dotenv load_dotenv() import os # Define a tool that adds an item to the shopping list def add_item(agent: Agent, item: str) -> str: """Add an item to the shopping list.""" if item not in agent.session_state["shopping_list"]: agent.session_state["shopping_list"].append(item) return f"The shopping list is now {agent.session_state['shopping_list']}" agent = Agent( model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY")), # Fix the session id to continue the same session across execution cycles session_id="fixed_id_for_demo", # Initialize the session state with an empty shopping list session_state={"shopping_list": []}, # Add a tool that adds an item to the shopping list tools=[add_item], # Store the session state in a SQLite database storage=SqliteStorage(table_name="agent_sessions", db_file="tmp/data.db"), # Add the current shopping list from the state in the instructions instructions="Current shopping list is: {shopping_list}", # Important: Set `add_state_in_messages=True` # to make `{shopping_list}` available in the instructions add_state_in_messages=True, markdown=True, ) # Example usage agent.print_response("What's on my shopping list?", stream=True) print(f"Session state: {agent.session_state}") agent.print_response("Add milk, eggs, and bread", stream=True) print(f"Session state: {agent.session_state}") # this can be used as todo list for exercises/medications/doctor appointments preps [V0_FILE]python:file="agno/examples/sessions.py" from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.memory.v2 import Memory # load environment variables from dotenv import load_dotenv load_dotenv() import os agent = Agent( model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY")), # Multi-user, multi-session only work with Memory.v2 memory=Memory(), add_history_to_messages=True, num_history_runs=3, ) user_1_id = "user_101" user_2_id = "user_102" user_1_session_id = "session_101" user_2_session_id = "session_102" # Start the session with user 1 agent.print_response( "Tell me a 5 second short story about a robot.", user_id=user_1_id, session_id=user_1_session_id, ) # Continue the session with user 1 agent.print_response("Now tell me a joke.", user_id=user_1_id, session_id=user_1_session_id) # Start the session with user 2 agent.print_response("Tell me about quantum physics.", user_id=user_2_id, session_id=user_2_session_id) # Continue the session with user 2 agent.print_response("What is the speed of light?", user_id=user_2_id, session_id=user_2_session_id) # Ask the agent to give a summary of the conversation, this will use the history from the previous messages agent.print_response( "Give me a summary of our conversation.", user_id=user_1_id, session_id=user_1_session_id, ) [V0_FILE]python:file="agno/examples/shopping_list_maintain_state_across_runs.py" from textwrap import dedent from agno.agent import Agent from agno.models.openai import OpenAIChat # load environment variables from dotenv import load_dotenv load_dotenv() import os # Define tools to manage our shopping list def add_item(agent: Agent, item: str) -> str: """Add an item to the shopping list and return confirmation.""" # Add the item if it's not already in the list if item.lower() not in [i.lower() for i in agent.session_state["shopping_list"]]: agent.session_state["shopping_list"].append(item) return f"Added '{item}' to the shopping list" else: return f"'{item}' is already in the shopping list" def remove_item(agent: Agent, item: str) -> str: """Remove an item from the shopping list by name.""" # Case-insensitive search for i, list_item in enumerate(agent.session_state["shopping_list"]): if list_item.lower() == item.lower(): agent.session_state["shopping_list"].pop(i) return f"Removed '{list_item}' from the shopping list" return f"'{item}' was not found in the shopping list" def list_items(agent: Agent) -> str: """List all items in the shopping list.""" shopping_list = agent.session_state["shopping_list"] if not shopping_list: return "The shopping list is empty." items_text = "\n".join([f"- {item}" for item in shopping_list]) return f"Current shopping list:\n{items_text}" # Create a Shopping List Manager Agent that maintains state agent = Agent( model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY") ), # Initialize the session state with an empty shopping list session_state={"shopping_list": []}, tools=[add_item, remove_item, list_items], # You can use variables from the session state in the instructions instructions=dedent("""\ Your job is to manage a shopping list. The shopping list starts empty. You can add items, remove items by name, and list all items. Current shopping list: {shopping_list} """), show_tool_calls=True, add_state_in_messages=True, markdown=True, ) # Example usage agent.print_response("Add milk, eggs, and bread to the shopping list", stream=True) print(f"Session state: {agent.session_state}") agent.print_response("I got bread", stream=True) print(f"Session state: {agent.session_state}") agent.print_response("I need apples and oranges", stream=True) print(f"Session state: {agent.session_state}") agent.print_response("whats on my list?", stream=True) print(f"Session state: {agent.session_state}") agent.print_response("Clear everything from my list and start over with just bananas and yogurt", stream=True) print(f"Session state: {agent.session_state}") [V0_FILE]python:file="agno/examples/state_in_instructions.py" from textwrap import dedent from agno.agent import Agent from agno.models.openai import OpenAIChat # load environment variables from dotenv import load_dotenv load_dotenv() import os agent = Agent( model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY") ), # Initialize the session state with a variable session_state={"user_name": "John"}, # You can use variables from the session state in the instructions instructions="Users name is {user_name}", show_tool_calls=True, add_state_in_messages=True, markdown=True, ) agent.print_response("What is my name?", stream=True) # this can be used for checking the user static profiles [V0_FILE]python:file="agno/examples/streaming_response.py" from typing import Iterator from agno.agent import Agent, RunResponseEvent from agno.models.openai import OpenAIChat from agno.utils.pprint import pprint_run_response # load environment variables from dotenv import load_dotenv load_dotenv() import os agent = Agent(model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY"))) # Run agent and return the response as a stream response_stream: Iterator[RunResponseEvent] = agent.run( "Tell me a 5 second short story about a lion", stream=True, stream_intermediate_steps=True #doesnt' work for gpt-4o-mini ) # Print the response stream in markdown format pprint_run_response(response_stream, markdown=True) [V0_FILE]python:file="agno/examples/streaming_response_claude.py" from agno.agent import Agent from agno.models.anthropic import Claude from agno.tools.reasoning import ReasoningTools from agno.tools.yfinance import YFinanceTools # Claude with reasoning tools for complex analysis reasoning_agent = Agent( model=Claude(id="claude-3-7-sonnet-latest"), #controls the flow of execution, decides whether to reason, act or respond tools=[ ReasoningTools( think=True, analyze=True, add_instructions=True, ), YFinanceTools( stock_price=True, analyst_recommendations=True, company_info=True, ), ], #enable an Agent to take actions and interacts with external systems instructions="Use structured reasoning for complex problems", # how Agno program in the agent, teaching it how to use tools and respond stream_intermediate_steps=True, show_tool_calls=True, markdown=True, ) # This question will trigger multiple intermediate steps response_stream = reasoning_agent.run( "What is the stock price of NVDA and should I invest in it? " "Consider market trends, financial health, and risk factors.", stream=True, stream_intermediate_steps=True ) # for chunk in response_stream: # print(chunk) # print("---" * 20) # response_stream = agent.run("Your prompt", stream=True, stream_intermediate_steps=True) for event in response_stream: if event.event == "RunResponseContent": print(f"Content: {event.content}") elif event.event == "ToolCallStarted": print(f"Tool call started: {event.tool}") elif event.event == "ReasoningStep": print(f"Reasoning step: {event.content}") ... [V0_FILE]python:file="agno/examples/streaming_response_gemini.py" from agno.agent import Agent from agno.models.google import Gemini from agno.tools.reasoning import ReasoningTools #load environment variables from dotenv import load_dotenv load_dotenv() import os # Gemini with structured reasoning reasoning_agent = Agent( model=Gemini(id="gemini-2.5-pro-preview-03-25", api_key=os.getenv("GEMINI_API_KEY")), tools=[ ReasoningTools( think=True, analyze=True, add_instructions=True, ) ], instructions="Use structured reasoning for complex problems", stream_intermediate_steps=True, show_tool_calls=True, markdown=True, debug_mode=True, ) # Stream with intermediate steps for your lion story response_stream = reasoning_agent.run( "Tell me a 5 second short story about a lion", stream=True, stream_intermediate_steps=True ) for chunk in response_stream: print(chunk) print("---" * 20) [V0_FILE]python:file="agno/examples/structured_input.py" from typing import List from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.hackernews import HackerNewsTools from pydantic import BaseModel, Field #load environment variables from dotenv import load_dotenv load_dotenv() import os class ResearchTopic(BaseModel): """Structured research topic with specific requirements""" topic: str focus_areas: List[str] = Field(description="Specific areas to focus on") target_audience: str = Field(description="Who this research is for") sources_required: int = Field(description="Number of sources needed", default=5) # Define agents hackernews_agent = Agent( name="Hackernews Agent", model=OpenAIChat(id="gpt-4o-mini-2024-07-18", api_key=os.getenv("OPENAI_API_KEY")), tools=[HackerNewsTools()], role="Extract key insights and content from Hackernews posts", ) hackernews_agent.print_response( message=ResearchTopic( topic="AI", focus_areas=["AI", "Machine Learning"], target_audience="Developers", sources_required=5, ) ) [V0_FILE]markdown:file="agno/index.md" # Agno Documentation Index *Generated from llms-full.txt* ## Core - [Agent API](core/agent_api.md) (lines 1-27) - [What are Agents?](core/what_are_agents.md) (lines 322-470) - [Running your Agent](core/running_your_agent.md) (lines 1513-1534) - [Sessions](core/sessions.md) (lines 1723-1751) - [What are Agents?](core/what_are_agents.md) (lines 59929-60038) - [Store agent sessions in a SQLite database](core/store_agent_sessions_in_a_sqlite_database.md) (lines 60051-60168) - [Running the Agent API on AWS](core/running_the_agent_api_on_aws.md) (lines 85213-85246) - [Agent API: FastAPI and Postgres](core/agent_api_fastapi_and_postgres.md) (lines 85246-85277) ## Ui - [A beautiful UI for your Agents](ui/a_beautiful_ui_for_your_agents.md) (lines 27-44) - [Get Started with Agent UI](ui/get_started_with_agent_ui.md) (lines 44-190) ## Knowledge - [Knowledge](knowledge/knowledge.md) (lines 470-492) - [Create a knowledge base for the Agent](knowledge/create_a_knowledge_base_for_the_agent.md) (lines 492-495) - [Add information to the knowledge base](knowledge/add_information_to_the_knowledge_base.md) (lines 495-498) - [Add the knowledge base to the Agent and](knowledge/add_the_knowledge_base_to_the_agent_and.md) (lines 498-499) - [give it a tool to search the knowledge base as needed](knowledge/give_it_a_tool_to_search_the_knowledge_base_as_needed.md) (lines 499-736) - [Memory](knowledge/memory.md) (lines 736-778) - [Database file for memory and storage](knowledge/database_file_for_memory_and_storage.md) (lines 780-783) - [Initialize memory.v2](knowledge/initialize_memory_v2.md) (lines 783-789) - [Agentic Chunking](knowledge/agentic_chunking.md) (lines 5316-5353) - [Document Chunking](knowledge/document_chunking.md) (lines 5353-5391) - [Fixed Size Chunking](knowledge/fixed_size_chunking.md) (lines 5391-5428) - [Recursive Chunking](knowledge/recursive_chunking.md) (lines 5428-5464) - [Semantic Chunking](knowledge/semantic_chunking.md) (lines 5464-5500) - [Example usage with a PDF knowledge base](knowledge/example_usage_with_a_pdf_knowledge_base.md) (lines 5547-5562) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 5629-5665) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 5687-5715) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 5736-5761) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 5782-5811) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 5832-5858) - [Create knowledge base](knowledge/create_knowledge_base.md) (lines 5870-5881) - [Add information to the knowledge base](knowledge/add_information_to_the_knowledge_base.md) (lines 5881-5884) - [Add the knowledge base to the Agent](knowledge/add_the_knowledge_base_to_the_agent.md) (lines 5884-5903) - [Custom configuration with late chunking for long documents](knowledge/custom_configuration_with_late_chunking_for_long_documents.md) (lines 5948-5955) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 5963-5997) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 6018-6048) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 6071-6099) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 6120-6151) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 6172-6195) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 6216-6240) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 6261-6286) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 6307-6337) - [Agent with User Memory](knowledge/agent_with_user_memory.md) (lines 8315-8353) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 8353-8411) - [Initialize knowledge base](knowledge/initialize_knowledge_base.md) (lines 8773-8784) - [knowledge_agent prompt](knowledge/knowledge_agent_prompt.md) (lines 9819-9820) - [Agent with User Memory](knowledge/agent_with_user_memory.md) (lines 10264-10301) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 10301-10368) - [Agent With User Memory](knowledge/agent_with_user_memory.md) (lines 10576-10616) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 10616-10683) - [Load the knowledge base: Comment out after first run](knowledge/load_the_knowledge_base_comment_out_after_first_run.md) (lines 12704-12757) - [ArXiv Knowledge Base](knowledge/arxiv_knowledge_base.md) (lines 12917-12931) - [Create a knowledge base with the ArXiv documents](knowledge/create_a_knowledge_base_with_the_arxiv_documents.md) (lines 12931-12940) - [Load the knowledge base](knowledge/load_the_knowledge_base.md) (lines 12940-12943) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 12943-12949) - [Ask the agent about the knowledge base](knowledge/ask_the_agent_about_the_knowledge_base.md) (lines 12949-12994) - [Combined Knowledge Base](knowledge/combined_knowledge_base.md) (lines 12994-13014) - [Create CSV knowledge base](knowledge/create_csv_knowledge_base.md) (lines 13014-13023) - [Create PDF URL knowledge base](knowledge/create_pdf_url_knowledge_base.md) (lines 13023-13032) - [Create Website knowledge base](knowledge/create_website_knowledge_base.md) (lines 13032-13042) - [Create Local PDF knowledge base](knowledge/create_local_pdf_knowledge_base.md) (lines 13042-13051) - [Combine knowledge bases](knowledge/combine_knowledge_bases.md) (lines 13051-13065) - [Initialize the Agent with the combined knowledge base](knowledge/initialize_the_agent_with_the_combined_knowledge_base.md) (lines 13065-13073) - [CSV Knowledge Base](knowledge/csv_knowledge_base.md) (lines 13116-13140) - [Load the knowledge base](knowledge/load_the_knowledge_base.md) (lines 13140-13143) - [Initialize the Agent with the knowledge_base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 13143-13149) - [CSV URL Knowledge Base](knowledge/csv_url_knowledge_base.md) (lines 13192-13262) - [Document Knowledge Base](knowledge/document_knowledge_base.md) (lines 13262-13288) - [Create a knowledge base with the loaded documents](knowledge/create_a_knowledge_base_with_the_loaded_documents.md) (lines 13294-13303) - [Load the knowledge base](knowledge/load_the_knowledge_base.md) (lines 13303-13306) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 13306-13311) - [Ask the agent about the knowledge base](knowledge/ask_the_agent_about_the_knowledge_base.md) (lines 13311-13356) - [DOCX Knowledge Base](knowledge/docx_knowledge_base.md) (lines 13356-13372) - [Create a knowledge base with the DOCX files from the data/docs directory](knowledge/create_a_knowledge_base_with_the_docx_files_from_the_data_docs_directory.md) (lines 13372-13380) - [Load the knowledge base](knowledge/load_the_knowledge_base.md) (lines 13380-13383) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 13383-13389) - [Ask the agent about the knowledge base](knowledge/ask_the_agent_about_the_knowledge_base.md) (lines 13389-13432) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 13462-13463) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 13464-13465) - [Step 2: Query the knowledge base with Agent using filters from query automatically](knowledge/step_2_query_the_knowledge_base_with_agent_using_filters_from_query_automatically.md) (lines 13516-13517) - [Now use the downloaded paths in knowledge base initialization](knowledge/now_use_the_downloaded_paths_in_knowledge_base_initialization.md) (lines 13600-13648) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 13651-13652) - [Initialize the Agent with the knowledge base and filters](knowledge/initialize_the_agent_with_the_knowledge_base_and_filters.md) (lines 13655-13662) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 13670-13671) - [search_knowledge=True,](knowledge/search_knowledge_true.md) (lines 13671-13672) - [knowledge_filters={"user_id": "taylor_brooks"},](knowledge/knowledge_filters_user_id_taylor_brooks.md) (lines 13677-13678) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 13733-13734) - [When loading the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_loading_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 13735-13736) - [Initialize the DocxKnowledgeBase](knowledge/initialize_the_docxknowledgebase.md) (lines 13738-13750) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 13774-13775) - [Initialize the Agent with the knowledge base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 13779-13790) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 13791-13792) - [search_knowledge=True,](knowledge/search_knowledge_true.md) (lines 13792-13793) - [knowledge_filters = {"user_id": "jordan_mitchell"},](knowledge/knowledge_filters_user_id_jordan_mitchell.md) (lines 13796-13797) - [Knowledge filtering using Traditional RAG](knowledge/knowledge_filtering_using_traditional_rag.md) (lines 13824-13840) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 13852-13899) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 13902-13904) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 13920-13921) - [search_knowledge=False,](knowledge/search_knowledge_false.md) (lines 13922-13923) - [knowledge_filters={"user_id": "taylor_brooks"},](knowledge/knowledge_filters_user_id_taylor_brooks.md) (lines 13928-13929) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 13980-13981) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 13982-13983) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 14034-14035) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 14100-14101) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 14102-14103) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 14154-14155) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 14219-14220) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 14221-14222) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 14273-14274) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 14334-14335) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 14336-14337) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 14392-14393) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 14456-14457) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 14458-14459) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 14510-14511) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 14584-14631) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 14634-14635) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 14696-14697) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 14698-14699) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 14750-14751) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 14831-14832) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 14833-14834) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 14885-14886) - [Initialize the Agent with the knowledge base and filters](knowledge/initialize_the_agent_with_the_knowledge_base_and_filters.md) (lines 14889-14926) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 14948-14949) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 14950-14951) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 15009-15010) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 15077-15078) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 15079-15080) - [Step 2: Query the knowledge base with Agent using filters from query automatically](knowledge/step_2_query_the_knowledge_base_with_agent_using_filters_from_query_automatically.md) (lines 15131-15132) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 15216-15217) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 15218-15219) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 15270-15271) - [Initialize the Agent with the knowledge base and filters](knowledge/initialize_the_agent_with_the_knowledge_base_and_filters.md) (lines 15274-15281) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 15289-15290) - [search_knowledge=True,](knowledge/search_knowledge_true.md) (lines 15290-15291) - [knowledge_filters={"user_id": "taylor_brooks"},](knowledge/knowledge_filters_user_id_taylor_brooks.md) (lines 15296-15297) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 15352-15353) - [When loading the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_loading_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 15354-15355) - [Initialize the JSONKnowledgeBase](knowledge/initialize_the_jsonknowledgebase.md) (lines 15357-15369) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 15393-15394) - [Initialize the Agent with the knowledge base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 15398-15409) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 15410-15411) - [search_knowledge=True,](knowledge/search_knowledge_true.md) (lines 15411-15412) - [knowledge_filters = {"user_id": "jordan_mitchell"},](knowledge/knowledge_filters_user_id_jordan_mitchell.md) (lines 15415-15416) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 15473-15474) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 15475-15476) - [Step 2: Query the knowledge base with Agent using filters from query automatically](knowledge/step_2_query_the_knowledge_base_with_agent_using_filters_from_query_automatically.md) (lines 15527-15528) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 15595-15642) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 15645-15647) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 15711-15712) - [When loading the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_loading_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 15713-15714) - [Initialize the PDFKnowledgeBase](knowledge/initialize_the_pdfknowledgebase.md) (lines 15716-15728) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 15752-15753) - [Initialize the Agent with the knowledge base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 15757-15792) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 15811-15812) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 15813-15814) - [Step 2: Query the knowledge base with Agent using filters from query automatically](knowledge/step_2_query_the_knowledge_base_with_agent_using_filters_from_query_automatically.md) (lines 15841-15842) - [Step 1: Initialize knowledge base with URLs and metadata](knowledge/step_1_initialize_knowledge_base_with_urls_and_metadata.md) (lines 15918-15919) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 15920-15921) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 15948-15949) - [Initialize the Agent with the knowledge base and filters](knowledge/initialize_the_agent_with_the_knowledge_base_and_filters.md) (lines 15952-15960) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 15968-15969) - [search_knowledge=True,](knowledge/search_knowledge_true.md) (lines 15969-15970) - [knowledge_filters={"cuisine": "Cape"},](knowledge/knowledge_filters_cuisine_cape.md) (lines 15975-15976) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 16022-16023) - [When loading the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_loading_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 16024-16025) - [Initialize knowledge base](knowledge/initialize_knowledge_base.md) (lines 16027-16044) - [Initialize the Agent with the knowledge base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 16045-16058) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 16060-16061) - [search_knowledge=True,](knowledge/search_knowledge_true.md) (lines 16061-16062) - [knowledge_filters={"cuisine": "Cape"},](knowledge/knowledge_filters_cuisine_cape.md) (lines 16065-16066) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 16123-16124) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 16125-16126) - [Step 2: Query the knowledge base with Agent using filters from query automatically](knowledge/step_2_query_the_knowledge_base_with_agent_using_filters_from_query_automatically.md) (lines 16177-16178) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 16262-16263) - [When initializing the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 16264-16265) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 16316-16317) - [Initialize the Agent with the knowledge base and filters](knowledge/initialize_the_agent_with_the_knowledge_base_and_filters.md) (lines 16320-16327) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 16335-16336) - [search_knowledge=True,](knowledge/search_knowledge_true.md) (lines 16336-16337) - [knowledge_filters={"user_id": "taylor_brooks"},](knowledge/knowledge_filters_user_id_taylor_brooks.md) (lines 16342-16343) - [Step 1: Initialize knowledge base with documents and metadata](knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md) (lines 16398-16399) - [When loading the knowledge base, we can attach metadata that will be used for filtering](knowledge/when_loading_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md) (lines 16400-16401) - [Initialize the TextKnowledgeBase](knowledge/initialize_the_textknowledgebase.md) (lines 16403-16415) - [Step 2: Query the knowledge base with different filter combinations](knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md) (lines 16439-16440) - [Initialize the Agent with the knowledge base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 16444-16455) - [knowledge=knowledge_base,](knowledge/knowledge_knowledge_base.md) (lines 16456-16457) - [search_knowledge=True,](knowledge/search_knowledge_true.md) (lines 16457-16458) - [knowledge_filters = {"user_id": "jordan_mitchell"},](knowledge/knowledge_filters_user_id_jordan_mitchell.md) (lines 16461-16462) - [Built-in Memory](knowledge/built-in_memory.md) (lines 16489-16510) - [-*- Print the messages in the memory](knowledge/-_-_print_the_messages_in_the_memory.md) (lines 16512-16520) - [-*- Print the messages in the memory](knowledge/-_-_print_the_messages_in_the_memory.md) (lines 16522-16556) - [Standalone Memory Operations](knowledge/standalone_memory_operations.md) (lines 16556-16571) - [Add a memory for the default user](knowledge/add_a_memory_for_the_default_user.md) (lines 16571-16578) - [Delete a memory](knowledge/delete_a_memory.md) (lines 16593-16601) - [Replace a memory](knowledge/replace_a_memory.md) (lines 16601-16639) - [Persistent Memory with SQLite](knowledge/persistent_memory_with_sqlite.md) (lines 16639-16663) - [memory.add_user_memory(](knowledge/memory_add_user_memory.md) (lines 16670-16671) - [memory=UserMemory(memory="The user works at a softward company called Agno", topics=["name"]),](knowledge/memory_usermemory_memory_the_user_works_at_a_softward_company_called_agno_topics_name.md) (lines 16671-16672) - [Custom Memory Creation](knowledge/custom_memory_creation.md) (lines 16707-16726) - [Use default memory manager](knowledge/use_default_memory_manager.md) (lines 16764-16768) - [Memory Search](knowledge/memory_search.md) (lines 16827-16847) - [Agent With Memory](knowledge/agent_with_memory.md) (lines 16912-16936) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 16939-16965) - [Agentic Memory](knowledge/agentic_memory.md) (lines 17021-17043) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 17046-17130) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 17157-17160) - [We can get the session summary from memory as well](knowledge/we_can_get_the_session_summary_from_memory_as_well.md) (lines 17209-17247) - [Multiple Agents Sharing Memory](knowledge/multiple_agents_sharing_memory.md) (lines 17247-17267) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 17270-17341) - [Custom Memory](knowledge/custom_memory.md) (lines 17341-17365) - [You can also override the entire `system_message` for the memory manager](knowledge/you_can_also_override_the_entire_system_message_for_the_memory_manager.md) (lines 17365-17373) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 17387-17441) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 17473-17592) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 17622-17757) - [Memory References](knowledge/memory_references.md) (lines 17757-17788) - [memory=memory,](knowledge/memory_memory.md) (lines 17791-17792) - [add_memory_references=False,](knowledge/add_memory_references_false.md) (lines 17793-17794) - [MongoDB Memory Storage](knowledge/mongodb_memory_storage.md) (lines 17904-17924) - [Create MongoDB memory database](knowledge/create_mongodb_memory_database.md) (lines 17929-17936) - [Create memory instance with MongoDB backend](knowledge/create_memory_instance_with_mongodb_backend.md) (lines 17936-17939) - [Create agent with memory](knowledge/create_agent_with_memory.md) (lines 17942-18002) - [PostgreSQL Memory Storage](knowledge/postgresql_memory_storage.md) (lines 18002-18079) - [Redis Memory Storage](knowledge/redis_memory_storage.md) (lines 18079-18097) - [Create Redis memory database](knowledge/create_redis_memory_database.md) (lines 18097-18104) - [Create memory instance with Redis backend](knowledge/create_memory_instance_with_redis_backend.md) (lines 18104-18107) - [Create agent with memory and Redis storage](knowledge/create_agent_with_memory_and_redis_storage.md) (lines 18114-18123) - [Second interaction - testing if memory was stored](knowledge/second_interaction_-_testing_if_memory_was_stored.md) (lines 18131-18139) - [SQLite Memory Storage](knowledge/sqlite_memory_storage.md) (lines 18183-18201) - [Create SQLite memory database](knowledge/create_sqlite_memory_database.md) (lines 18201-18207) - [Create memory instance with SQLite backend](knowledge/create_memory_instance_with_sqlite_backend.md) (lines 18207-18210) - [Create agent with memory and SQLite storage](knowledge/create_agent_with_memory_and_sqlite_storage.md) (lines 18217-18229) - [Second interaction - testing if memory was stored](knowledge/second_interaction_-_testing_if_memory_was_stored.md) (lines 18237-18245) - [Mem0 Memory](knowledge/mem0_memory.md) (lines 18283-18304) - [Create a knowledge base of PDFs from URLs](knowledge/create_a_knowledge_base_of_pdfs_from_urls.md) (lines 21014-21102) - [Create a knowledge base of PDFs from URLs](knowledge/create_a_knowledge_base_of_pdfs_from_urls.md) (lines 21116-21127) - [Load the knowledge base: Comment after first run as the knowledge base is already loaded](knowledge/load_the_knowledge_base_comment_after_first_run_as_the_knowledge_base_is_already_loaded.md) (lines 21127-21175) - [Create a knowledge base of PDFs from URLs](knowledge/create_a_knowledge_base_of_pdfs_from_urls.md) (lines 21190-21201) - [Load the knowledge base: Comment after first run as the knowledge base is already loaded](knowledge/load_the_knowledge_base_comment_after_first_run_as_the_knowledge_base_is_already_loaded.md) (lines 21201-21263) - [Create a knowledge base of PDFs from URLs](knowledge/create_a_knowledge_base_of_pdfs_from_urls.md) (lines 21278-21290) - [Load the knowledge base: Comment after first run as the knowledge base is already loaded](knowledge/load_the_knowledge_base_comment_after_first_run_as_the_knowledge_base_is_already_loaded.md) (lines 21290-21339) - [Create a knowledge base from a PDF URL using LanceDb for vector storage and OllamaEmbedder for embedding](knowledge/create_a_knowledge_base_from_a_pdf_url_using_lancedb_for_vector_storage_and_ollamaembedder_for_embedding.md) (lines 21370-21376) - [Load the knowledge base without recreating it if it already exists in Vector LanceDB](knowledge/load_the_knowledge_base_without_recreating_it_if_it_already_exists_in_vector_lancedb.md) (lines 21376-21379) - [Initialize the Agent with various configurations including the knowledge base and storage](knowledge/initialize_the_agent_with_various_configurations_including_the_knowledge_base_and_storage.md) (lines 21383-21394) - [Create a knowledge base of PDFs from URLs](knowledge/create_a_knowledge_base_of_pdfs_from_urls.md) (lines 21444-21455) - [Load the knowledge base: Comment after first run as the knowledge base is already loaded](knowledge/load_the_knowledge_base_comment_after_first_run_as_the_knowledge_base_is_already_loaded.md) (lines 21455-21504) - [Create a knowledge base of PDFs from URLs](knowledge/create_a_knowledge_base_of_pdfs_from_urls.md) (lines 21519-21530) - [Load the knowledge base: Comment after first run as the knowledge base is already loaded](knowledge/load_the_knowledge_base_comment_after_first_run_as_the_knowledge_base_is_already_loaded.md) (lines 21530-21592) - [Team with Knowledge Tools](knowledge/team_with_knowledge_tools.md) (lines 22569-22699) - [Reasoning Agent with Knowledge Tools](knowledge/reasoning_agent_with_knowledge_tools.md) (lines 22970-22986) - [Create a knowledge base containing information from a URL](knowledge/create_a_knowledge_base_containing_information_from_a_url.md) (lines 22986-23052) - [In-memory database to store user shopping lists](knowledge/in-memory_database_to_store_user_shopping_lists.md) (lines 23451-23452) - [Mem0 Memory Tools](knowledge/mem0_memory_tools.md) (lines 26141-26213) - [Zep Memory Tools](knowledge/zep_memory_tools.md) (lines 26383-26396) - [Zep Async Memory Tools](knowledge/zep_async_memory_tools.md) (lines 26454-26536) - [Create knowledge base](knowledge/create_knowledge_base.md) (lines 32661-32669) - [Create knowledge base](knowledge/create_knowledge_base.md) (lines 33345-33352) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 34315-34343) - [Create a Recipe Expert Agent with knowledge of Thai recipes](knowledge/create_a_recipe_expert_agent_with_knowledge_of_thai_recipes.md) (lines 34343-34405) - [Comment out after the knowledge base is loaded](knowledge/comment_out_after_the_knowledge_base_is_loaded.md) (lines 34405-34435) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 37089-37170) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 37954-38030) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 38351-38427) - [Agent with Knowledge Base](knowledge/agent_with_knowledge_base.md) (lines 38743-38824) - [Agent with Knowledge Base](knowledge/agent_with_knowledge_base.md) (lines 39141-39223) - [Agent with Knowledge Base](knowledge/agent_with_knowledge_base.md) (lines 39537-39606) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 40118-40192) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 41540-41619) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 42302-42376) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 43665-43725) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 44244-44322) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 45452-45526) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 46055-46136) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 46757-46832) - [Agent with Knowledge](knowledge/agent_with_knowledge.md) (lines 47340-47414) - [Agent with Memory](knowledge/agent_with_memory.md) (lines 48505-48534) - [Image Agent with Memory](knowledge/image_agent_with_memory.md) (lines 49267-49325) - [Memory V2](knowledge/memory_v2.md) (lines 58755-58838) - [Use an embedder in a knowledge base](knowledge/use_an_embedder_in_a_knowledge_base.md) (lines 58887-58901) - [Agentic Knowledge Filters](knowledge/agentic_knowledge_filters.md) (lines 59086-59192) - [Knowledge Filters](knowledge/knowledge_filters.md) (lines 59197-59277) - [Manual Knowledge Filters](knowledge/manual_knowledge_filters.md) (lines 59282-59407) - [Load Agno documentation in a knowledge base](knowledge/load_agno_documentation_in_a_knowledge_base.md) (lines 60038-60039) - [ArXiv Knowledge Base](knowledge/arxiv_knowledge_base.md) (lines 60594-60659) - [Create a knowledge base with the ArXiv documents](knowledge/create_a_knowledge_base_with_the_arxiv_documents.md) (lines 60659-60664) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 60664-60697) - [Combined Knowledge Base](knowledge/combined_knowledge_base.md) (lines 60697-60794) - [CSV Knowledge Base](knowledge/csv_knowledge_base.md) (lines 60794-60863) - [Initialize the Agent with the knowledge_base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 60863-60892) - [CSV URL Knowledge Base](knowledge/csv_url_knowledge_base.md) (lines 60892-60960) - [Initialize the Agent with the knowledge_base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 60960-60991) - [This section loads the knowledge base. Skip if your knowledge base was populated elsewhere.](knowledge/this_section_loads_the_knowledge_base_skip_if_your_knowledge_base_was_populated_elsewhere.md) (lines 61015-61016) - [Load the knowledge base](knowledge/load_the_knowledge_base.md) (lines 61020-61026) - [Load the knowledge base](knowledge/load_the_knowledge_base.md) (lines 61026-61028) - [Knowledge base is now loaded](knowledge/knowledge_base_is_now_loaded.md) (lines 61028-61029) - [Knowledge base setup (same as synchronous example)](knowledge/knowledge_base_setup_same_as_synchronous_example.md) (lines 61097-61104) - [Document Knowledge Base](knowledge/document_knowledge_base.md) (lines 61168-61228) - [Docx Knowledge Base](knowledge/docx_knowledge_base.md) (lines 61228-61290) - [Create a knowledge base with the DOCX files from the data/docs directory](knowledge/create_a_knowledge_base_with_the_docx_files_from_the_data_docs_directory.md) (lines 61290-61300) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 61300-61332) - [Load PDF knowledge base using hybrid search](knowledge/load_pdf_knowledge_base_using_hybrid_search.md) (lines 61400-61406) - [What is Knowledge?](knowledge/what_is_knowledge.md) (lines 61419-61441) - [Create a knowledge base for the Agent](knowledge/create_a_knowledge_base_for_the_agent.md) (lines 61441-61444) - [Add information to the knowledge base](knowledge/add_information_to_the_knowledge_base.md) (lines 61444-61447) - [Add the knowledge base to the Agent and](knowledge/add_the_knowledge_base_to_the_agent_and.md) (lines 61447-61448) - [give it a tool to search the knowledge base as needed](knowledge/give_it_a_tool_to_search_the_knowledge_base_as_needed.md) (lines 61448-61509) - [Create a knowledge base with the PDFs from the data/pdfs directory](knowledge/create_a_knowledge_base_with_the_pdfs_from_the_data_pdfs_directory.md) (lines 61509-61516) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 61516-61552) - [JSON Knowledge Base](knowledge/json_knowledge_base.md) (lines 61552-61620) - [Initialize the Agent with the knowledge_base](knowledge/initialize_the_agent_with_the_knowledge_base.md) (lines 61620-61653) - [LangChain Knowledge Base](knowledge/langchain_knowledge_base.md) (lines 61653-61689) - [-*- Create a knowledge base from the vector store](knowledge/-_-_create_a_knowledge_base_from_the_vector_store.md) (lines 61694-61717) - [LightRAG Knowledge Base](knowledge/lightrag_knowledge_base.md) (lines 61717-61733) - [Create a knowledge base, loaded with documents from a URL](knowledge/create_a_knowledge_base_loaded_with_documents_from_a_url.md) (lines 61733-61740) - [Load the knowledge base with the documents from the local directory and the URL](knowledge/load_the_knowledge_base_with_the_documents_from_the_local_directory_and_the_url.md) (lines 61740-61743) - [Load the knowledge base with the text](knowledge/load_the_knowledge_base_with_the_text.md) (lines 61743-61753) - [Create an agent with the knowledge base and the retriever](knowledge/create_an_agent_with_the_knowledge_base_and_the_retriever.md) (lines 61753-61786) - [LlamaIndex Knowledge Base](knowledge/llamaindex_knowledge_base.md) (lines 61786-61844) - [Create a knowledge base from the vector store](knowledge/create_a_knowledge_base_from_the_vector_store.md) (lines 61844-61847) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 61847-61850) - [Markdown Knowledge Base](knowledge/markdown_knowledge_base.md) (lines 61868-61929) - [Initialize the MarkdownKnowledgeBase](knowledge/initialize_the_markdownknowledgebase.md) (lines 61929-61936) - [Initialize the Assistant with the knowledge_base](knowledge/initialize_the_assistant_with_the_knowledge_base.md) (lines 61936-61969) - [PDF Knowledge Base](knowledge/pdf_knowledge_base.md) (lines 61969-62034) - [Create a knowledge base with the PDFs from the data/pdfs directory](knowledge/create_a_knowledge_base_with_the_pdfs_from_the_data_pdfs_directory.md) (lines 62034-62041) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 62041-62070) - [PDF Bytes Knowledge Base](knowledge/pdf_bytes_knowledge_base.md) (lines 62070-62130) - [PDF URL Knowledge Base](knowledge/pdf_url_knowledge_base.md) (lines 62130-62195) - [Create a knowledge base with the PDFs from the data/pdfs directory](knowledge/create_a_knowledge_base_with_the_pdfs_from_the_data_pdfs_directory.md) (lines 62195-62202) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 62202-62231) - [S3 PDF Knowledge Base](knowledge/s3_pdf_knowledge_base.md) (lines 62231-62287) - [S3 Text Knowledge Base](knowledge/s3_text_knowledge_base.md) (lines 62287-62348) - [Create a knowledge base, loaded with documents from a URL](knowledge/create_a_knowledge_base_loaded_with_documents_from_a_url.md) (lines 62395-62452) - [Create a knowledge base, loaded with documents from a URL](knowledge/create_a_knowledge_base_loaded_with_documents_from_a_url.md) (lines 62452-62493) - [Text Knowledge Base](knowledge/text_knowledge_base.md) (lines 62493-62555) - [Initialize the TextKnowledgeBase](knowledge/initialize_the_textknowledgebase.md) (lines 62555-62562) - [Initialize the Assistant with the knowledge_base](knowledge/initialize_the_assistant_with_the_knowledge_base.md) (lines 62562-62595) - [Website Knowledge Base](knowledge/website_knowledge_base.md) (lines 62595-62665) - [Create a knowledge base with the seed URLs](knowledge/create_a_knowledge_base_with_the_seed_urls.md) (lines 62665-62674) - [Create an agent with the knowledge base](knowledge/create_an_agent_with_the_knowledge_base.md) (lines 62674-62702) - [Wikipedia KnowledgeBase](knowledge/wikipedia_knowledgebase.md) (lines 62702-62762) - [Youtube KnowledgeBase](knowledge/youtube_knowledgebase.md) (lines 62762-62869) - [What is Memory?](knowledge/what_is_memory.md) (lines 62869-62891) - [Database file for memory and storage](knowledge/database_file_for_memory_and_storage.md) (lines 62918-62921) - [Initialize memory.v2](knowledge/initialize_memory_v2.md) (lines 62921-62927) - [Reset the memory for this example](knowledge/reset_the_memory_for_this_example.md) (lines 62988-62991) - [User ID for the memory](knowledge/user_id_for_the_memory.md) (lines 62991-63000) - [Send a message to the agent that would require the memory to be used](knowledge/send_a_message_to_the_agent_that_would_require_the_memory_to_be_used.md) (lines 63000-63007) - [Send a message to the agent that checks the memory is working](knowledge/send_a_message_to_the_agent_that_checks_the_memory_is_working.md) (lines 63007-63010) - [Create a memory instance with persistent storage](knowledge/create_a_memory_instance_with_persistent_storage.md) (lines 63039-63052) - [Create a user memory manually](knowledge/create_a_user_memory_manually.md) (lines 63052-63070) - [Replace a user memory](knowledge/replace_a_user_memory.md) (lines 63070-63090) - [Delete a user memory](knowledge/delete_a_user_memory.md) (lines 63090-63139) - [Get the most recent memory](knowledge/get_the_most_recent_memory.md) (lines 63182-63190) - [Get the oldest memory](knowledge/get_the_oldest_memory.md) (lines 63190-63207) - [Initialize memory with a model for agentic search](knowledge/initialize_memory_with_a_model_for_agentic_search.md) (lines 63207-63223) - [Memory Storage](knowledge/memory_storage.md) (lines 63242-63247) - [Memory Storage](knowledge/memory_storage.md) (lines 63247-63272) - [Create a SQLite database for memory](knowledge/create_a_sqlite_database_for_memory.md) (lines 63272-63278) - [Initialize Memory with the storage backend](knowledge/initialize_memory_with_the_storage_backend.md) (lines 63278-63297) - [Create a SQLite memory database](knowledge/create_a_sqlite_memory_database.md) (lines 63297-63303) - [Initialize Memory with the storage backend](knowledge/initialize_memory_with_the_storage_backend.md) (lines 63303-63306) - [Add a user memory that will persist across restarts](knowledge/add_a_user_memory_that_will_persist_across_restarts.md) (lines 63306-63316) - [Create a PostgreSQL memory database](knowledge/create_a_postgresql_memory_database.md) (lines 63329-63335) - [Initialize Memory with the storage backend](knowledge/initialize_memory_with_the_storage_backend.md) (lines 63335-63338) - [Memory operations work the same regardless of the backend](knowledge/memory_operations_work_the_same_regardless_of_the_backend.md) (lines 63348-63363) - [Create memory storage](knowledge/create_memory_storage.md) (lines 63363-63370) - [Create agent with both memory and storage](knowledge/create_agent_with_both_memory_and_storage.md) (lines 63376-63390) - [Delete a specific memory](knowledge/delete_a_specific_memory.md) (lines 63390-63393) - [Replace/update a memory](knowledge/replace_update_a_memory.md) (lines 63393-63400) - [Agentic Chunking](knowledge/agentic_chunking.md) (lines 67274-67285) - [Document Chunking](knowledge/document_chunking.md) (lines 67285-67296) - [Fixed Size Chunking](knowledge/fixed_size_chunking.md) (lines 67296-67307) - [Recursive Chunking](knowledge/recursive_chunking.md) (lines 67307-67318) - [Semantic Chunking](knowledge/semantic_chunking.md) (lines 67318-67330) - [Arxiv Knowledge Base](knowledge/arxiv_knowledge_base.md) (lines 67590-67600) - [AgentKnowledge](knowledge/agentknowledge.md) (lines 67600-67614) - [Combined Knowledge Base](knowledge/combined_knowledge_base.md) (lines 67614-67624) - [CSV Knowledge Base](knowledge/csv_knowledge_base.md) (lines 67624-67634) - [CSV URL Knowledge Base](knowledge/csv_url_knowledge_base.md) (lines 67634-67644) - [Docx Knowledge Base](knowledge/docx_knowledge_base.md) (lines 67644-67654) - [JSON Knowledge Base](knowledge/json_knowledge_base.md) (lines 67654-67664) - [Langchain Knowledge Base](knowledge/langchain_knowledge_base.md) (lines 67664-67674) - [LlamaIndex Knowledge Base](knowledge/llamaindex_knowledge_base.md) (lines 67674-67684) - [PDF Knowledge Base](knowledge/pdf_knowledge_base.md) (lines 67684-67694) - [PDF URL Knowledge Base](knowledge/pdf_url_knowledge_base.md) (lines 67694-67704) - [Text Knowledge Base](knowledge/text_knowledge_base.md) (lines 67704-67714) - [Website Knowledge Base](knowledge/website_knowledge_base.md) (lines 67714-67724) - [Wikipedia Knowledge Base](knowledge/wikipedia_knowledge_base.md) (lines 67724-67734) - [YouTube Knowledge Base](knowledge/youtube_knowledge_base.md) (lines 67734-67744) - [Memory](knowledge/memory.md) (lines 67744-67754) - [MongoMemoryDb](knowledge/mongomemorydb.md) (lines 67754-67764) - [PostgresMemoryDb](knowledge/postgresmemorydb.md) (lines 67764-67774) - [RedisMemoryDb](knowledge/redismemorydb.md) (lines 67774-67784) - [SqliteMemoryDb](knowledge/sqlitememorydb.md) (lines 67784-67794) - [Create a memory instance with persistent storage](knowledge/create_a_memory_instance_with_persistent_storage.md) (lines 69623-69653) - [Initialize knowledge base](knowledge/initialize_knowledge_base.md) (lines 69658-69702) - [Create a memory instance with persistent storage](knowledge/create_a_memory_instance_with_persistent_storage.md) (lines 69702-69804) - [Knowledge Tools](knowledge/knowledge_tools.md) (lines 72526-72559) - [Create a knowledge base containing information from a URL](knowledge/create_a_knowledge_base_containing_information_from_a_url.md) (lines 72559-72609) - [Azure Cosmos DB MongoDB vCore Agent Knowledge](knowledge/azure_cosmos_db_mongodb_vcore_agent_knowledge.md) (lines 79903-79926) - [Cassandra Agent Knowledge](knowledge/cassandra_agent_knowledge.md) (lines 79963-79995) - [knowledge_base.load(recreate=False) # Comment out after first run](knowledge/knowledge_base_load_recreate_false_comment_out_after_first_run.md) (lines 80013-80099) - [ChromaDB Agent Knowledge](knowledge/chromadb_agent_knowledge.md) (lines 80099-80207) - [Clickhouse Agent Knowledge](knowledge/clickhouse_agent_knowledge.md) (lines 80207-80256) - [Couchbase Agent Knowledge](knowledge/couchbase_agent_knowledge.md) (lines 80317-80379) - [Load the knowledge base](knowledge/load_the_knowledge_base.md) (lines 80409-80412) - [LanceDB Agent Knowledge](knowledge/lancedb_agent_knowledge.md) (lines 80574-80597) - [Knowledge Base](knowledge/knowledge_base.md) (lines 80604-80691) - [Milvus Agent Knowledge](knowledge/milvus_agent_knowledge.md) (lines 80691-80722) - [Create knowledge base](knowledge/create_knowledge_base.md) (lines 80722-80730) - [MongoDB Agent Knowledge](knowledge/mongodb_agent_knowledge.md) (lines 80791-80813) - [knowledge_base.load(recreate=True) # Comment out after first run](knowledge/knowledge_base_load_recreate_true_comment_out_after_first_run.md) (lines 80831-80892) - [PgVector Agent Knowledge](knowledge/pgvector_agent_knowledge.md) (lines 80892-80924) - [Load the knowledge base: Comment out after first run](knowledge/load_the_knowledge_base_comment_out_after_first_run.md) (lines 80924-80988) - [Pinecone Agent Knowledge](knowledge/pinecone_agent_knowledge.md) (lines 80988-81131) - [Qdrant Agent Knowledge](knowledge/qdrant_agent_knowledge.md) (lines 81131-81252) - [SingleStore Agent Knowledge](knowledge/singlestore_agent_knowledge.md) (lines 81252-81356) - [SurrealDB Agent Knowledge](knowledge/surrealdb_agent_knowledge.md) (lines 81356-81388) - [Weaviate Agent Knowledge](knowledge/weaviate_agent_knowledge.md) (lines 81511-81557) - [Create knowledge base](knowledge/create_knowledge_base.md) (lines 81557-81564) ## Multimodal - [Multimodal Agents](multimodal/multimodal_agents.md) (lines 1200-1254) - [Multimodal Agents](multimodal/multimodal_agents.md) (lines 9183-9417) ## Tools - [Define tools to manage our shopping list](tools/define_tools_to_manage_our_shopping_list.md) (lines 1933-1966) - [Tools](tools/tools.md) (lines 2555-2672) - [Agent with Tools](tools/agent_with_tools.md) (lines 8119-8159) - [Image Generation Tools](tools/image_generation_tools.md) (lines 10816-10886) - [Team with Reasoning Tools](tools/team_with_reasoning_tools.md) (lines 22699-22831) - [Gemini with Reasoning Tools](tools/gemini_with_reasoning_tools.md) (lines 22831-22902) - [Gemini with Thinking Tools](tools/gemini_with_thinking_tools.md) (lines 22902-22970) - [Reasoning Agent with Reasoning Tools](tools/reasoning_agent_with_reasoning_tools.md) (lines 23052-23115) - [Reasoning Agent with Thinking Tools](tools/reasoning_agent_with_thinking_tools.md) (lines 23157-23227) - [Define tools to manage our shopping list](tools/define_tools_to_manage_our_shopping_list.md) (lines 23607-23640) - [Define tools to manage our shopping list](tools/define_tools_to_manage_our_shopping_list.md) (lines 23726-23806) - [CSV Tools](tools/csv_tools.md) (lines 26024-26089) - [DuckDB Tools](tools/duckdb_tools.md) (lines 26089-26141) - [Postgres Tools](tools/postgres_tools.md) (lines 26277-26333) - [SQL Tools](tools/sql_tools.md) (lines 26333-26383) - [Initialize the ZepTools](tools/initialize_the_zeptools.md) (lines 26396-26399) - [Docker Tools](tools/docker_tools.md) (lines 26597-26716) - [File Tools](tools/file_tools.md) (lines 26716-26766) - [Python Tools](tools/python_tools.md) (lines 26766-26816) - [Shell Tools](tools/shell_tools.md) (lines 26816-26866) - [Sleep Tools](tools/sleep_tools.md) (lines 26866-26916) - [Airflow Tools](tools/airflow_tools.md) (lines 27876-27908) - [Apify Tools](tools/apify_tools.md) (lines 27963-28010) - [AWS Lambda Tools](tools/aws_lambda_tools.md) (lines 28010-28066) - [AWS SES Tools](tools/aws_ses_tools.md) (lines 28066-28079) - [Cal.com Tools](tools/cal_com_tools.md) (lines 28175-28243) - [Composio Tools](tools/composio_tools.md) (lines 28243-28295) - [Confluence Tools](tools/confluence_tools.md) (lines 28295-28355) - [DALL-E Tools](tools/dall-e_tools.md) (lines 28355-28428) - [Daytona Tools](tools/daytona_tools.md) (lines 28428-28495) - [Desi Vocal Tools](tools/desi_vocal_tools.md) (lines 28495-28560) - [Fal Tools](tools/fal_tools.md) (lines 28664-28726) - [Financial Datasets Tools](tools/financial_datasets_tools.md) (lines 28726-28755) - [Giphy Tools](tools/giphy_tools.md) (lines 28802-28861) - [GitHub Tools](tools/github_tools.md) (lines 28861-28915) - [Google Calendar Tools](tools/google_calendar_tools.md) (lines 28915-28973) - [Google Maps Tools](tools/google_maps_tools.md) (lines 28973-28996) - [Jira Tools](tools/jira_tools.md) (lines 29093-29148) - [Linear Tools](tools/linear_tools.md) (lines 29148-29201) - [Luma Labs Tools](tools/luma_labs_tools.md) (lines 29201-29252) - [MLX Transcribe Tools](tools/mlx_transcribe_tools.md) (lines 29252-29302) - [Models Labs Tools](tools/models_labs_tools.md) (lines 29302-29353) - [OpenBB Tools](tools/openbb_tools.md) (lines 29353-29404) - [Replicate Tools](tools/replicate_tools.md) (lines 29404-29455) - [Resend Tools](tools/resend_tools.md) (lines 29455-29506) - [Todoist Tools](tools/todoist_tools.md) (lines 29506-29550) - [YFinance Tools](tools/yfinance_tools.md) (lines 29599-29649) - [YouTube Tools](tools/youtube_tools.md) (lines 29649-29700) - [Zendesk Tools](tools/zendesk_tools.md) (lines 29700-29758) - [ArXiv Tools](tools/arxiv_tools.md) (lines 29758-29804) - [Baidu Search Tools](tools/baidu_search_tools.md) (lines 29804-29859) - [Brave Search Tools](tools/brave_search_tools.md) (lines 29859-29912) - [Crawl4ai Tools](tools/crawl4ai_tools.md) (lines 29912-29958) - [Exa Tools](tools/exa_tools.md) (lines 30012-30062) - [Google Search Tools](tools/google_search_tools.md) (lines 30062-30114) - [Hacker News Tools](tools/hacker_news_tools.md) (lines 30114-30164) - [Linkup Tools](tools/linkup_tools.md) (lines 30164-30205) - [PubMed Tools](tools/pubmed_tools.md) (lines 30205-30255) - [SearxNG Tools](tools/searxng_tools.md) (lines 30255-30305) - [SerpAPI Tools](tools/serpapi_tools.md) (lines 30305-30356) - [Serper Tools](tools/serper_tools.md) (lines 30356-30437) - [Tavily Tools](tools/tavily_tools.md) (lines 30437-30488) - [Wikipedia Tools](tools/wikipedia_tools.md) (lines 30488-30538) - [Discord Tools](tools/discord_tools.md) (lines 30538-30616) - [Email Tools](tools/email_tools.md) (lines 30616-30683) - [Slack Tools](tools/slack_tools.md) (lines 30683-30739) - [Twilio Tools](tools/twilio_tools.md) (lines 30739-30797) - [Webex Tools](tools/webex_tools.md) (lines 30797-30822) - [WhatsApp Tools](tools/whatsapp_tools.md) (lines 30873-30891) - [X (Twitter) Tools](tools/x_twitter_tools.md) (lines 30937-30997) - [BrightData Tools](tools/brightdata_tools.md) (lines 30997-31021) - [Firecrawl Tools](tools/firecrawl_tools.md) (lines 31059-31110) - [Jina Reader Tools](tools/jina_reader_tools.md) (lines 31110-31160) - [Newspaper Tools](tools/newspaper_tools.md) (lines 31160-31210) - [Newspaper4k Tools](tools/newspaper4k_tools.md) (lines 31210-31260) - [Oxylabs Tools](tools/oxylabs_tools.md) (lines 31260-31310) - [Spider Tools](tools/spider_tools.md) (lines 31310-31360) - [Website Tools](tools/website_tools.md) (lines 31360-31410) - [Accuracy with Tools](tools/accuracy_with_tools.md) (lines 33470-33502) - [Reliability with Multiple Tools](tools/reliability_with_multiple_tools.md) (lines 33747-33785) - [Agent with Tools](tools/agent_with_tools.md) (lines 34599-34623) - [Custom Tools](tools/custom_tools.md) (lines 34869-34923) - [Agent with Tools](tools/agent_with_tools.md) (lines 37603-37655) - [Async Agent with Tools](tools/async_agent_with_tools.md) (lines 37713-37767) - [Agent with Tools](tools/agent_with_tools.md) (lines 38183-38237) - [Agent with Tools](tools/agent_with_tools.md) (lines 38580-38634) - [Agent with Tools](tools/agent_with_tools.md) (lines 38976-39030) - [Agent with Tools](tools/agent_with_tools.md) (lines 39377-39432) - [Agent with Tools](tools/agent_with_tools.md) (lines 39741-39795) - [Agent with Tools](tools/agent_with_tools.md) (lines 39901-39919) - [Agent with Tools](tools/agent_with_tools.md) (lines 40342-40395) - [Agent with Tools](tools/agent_with_tools.md) (lines 40587-40640) - [Agent with Tools](tools/agent_with_tools.md) (lines 40824-40877) - [Agent with Tools](tools/agent_with_tools.md) (lines 41069-41121) - [Agent with Tools](tools/agent_with_tools.md) (lines 41883-41935) - [Agent with Tools](tools/agent_with_tools.md) (lines 42521-42581) - [Agent with Tools](tools/agent_with_tools.md) (lines 43322-43375) - [Finance Agent with Tools](tools/finance_agent_with_tools.md) (lines 43494-43523) - [Agent with Tools](tools/agent_with_tools.md) (lines 43865-43886) - [Agent with Tools](tools/agent_with_tools.md) (lines 44032-44092) - [Agent with Tools](tools/agent_with_tools.md) (lines 44467-44518) - [Agent with Tools](tools/agent_with_tools.md) (lines 45291-45346) - [Agent with Tools](tools/agent_with_tools.md) (lines 45675-45728) - [Agent with Tools](tools/agent_with_tools.md) (lines 45833-45886) - [Agent with Tools](tools/agent_with_tools.md) (lines 46351-46405) - [Agent with Tools](tools/agent_with_tools.md) (lines 47047-47099) - [Agent with Tools](tools/agent_with_tools.md) (lines 47701-47753) - [Agent with Tools](tools/agent_with_tools.md) (lines 48048-48100) - [Agent with Tools](tools/agent_with_tools.md) (lines 48206-48258) - [Async Agent with Tools](tools/async_agent_with_tools.md) (lines 48352-48407) - [Agent with Tools](tools/agent_with_tools.md) (lines 48738-48791) - [Async Agent with Tools](tools/async_agent_with_tools.md) (lines 48791-48841) - [Agent with Tools](tools/agent_with_tools.md) (lines 49561-49607) - ["[Bug] Async tools in team of agents not awaited properly, causing runtime errors ",](tools/bug_async_tools_in_team_of_agents_not_awaited_properly_causing_runtime_errors.md) (lines 51029-51030) - [Define tools to manage our shopping list](tools/define_tools_to_manage_our_shopping_list.md) (lines 51227-51279) - [--- Simulation tools ---](tools/---_simulation_tools_---.md) (lines 57768-57801) - [Reasoning Tools](tools/reasoning_tools.md) (lines 66967-67096) - [Ollama Tools](tools/ollama_tools.md) (lines 68008-68018) - [Define tools that work with shared team state](tools/define_tools_that_work_with_shared_team_state.md) (lines 70598-70620) - [Define team-level tools](tools/define_team-level_tools.md) (lines 70629-70652) - [Async Tools](tools/async_tools.md) (lines 70988-71083) - [Updating Tools](tools/updating_tools.md) (lines 71083-71125) - [What are Tools?](tools/what_are_tools.md) (lines 71648-71720) - [Reasoning Tools](tools/reasoning_tools.md) (lines 72609-72743) - [Thinking Tools](tools/thinking_tools.md) (lines 72743-72856) - [Selecting tools](tools/selecting_tools.md) (lines 72856-72910) - [Writing your own tools](tools/writing_your_own_tools.md) (lines 72910-73046) - [Initialize the Agent with Mem0Tools](tools/initialize_the_agent_with_mem0tools.md) (lines 73260-73282) - [Create an agent with PandasTools](tools/create_an_agent_with_pandastools.md) (lines 73362-73365) - [Initialize PostgresTools with connection details](tools/initialize_postgrestools_with_connection_details.md) (lines 73435-73444) - [Create an agent with the PostgresTools](tools/create_an_agent_with_the_postgrestools.md) (lines 73444-73447) - [Initialize the ZepTools](tools/initialize_the_zeptools.md) (lines 73610-73613) - [Create an agent with ApifyTools](tools/create_an_agent_with_apifytools.md) (lines 74539-74551) - [MoviePy Video Tools](tools/moviepy_video_tools.md) (lines 76483-76570) - [Create an agent with OpenWeatherTools](tools/create_an_agent_with_openweathertools.md) (lines 76647-76657) - [Web Browser Tools](tools/web_browser_tools.md) (lines 77005-77038) - [Initialize Zoom tools with credentials](tools/initialize_zoom_tools_with_credentials.md) (lines 78727-78734) - [Website Tools](tools/website_tools.md) (lines 79857-79903) - [Define tools to manage a shopping list in workflow session state](tools/define_tools_to_manage_a_shopping_list_in_workflow_session_state.md) (lines 85055-85126) - [Create agents with tools that use workflow session state](tools/create_agents_with_tools_that_use_workflow_session_state.md) (lines 85126-85151) ## Metrics - [Metrics](metrics/metrics.md) (lines 1099-1147) - [Print metrics per message](metrics/print_metrics_per_message.md) (lines 1147-1159) - [Print the aggregated metrics for the whole run](metrics/print_the_aggregated_metrics_for_the_whole_run.md) (lines 1159-1162) - [Print the aggregated metrics for the whole session](metrics/print_the_aggregated_metrics_for_the_whole_session.md) (lines 1162-1200) - [Agent Extra Metrics](metrics/agent_extra_metrics.md) (lines 20102-20117) - [Showing input audio, output audio and total audio tokens metrics](metrics/showing_input_audio_output_audio_and_total_audio_tokens_metrics.md) (lines 20136-20152) - [Showing reasoning tokens metrics](metrics/showing_reasoning_tokens_metrics.md) (lines 20152-20161) - [Showing cached tokens metrics](metrics/showing_cached_tokens_metrics.md) (lines 20161-20196) - [Agent Metrics](metrics/agent_metrics.md) (lines 20196-20225) - [Print metrics per message](metrics/print_metrics_per_message.md) (lines 20225-20237) - [Print the metrics](metrics/print_the_metrics.md) (lines 20237-20240) - [Print the session metrics](metrics/print_the_session_metrics.md) (lines 20240-20276) - [Metrics](metrics/metrics.md) (lines 69930-69972) - [Print team leader message metrics](metrics/print_team_leader_message_metrics.md) (lines 69994-70007) - [Print aggregated team leader metrics](metrics/print_aggregated_team_leader_metrics.md) (lines 70007-70011) - [Print team leader session metrics](metrics/print_team_leader_session_metrics.md) (lines 70011-70015) - [Print team member message metrics](metrics/print_team_member_message_metrics.md) (lines 70015-70030) - [Print full team session metrics (including all members)](metrics/print_full_team_session_metrics_including_all_members.md) (lines 70030-70102) - [Access metrics for a specific member](metrics/access_metrics_for_a_specific_member.md) (lines 70102-70131) ## Advanced - [Agent Context](advanced/agent_context.md) (lines 190-238) - [Create a Context-Aware Agent that can access real-time HackerNews data](advanced/create_a_context-aware_agent_that_can_access_real-time_hackernews_data.md) (lines 238-257) - [Create a Context-Aware Agent that can access real-time HackerNews data](advanced/create_a_context-aware_agent_that_can_access_real-time_hackernews_data.md) (lines 303-314) - [Prompts](advanced/prompts.md) (lines 1388-1513) - [Agent that uses structured outputs with streaming](advanced/agent_that_uses_structured_outputs_with_streaming.md) (lines 2462-2481) - [More example prompts to explore:](advanced/more_example_prompts_to_explore.md) (lines 6627-6681) - [More example prompts to explore:](advanced/more_example_prompts_to_explore.md) (lines 6787-6836) - [More example prompts to explore:](advanced/more_example_prompts_to_explore.md) (lines 6935-6989) - [More example prompts to explore:](advanced/more_example_prompts_to_explore.md) (lines 7087-7153) - [More example prompts to explore:](advanced/more_example_prompts_to_explore.md) (lines 7810-7864) - [More example prompts to explore:](advanced/more_example_prompts_to_explore.md) (lines 8066-8119) - [Example prompts to explore:](advanced/example_prompts_to_explore.md) (lines 9125-9183) - [Add Context](advanced/add_context.md) (lines 11439-11483) - [Create a Context-Aware Agent that can access real-time HackerNews data](advanced/create_a_context-aware_agent_that_can_access_real-time_hackernews_data.md) (lines 11483-11494) - [Agent Context](advanced/agent_context.md) (lines 11526-11568) - [Create a Context-Aware Agent that can access real-time HackerNews data](advanced/create_a_context-aware_agent_that_can_access_real-time_hackernews_data.md) (lines 11568-11586) - [Context In Instructions](advanced/context_in_instructions.md) (lines 11618-11641) - [Audio Streaming Agent](advanced/audio_streaming_agent.md) (lines 18553-18567) - [Question with streaming](advanced/question_with_streaming.md) (lines 18585-18640) - [Run the agent (non-streaming)](advanced/run_the_agent_non-streaming.md) (lines 21675-21679) - [Run the agent (non-streaming)](advanced/run_the_agent_non-streaming.md) (lines 21696-21702) - [Create a fresh agent for streaming](advanced/create_a_fresh_agent_for_streaming.md) (lines 21712-21719) - [Print response (which includes processing streaming responses)](advanced/print_response_which_includes_processing_streaming_responses.md) (lines 21719-21727) - [Access reasoning_content from the agent's run_response after streaming](advanced/access_reasoning_content_from_the_agent_s_run_response_after_streaming.md) (lines 21727-21742) - [Create a fresh agent with reasoning_model for streaming](advanced/create_a_fresh_agent_with_reasoning_model_for_streaming.md) (lines 21742-21749) - [Print response (which includes processing streaming responses)](advanced/print_response_which_includes_processing_streaming_responses.md) (lines 21749-21757) - [Access reasoning_content from the agent's run_response after streaming](advanced/access_reasoning_content_from_the_agent_s_run_response_after_streaming.md) (lines 21757-21802) - [Refresh the context](advanced/refresh_the_context.md) (lines 26415-26418) - [More example prompts to explore:](advanced/more_example_prompts_to_explore.md) (lines 27038-27064) - [Example prompts:](advanced/example_prompts.md) (lines 27587-27614) - [Example prompts to try:](advanced/example_prompts_to_try.md) (lines 27696-27709) - [Streaming User Confirmation](advanced/streaming_user_confirmation.md) (lines 31644-31768) - [Streaming User Input](advanced/streaming_user_input.md) (lines 31974-32078) - [Streaming External Tool Execution](advanced/streaming_external_tool_execution.md) (lines 32397-32412) - [Agent Context](advanced/agent_context.md) (lines 33838-33889) - [Create a Context-Aware Agent that can access real-time HackerNews data](advanced/create_a_context-aware_agent_that_can_access_real-time_hackernews_data.md) (lines 33889-33907) - [More example prompts to try:](advanced/more_example_prompts_to_try.md) (lines 34125-34157) - [More example prompts to try:](advanced/more_example_prompts_to_try.md) (lines 34285-34315) - [More example prompts to try:](advanced/more_example_prompts_to_try.md) (lines 34660-34690) - [More example prompts to try:](advanced/more_example_prompts_to_try.md) (lines 34839-34869) - [More example prompts to try:](advanced/more_example_prompts_to_try.md) (lines 35296-35326) - [More example prompts to try:](advanced/more_example_prompts_to_try.md) (lines 35548-35578) - [More example prompts to try:](advanced/more_example_prompts_to_try.md) (lines 36530-36566) - [Streaming Agent](advanced/streaming_agent.md) (lines 36773-36787) - [Async Streaming Agent](advanced/async_streaming_agent.md) (lines 37655-37671) - [Streaming Agent](advanced/streaming_agent.md) (lines 37823-37839) - [Streaming Agent](advanced/streaming_agent.md) (lines 38293-38309) - [Basic Streaming](advanced/basic_streaming.md) (lines 38687-38702) - [Basic Streaming](advanced/basic_streaming.md) (lines 39084-39099) - [Streaming Agent](advanced/streaming_agent.md) (lines 39484-39537) - [Streaming Agent](advanced/streaming_agent.md) (lines 39848-39866) - [Print the response in the terminal (streaming)](advanced/print_the_response_in_the_terminal_streaming.md) (lines 39866-39901) - [Streaming Agent](advanced/streaming_agent.md) (lines 40006-40019) - [Streaming Agent](advanced/streaming_agent.md) (lines 40450-40468) - [Streaming Agent](advanced/streaming_agent.md) (lines 40692-40705) - [Streaming Agent](advanced/streaming_agent.md) (lines 40932-40949) - [Streaming Agent](advanced/streaming_agent.md) (lines 41372-41386) - [Streaming Agent](advanced/streaming_agent.md) (lines 42194-42208) - [Streaming Agent](advanced/streaming_agent.md) (lines 42633-42686) - [Async Streaming Agent](advanced/async_streaming_agent.md) (lines 42798-42815) - [Streaming Basic Agent](advanced/streaming_basic_agent.md) (lines 42969-42983) - [Streaming Agent](advanced/streaming_agent.md) (lines 43435-43453) - [Streaming Agent](advanced/streaming_agent.md) (lines 43612-43665) - [Streaming Agent](advanced/streaming_agent.md) (lines 43979-44032) - [Streaming Agent](advanced/streaming_agent.md) (lines 44139-44153) - [Streaming Agent](advanced/streaming_agent.md) (lines 45141-45162) - [Streaming Agent](advanced/streaming_agent.md) (lines 45399-45417) - [Streaming Agent](advanced/streaming_agent.md) (lines 45780-45793) - [Streaming Agent](advanced/streaming_agent.md) (lines 45940-45954) - [Streaming Agent](advanced/streaming_agent.md) (lines 46581-46595) - [Streaming Agent](advanced/streaming_agent.md) (lines 47153-47167) - [Streaming Agent](advanced/streaming_agent.md) (lines 47805-47818) - [Streaming Agent](advanced/streaming_agent.md) (lines 47912-47928) - [Streaming Agent](advanced/streaming_agent.md) (lines 48152-48166) - [Async Agent with Streaming](advanced/async_agent_with_streaming.md) (lines 48305-48352) - [Agent with Streaming](advanced/agent_with_streaming.md) (lines 48407-48455) - [Async Streaming Agent](advanced/async_streaming_agent.md) (lines 48935-48951) - [Streaming Agent](advanced/streaming_agent.md) (lines 48985-48999) - [Live Search Streaming Agent](advanced/live_search_streaming_agent.md) (lines 49375-49427) - [Streaming responses](advanced/streaming_responses.md) (lines 70368-70372) - [Asynchronous streaming](advanced/asynchronous_streaming.md) (lines 70372-70384) - [Create streaming team](advanced/create_streaming_team.md) (lines 70867-70878) - [Model Context Protocol (MCP)](advanced/model_context_protocol_mcp.md) (lines 71881-71992) - [Refresh the context](advanced/refresh_the_context.md) (lines 73629-73632) ## Miscellaneous - [Example usage](misc/example_usage.md) (lines 257-303) - [Example usage](misc/example_usage.md) (lines 314-322) - [UserId for the memories](misc/userid_for_the_memories.md) (lines 778-780) - [Initialize storage](misc/initialize_storage.md) (lines 789-792) - [Initialize Agent](misc/initialize_agent.md) (lines 792-1099) - [Fetch the audio file and convert it to a base64 encoded string](misc/fetch_the_audio_file_and_convert_it_to_a_base64_encoded_string.md) (lines 1254-1285) - [Please download "GreatRedSpot.mp4" using](misc/please_download_greatredspot_mp4_using.md) (lines 1285-1286) - [wget https://storage.googleapis.com/generativeai-downloads/images/GreatRedSpot.mp4](misc/wget_https_storage_googleapis_com_generativeai-downloads_images_greatredspot_mp4.md) (lines 1286-1342) - [Save the response audio to a file](misc/save_the_response_audio_to_a_file.md) (lines 1342-1364) - [Fetch the audio file and convert it to a base64 encoded string](misc/fetch_the_audio_file_and_convert_it_to_a_base64_encoded_string.md) (lines 1364-1388) - [Run agent and return the response as a variable](misc/run_agent_and_return_the_response_as_a_variable.md) (lines 1534-1537) - [Print the response in markdown format](misc/print_the_response_in_markdown_format.md) (lines 1537-1569) - [Run agent and return the response as a stream](misc/run_agent_and_return_the_response_as_a_stream.md) (lines 1569-1575) - [Print the response stream in markdown format](misc/print_the_response_stream_in_markdown_format.md) (lines 1575-1584) - [Stream with intermediate steps](misc/stream_with_intermediate_steps.md) (lines 1584-1704) - [Define agents](misc/define_agents.md) (lines 1704-1723) - [Run agent and return the response as a variable](misc/run_agent_and_return_the_response_as_a_variable.md) (lines 1751-1786) - [Start the session with user 1](misc/start_the_session_with_user_1.md) (lines 1786-1792) - [Continue the session with user 1](misc/continue_the_session_with_user_1.md) (lines 1792-1795) - [Start the session with user 2](misc/start_the_session_with_user_2.md) (lines 1795-1797) - [Continue the session with user 2](misc/continue_the_session_with_user_2.md) (lines 1797-1800) - [Ask the agent to give a summary of the conversation, this will use the history from the previous messages](misc/ask_the_agent_to_give_a_summary_of_the_conversation_this_will_use_the_history_from_the_previous_messages.md) (lines 1800-1815) - [Remove the tmp db file before running the script](misc/remove_the_tmp_db_file_before_running_the_script.md) (lines 1815-1859) - [Agent State](misc/agent_state.md) (lines 1859-1887) - [Define a tool that adds an item to the shopping list](misc/define_a_tool_that_adds_an_item_to_the_shopping_list.md) (lines 1887-1894) - [Create an Agent that maintains state](misc/create_an_agent_that_maintains_state.md) (lines 1894-1907) - [Example usage](misc/example_usage.md) (lines 1907-1933) - [Create a Shopping List Manager Agent that maintains state](misc/create_a_shopping_list_manager_agent_that_maintains_state.md) (lines 1966-1985) - [Example usage](misc/example_usage.md) (lines 1985-2051) - [Sets the session state for the session with the id "user_1_session_1"](misc/sets_the_session_state_for_the_session_with_the_id_user_1_session_1.md) (lines 2051-2054) - [Will load the session state from the session with the id "user_1_session_1"](misc/will_load_the_session_state_from_the_session_with_the_id_user_1_session_1.md) (lines 2054-2057) - [Sets the session state for the session with the id "user_2_session_1"](misc/sets_the_session_state_for_the_session_with_the_id_user_2_session_1.md) (lines 2057-2060) - [Will load the session state from the session with the id "user_2_session_1"](misc/will_load_the_session_state_from_the_session_with_the_id_user_2_session_1.md) (lines 2060-2078) - [Define a tool that adds an item to the shopping list](misc/define_a_tool_that_adds_an_item_to_the_shopping_list.md) (lines 2078-2104) - [Example usage](misc/example_usage.md) (lines 2104-2112) - [Session Storage](misc/session_storage.md) (lines 2112-2332) - [Structured Output](misc/structured_output.md) (lines 2332-2360) - [Agent that uses JSON mode](misc/agent_that_uses_json_mode.md) (lines 2360-2369) - [Agent that uses structured outputs](misc/agent_that_uses_structured_outputs.md) (lines 2369-2390) - [Using JSON mode](misc/using_json_mode.md) (lines 2390-2400) - [Use the structured output](misc/use_the_structured_output.md) (lines 2400-2462) - [Agent Teams [Deprecated]](misc/agent_teams_deprecated.md) (lines 2481-2555) - [User Control Flows](misc/user_control_flows.md) (lines 2672-2739) - [Run the agent](misc/run_the_agent.md) (lines 2739-2742) - [Handle confirmation](misc/handle_confirmation.md) (lines 2742-2800) - [We still provide a docstring to the tool; This will be used to populate the `user_input_schema`](misc/we_still_provide_a_docstring_to_the_tool_this_will_be_used_to_populate_the_user_input_schema.md) (lines 2800-2853) - [You can either specify the user_input_fields leave empty for all fields to be provided by the user](misc/you_can_either_specify_the_user_input_fields_leave_empty_for_all_fields_to_be_provided_by_the_user.md) (lines 2853-2911) - [Example toolkit for handling emails](misc/example_toolkit_for_handling_emails.md) (lines 2911-2960) - [We use a while loop to continue the running until the agent is satisfied with the user input](misc/we_use_a_while_loop_to_continue_the_running_until_the_agent_is_satisfied_with_the_user_input.md) (lines 2960-3000) - [We have to create a tool with the correct name, arguments and docstring for the agent to know what to call.](misc/we_have_to_create_a_tool_with_the_correct_name_arguments_and_docstring_for_the_agent_to_know_what_to_call.md) (lines 3000-3052) - [AG-UI App](misc/ag-ui_app.md) (lines 3052-3150) - [Discord Bot](misc/discord_bot.md) (lines 3150-3273) - [FastAPI App](misc/fastapi_app.md) (lines 3273-3299) - [Async router by default (use_async=True)](misc/async_router_by_default_use_async_true.md) (lines 3299-3309) - [For synchronous router:](misc/for_synchronous_router.md) (lines 3309-3310) - [app = fastapi_app.get_app(use_async=False)](misc/app_fastapi_app_get_app_use_async_false.md) (lines 3310-3393) - [Playground App](misc/playground_app.md) (lines 3393-3504) - [Slack App](misc/slack_app.md) (lines 3504-3603) - [Whatsapp App](misc/whatsapp_app.md) (lines 3603-3631) - [Async router by default (use_async=True)](misc/async_router_by_default_use_async_true.md) (lines 3631-3715) - [Product updates](misc/product_updates.md) (lines 3715-5316) - [AWS Bedrock Embedder](misc/aws_bedrock_embedder.md) (lines 5500-5507) - [Setup](misc/setup.md) (lines 5507-5535) - [Usage](misc/usage.md) (lines 5535-5539) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 5539-5543) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 5543-5547) - [Params](misc/params.md) (lines 5562-5579) - [Developer Resources](misc/developer_resources.md) (lines 5579-5584) - [Azure OpenAI Embedder](misc/azure_openai_embedder.md) (lines 5584-5622) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 5622-5625) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 5625-5629) - [Cohere Embedder](misc/cohere_embedder.md) (lines 5665-5681) - [Add embedding to database](misc/add_embedding_to_database.md) (lines 5681-5683) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 5683-5687) - [Fireworks Embedder](misc/fireworks_embedder.md) (lines 5715-5729) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 5729-5732) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 5732-5736) - [Gemini Embedder](misc/gemini_embedder.md) (lines 5761-5775) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 5775-5778) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 5778-5782) - [HuggingFace Embedder](misc/huggingface_embedder.md) (lines 5811-5825) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 5825-5828) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 5828-5832) - [What are Embedders?](misc/what_are_embedders.md) (lines 5858-5870) - [Jina Embedder](misc/jina_embedder.md) (lines 5903-5939) - [Basic usage - automatically loads from JINA_API_KEY environment variable](misc/basic_usage_-_automatically_loads_from_jina_api_key_environment_variable.md) (lines 5939-5944) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 5944-5948) - [Get embedding with usage information](misc/get_embedding_with_usage_information.md) (lines 5955-5963) - [Mistral Embedder](misc/mistral_embedder.md) (lines 5997-6011) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 6011-6014) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 6014-6018) - [Ollama Embedder](misc/ollama_embedder.md) (lines 6048-6064) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 6064-6067) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 6067-6071) - [OpenAI Embedder](misc/openai_embedder.md) (lines 6099-6113) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 6113-6116) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 6116-6120) - [Qdrant FastEmbed Embedder](misc/qdrant_fastembed_embedder.md) (lines 6151-6165) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 6165-6168) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 6168-6172) - [SentenceTransformers Embedder](misc/sentencetransformers_embedder.md) (lines 6195-6209) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 6209-6212) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 6212-6216) - [Together Embedder](misc/together_embedder.md) (lines 6240-6254) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 6254-6257) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 6257-6261) - [Voyage AI Embedder](misc/voyage_ai_embedder.md) (lines 6286-6300) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 6300-6303) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 6303-6307) - [Simple Agent Evals](misc/simple_agent_evals.md) (lines 6337-6479) - [Evals on the Agno platform](misc/evals_on_the_agno_platform.md) (lines 6479-6541) - [Books Recommender](misc/books_recommender.md) (lines 6541-6621) - [Example usage with different types of book queries](misc/example_usage_with_different_types_of_book_queries.md) (lines 6621-6627) - [Finance Agent](misc/finance_agent.md) (lines 6681-6758) - [Example usage with detailed market analysis request](misc/example_usage_with_detailed_market_analysis_request.md) (lines 6758-6763) - [Semiconductor market analysis example](misc/semiconductor_market_analysis_example.md) (lines 6763-6775) - [Automotive market analysis example](misc/automotive_market_analysis_example.md) (lines 6775-6787) - [Movie Recommender](misc/movie_recommender.md) (lines 6836-6928) - [Example usage with different types of movie queries](misc/example_usage_with_different_types_of_movie_queries.md) (lines 6928-6935) - [Recipe Creator](misc/recipe_creator.md) (lines 6989-7081) - [Example usage with different types of recipe queries](misc/example_usage_with_different_types_of_recipe_queries.md) (lines 7081-7087) - [Research Agent](misc/research_agent.md) (lines 7153-7189) - [Initialize the research agent with advanced journalistic capabilities](misc/initialize_the_research_agent_with_advanced_journalistic_capabilities.md) (lines 7189-7277) - [Example usage with detailed research request](misc/example_usage_with_detailed_research_request.md) (lines 7277-7284) - [Advanced research topics to explore:](misc/advanced_research_topics_to_explore.md) (lines 7284-7337) - [Research Agent using Exa](misc/research_agent_using_exa.md) (lines 7337-7373) - [Initialize the academic research agent with scholarly capabilities](misc/initialize_the_academic_research_agent_with_scholarly_capabilities.md) (lines 7373-7469) - [Example usage with academic research request](misc/example_usage_with_academic_research_request.md) (lines 7469-7476) - [Advanced research topics to explore:](misc/advanced_research_topics_to_explore.md) (lines 7476-7530) - [Startup Analyst Agent](misc/startup_analyst_agent.md) (lines 7530-7650) - [Teaching Assistant](misc/teaching_assistant.md) (lines 7650-7658) - [Travel Agent](misc/travel_agent.md) (lines 7658-7801) - [Example usage with different types of travel queries](misc/example_usage_with_different_types_of_travel_queries.md) (lines 7801-7810) - [Tweet Analysis Agent](misc/tweet_analysis_agent.md) (lines 7864-7988) - [Youtube Agent](misc/youtube_agent.md) (lines 7988-8060) - [Example usage with different types of videos](misc/example_usage_with_different_types_of_videos.md) (lines 8060-8066) - [Basic](misc/basic.md) (lines 8159-8197) - [Research Team](misc/research_team.md) (lines 8197-8255) - [Agent with Media](misc/agent_with_media.md) (lines 8255-8315) - [Basic](misc/basic.md) (lines 8411-8468) - [Basic](misc/basic.md) (lines 8468-8534) - [Study Friend](misc/study_friend.md) (lines 8534-8672) - [Agno Assist](misc/agno_assist.md) (lines 8672-8693) - [Setup paths](misc/setup_paths.md) (lines 8693-8773) - [Create the agent](misc/create_the_agent.md) (lines 8784-8833) - [Create and configure the playground app](misc/create_and_configure_the_playground_app.md) (lines 8833-8882) - [Audio Conversation Agent](misc/audio_conversation_agent.md) (lines 8882-8957) - [Basic](misc/basic.md) (lines 8957-9049) - [Mcp Demo](misc/mcp_demo.md) (lines 9049-9069) - [Allow nested event loops](misc/allow_nested_event_loops.md) (lines 9069-9125) - [Ollama Agents](misc/ollama_agents.md) (lines 9417-9564) - [Reasoning Demo](misc/reasoning_demo.md) (lines 9564-9813) - [reasoning_tool_agent](misc/reasoning_tool_agent.md) (lines 9813-9814) - [Solve this logic puzzle: A man has to take a fox, a chicken, and a sack of grain across a river.](misc/solve_this_logic_puzzle_a_man_has_to_take_a_fox_a_chicken_and_a_sack_of_grain_across_a_river.md) (lines 9814-9815) - [The boat is only big enough for the man and one item. If left unattended together,](misc/the_boat_is_only_big_enough_for_the_man_and_one_item_if_left_unattended_together.md) (lines 9815-9816) - [the fox will eat the chicken, and the chicken will eat the grain. How can the man get everything across safely?](misc/the_fox_will_eat_the_chicken_and_the_chicken_will_eat_the_grain_how_can_the_man_get_everything_across_safely.md) (lines 9816-9819) - [What does Paul Graham explain here with respect to need to read?](misc/what_does_paul_graham_explain_here_with_respect_to_need_to_read.md) (lines 9820-9855) - [Teams Demo](misc/teams_demo.md) (lines 9855-10130) - [Upload Files](misc/upload_files.md) (lines 10130-10264) - [Basic](misc/basic.md) (lines 10368-10432) - [Reasoning Agent](misc/reasoning_agent.md) (lines 10432-10508) - [Agent With Media](misc/agent_with_media.md) (lines 10508-10576) - [Basic](misc/basic.md) (lines 10683-10749) - [Image Generation Model](misc/image_generation_model.md) (lines 10749-10816) - [Reasoning Agent](misc/reasoning_agent.md) (lines 10886-10964) - [Study Friend](misc/study_friend.md) (lines 10964-11116) - [Basic Async](misc/basic_async.md) (lines 11116-11135) - [-*- Print a response to the cli](misc/-_-_print_a_response_to_the_cli.md) (lines 11135-11164) - [Data Analyst](misc/data_analyst.md) (lines 11164-11227) - [Gather Multiple Agents](misc/gather_multiple_agents.md) (lines 11227-11303) - [Reasoning Agent](misc/reasoning_agent.md) (lines 11303-11359) - [Structured Outputs](misc/structured_outputs.md) (lines 11359-11395) - [Agent that uses JSON mode](misc/agent_that_uses_json_mode.md) (lines 11395-11402) - [Agent that uses structured outputs](misc/agent_that_uses_structured_outputs.md) (lines 11402-11439) - [Example usage](misc/example_usage.md) (lines 11494-11526) - [Example usage](misc/example_usage.md) (lines 11586-11618) - [Create an Agent that has access to real-time SpaceX data](misc/create_an_agent_that_has_access_to_real-time_spacex_data.md) (lines 11641-11689) - [Azure OpenAI Embedder](misc/azure_openai_embedder.md) (lines 11689-11705) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 11705-11709) - [Example usage:](misc/example_usage.md) (lines 11709-11767) - [Cohere Embedder](misc/cohere_embedder.md) (lines 11767-11782) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 11782-11786) - [Example usage:](misc/example_usage.md) (lines 11786-11842) - [Fireworks Embedder](misc/fireworks_embedder.md) (lines 11842-11858) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 11858-11862) - [Example usage:](misc/example_usage.md) (lines 11862-11918) - [Gemini Embedder](misc/gemini_embedder.md) (lines 11918-11934) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 11934-11938) - [Example usage:](misc/example_usage.md) (lines 11938-11994) - [Huggingface Embedder](misc/huggingface_embedder.md) (lines 11994-12010) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 12010-12014) - [Example usage:](misc/example_usage.md) (lines 12014-12070) - [Jina Embedder](misc/jina_embedder.md) (lines 12070-12086) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 12086-12090) - [Example usage:](misc/example_usage.md) (lines 12090-12146) - [Mistral Embedder](misc/mistral_embedder.md) (lines 12146-12162) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 12162-12166) - [Example usage:](misc/example_usage.md) (lines 12166-12222) - [Ollama Embedder](misc/ollama_embedder.md) (lines 12222-12238) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 12238-12242) - [Example usage:](misc/example_usage.md) (lines 12242-12296) - [OpenAI Embedder](misc/openai_embedder.md) (lines 12296-12312) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 12312-12316) - [Example usage:](misc/example_usage.md) (lines 12316-12372) - [Qdrant FastEmbed Embedder](misc/qdrant_fastembed_embedder.md) (lines 12372-12388) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 12388-12392) - [Example usage:](misc/example_usage.md) (lines 12392-12442) - [LanceDB Hybrid Search](misc/lancedb_hybrid_search.md) (lines 12442-12521) - [MilvusDB Hybrid Search](misc/milvusdb_hybrid_search.md) (lines 12521-12597) - [MongoDB Hybrid Search](misc/mongodb_hybrid_search.md) (lines 12597-12612) - [MongoDB Atlas connection string](misc/mongodb_atlas_connection_string.md) (lines 12612-12684) - [PgVector Hybrid Search](misc/pgvector_hybrid_search.md) (lines 12684-12704) - [QdrantDB Hybrid Search](misc/qdrantdb_hybrid_search.md) (lines 12757-12836) - [Weaviate Hybrid Search](misc/weaviate_hybrid_search.md) (lines 12836-12917) - [Use the agent](misc/use_the_agent.md) (lines 13073-13116) - [Use the agent](misc/use_the_agent.md) (lines 13149-13192) - [Load documents from the data/docs directory](misc/load_documents_from_the_data_docs_directory.md) (lines 13288-13291) - [Database connection URL](misc/database_connection_url.md) (lines 13291-13294) - [Agentic filtering with Docx](misc/agentic_filtering_with_docx.md) (lines 13432-13450) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 13450-13455) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 13455-13456) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 13456-13462) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 13463-13464) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 13465-13513) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 13513-13516) - [-----------------------------------------------------------------------------------](misc/-----------------------------------------------------------------------------------.md) (lines 13517-13519) - [Enable agentic filtering](misc/enable_agentic_filtering.md) (lines 13519-13526) - [Query for Jordan Mitchell's experience and skills with filters in query so that Agent can automatically pick them up](misc/query_for_jordan_mitchell_s_experience_and_skills_with_filters_in_query_so_that_agent_can_automatically_pick_them_up.md) (lines 13526-13556) - [Filtering with Docx](misc/filtering_with_docx.md) (lines 13556-13589) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 13589-13594) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 13594-13600) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 13648-13651) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 13652-13654) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 13654-13655) - [Query for Jordan Mitchell's experience and skills](misc/query_for_jordan_mitchell_s_experience_and_skills.md) (lines 13662-13668) - [# Option 2: Filters on the run/print_response](misc/option_2_filters_on_the_run_print_response.md) (lines 13668-13669) - [agent = Agent(](misc/agent_agent.md) (lines 13669-13670) - [)](misc/.md) (lines 13672-13674) - [# Query for Taylor Brooks as a candidate](misc/query_for_taylor_brooks_as_a_candidate.md) (lines 13674-13675) - [agent.print_response(](misc/agent_print_response.md) (lines 13675-13676) - ["Tell me about Taylor Brooks as a candidate",](misc/tell_me_about_taylor_brooks_as_a_candidate.md) (lines 13676-13677) - [markdown=True,](misc/markdown_true.md) (lines 13678-13679) - [)](misc/.md) (lines 13679-13705) - [Filtering on load with Docx](misc/filtering_on_load_with_docx.md) (lines 13705-13721) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 13721-13726) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 13726-13727) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 13727-13733) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 13734-13735) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 13736-13738) - [Load second document with user_2 metadata](misc/load_second_document_with_user_2_metadata.md) (lines 13750-13756) - [Load second document with user_3 metadata](misc/load_second_document_with_user_3_metadata.md) (lines 13756-13762) - [Load second document with user_4 metadata](misc/load_second_document_with_user_4_metadata.md) (lines 13762-13768) - [Load second document with user_5 metadata](misc/load_second_document_with_user_5_metadata.md) (lines 13768-13774) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 13775-13776) - [Uncomment the example you want to run](misc/uncomment_the_example_you_want_to_run.md) (lines 13776-13778) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 13778-13779) - [agent = Agent(](misc/agent_agent.md) (lines 13790-13791) - [)](misc/.md) (lines 13793-13794) - [agent.print_response(](misc/agent_print_response.md) (lines 13794-13795) - ["Tell me about Jordan Mitchell's experience and skills",](misc/tell_me_about_jordan_mitchell_s_experience_and_skills.md) (lines 13795-13796) - [markdown=True,](misc/markdown_true.md) (lines 13797-13798) - [)](misc/.md) (lines 13798-13824) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 13840-13845) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 13845-13846) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 13846-13852) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 13899-13902) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 13904-13912) - [Query for Jordan Mitchell's experience and skills](misc/query_for_jordan_mitchell_s_experience_and_skills.md) (lines 13912-13918) - [Option 2: Filters on the run/print_response](misc/option_2_filters_on_the_run_print_response.md) (lines 13918-13919) - [agent = Agent(](misc/agent_agent.md) (lines 13919-13920) - [add_references=True,](misc/add_references_true.md) (lines 13921-13922) - [)](misc/.md) (lines 13923-13925) - [Query for Taylor Brooks as a candidate](misc/query_for_taylor_brooks_as_a_candidate.md) (lines 13925-13926) - [agent.print_response(](misc/agent_print_response.md) (lines 13926-13927) - ["Tell me about Taylor Brooks as a candidate",](misc/tell_me_about_taylor_brooks_as_a_candidate.md) (lines 13927-13928) - [markdown=True,](misc/markdown_true.md) (lines 13929-13930) - [)](misc/.md) (lines 13930-13956) - [Filtering on ChromaDB](misc/filtering_on_chromadb.md) (lines 13956-13972) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 13972-13977) - [Initialize ChromaDB](misc/initialize_chromadb.md) (lines 13977-13980) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 13981-13982) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 13983-14031) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 14031-14034) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14035-14072) - [Filtering on LanceDB](misc/filtering_on_lancedb.md) (lines 14072-14088) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 14088-14093) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 14093-14094) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 14094-14100) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14101-14102) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 14103-14151) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 14151-14154) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14155-14192) - [Filtering on MilvusDB](misc/filtering_on_milvusdb.md) (lines 14192-14208) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 14208-14213) - [Initialize Milvus vector db](misc/initialize_milvus_vector_db.md) (lines 14213-14219) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14220-14221) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 14222-14270) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 14270-14273) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14274-14311) - [Filtering on MongoDB](misc/filtering_on_mongodb.md) (lines 14311-14327) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 14327-14334) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14335-14336) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 14337-14389) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 14389-14392) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14393-14430) - [Filtering on PgVector](misc/filtering_on_pgvector.md) (lines 14430-14446) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 14446-14451) - [Initialize PgVector](misc/initialize_pgvector.md) (lines 14451-14456) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14457-14458) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 14459-14507) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 14507-14510) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14511-14548) - [Filtering on Pinecone](misc/filtering_on_pinecone.md) (lines 14548-14566) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 14566-14571) - [Initialize Pinecone](misc/initialize_pinecone.md) (lines 14571-14584) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 14631-14634) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14635-14672) - [Filtering on Qdrant](misc/filtering_on_qdrant.md) (lines 14672-14688) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 14688-14696) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14697-14698) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 14699-14747) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 14747-14750) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14751-14788) - [Filtering on SurrealDB](misc/filtering_on_surrealdb.md) (lines 14788-14805) - [SurrealDB connection parameters](misc/surrealdb_connection_parameters.md) (lines 14805-14812) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 14812-14817) - [Create a client](misc/create_a_client.md) (lines 14817-14831) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14832-14833) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 14834-14882) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 14882-14885) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14886-14888) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 14888-14889) - [Filtering on Weaviate](misc/filtering_on_weaviate.md) (lines 14926-14943) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 14943-14948) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 14949-14950) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 14951-15006) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 15006-15009) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15010-15047) - [Agentic filtering with Json](misc/agentic_filtering_with_json.md) (lines 15047-15065) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 15065-15070) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 15070-15071) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 15071-15077) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15078-15079) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 15080-15128) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 15128-15131) - [-----------------------------------------------------------------------------------](misc/-----------------------------------------------------------------------------------.md) (lines 15132-15134) - [Enable agentic filtering](misc/enable_agentic_filtering.md) (lines 15134-15141) - [Query for Jordan Mitchell's experience and skills with filters in query so that Agent can automatically pick them up](misc/query_for_jordan_mitchell_s_experience_and_skills_with_filters_in_query_so_that_agent_can_automatically_pick_them_up.md) (lines 15141-15171) - [Filtering with Json](misc/filtering_with_json.md) (lines 15171-15204) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 15204-15209) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 15209-15210) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 15210-15216) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15217-15218) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 15219-15267) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 15267-15270) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15271-15273) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 15273-15274) - [Query for Jordan Mitchell' experience and skills](misc/query_for_jordan_mitchell_experience_and_skills.md) (lines 15281-15287) - [# Option 2: Filters on the run/print_response](misc/option_2_filters_on_the_run_print_response.md) (lines 15287-15288) - [agent = Agent(](misc/agent_agent.md) (lines 15288-15289) - [)](misc/.md) (lines 15291-15293) - [# Query for Taylor Brooks as a candidate](misc/query_for_taylor_brooks_as_a_candidate.md) (lines 15293-15294) - [agent.print_response(](misc/agent_print_response.md) (lines 15294-15295) - ["Tell me about Taylor Brooks as a candidate",](misc/tell_me_about_taylor_brooks_as_a_candidate.md) (lines 15295-15296) - [markdown=True,](misc/markdown_true.md) (lines 15297-15298) - [)](misc/.md) (lines 15298-15324) - [Filtering on load with Json](misc/filtering_on_load_with_json.md) (lines 15324-15340) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 15340-15345) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 15345-15346) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 15346-15352) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15353-15354) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 15355-15357) - [Load second document with user_2 metadata](misc/load_second_document_with_user_2_metadata.md) (lines 15369-15375) - [Load second document with user_3 metadata](misc/load_second_document_with_user_3_metadata.md) (lines 15375-15381) - [Load second document with user_4 metadata](misc/load_second_document_with_user_4_metadata.md) (lines 15381-15387) - [Load second document with user_5 metadata](misc/load_second_document_with_user_5_metadata.md) (lines 15387-15393) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15394-15395) - [Uncomment the example you want to run](misc/uncomment_the_example_you_want_to_run.md) (lines 15395-15397) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 15397-15398) - [agent = Agent(](misc/agent_agent.md) (lines 15409-15410) - [)](misc/.md) (lines 15412-15413) - [agent.print_response(](misc/agent_print_response.md) (lines 15413-15414) - ["Tell me about Jordan Mitchell's experience and skills",](misc/tell_me_about_jordan_mitchell_s_experience_and_skills.md) (lines 15414-15415) - [markdown=True,](misc/markdown_true.md) (lines 15416-15417) - [)](misc/.md) (lines 15417-15443) - [Agentic filtering with Pdf](misc/agentic_filtering_with_pdf.md) (lines 15443-15461) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 15461-15466) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 15466-15467) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 15467-15473) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15474-15475) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 15476-15524) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 15524-15527) - [-----------------------------------------------------------------------------------](misc/-----------------------------------------------------------------------------------.md) (lines 15528-15530) - [Enable agentic filtering](misc/enable_agentic_filtering.md) (lines 15530-15537) - [Query for Jordan Mitchell's experience and skills with filters in query so that Agent can automatically pick them up](misc/query_for_jordan_mitchell_s_experience_and_skills_with_filters_in_query_so_that_agent_can_automatically_pick_them_up.md) (lines 15537-15567) - [Filtering with Pdf](misc/filtering_with_pdf.md) (lines 15567-15583) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 15583-15588) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 15588-15589) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 15589-15595) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 15642-15645) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 15647-15683) - [Filtering on load with Pdf](misc/filtering_on_load_with_pdf.md) (lines 15683-15699) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 15699-15704) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 15704-15705) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 15705-15711) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15712-15713) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 15714-15716) - [Load second document with user_2 metadata](misc/load_second_document_with_user_2_metadata.md) (lines 15728-15734) - [Load second document with user_3 metadata](misc/load_second_document_with_user_3_metadata.md) (lines 15734-15740) - [Load second document with user_4 metadata](misc/load_second_document_with_user_4_metadata.md) (lines 15740-15746) - [Load second document with user_5 metadata](misc/load_second_document_with_user_5_metadata.md) (lines 15746-15752) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15753-15754) - [Uncomment the example you want to run](misc/uncomment_the_example_you_want_to_run.md) (lines 15754-15756) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 15756-15757) - [Agentic filtering with Pdf-Url](misc/agentic_filtering_with_pdf-url.md) (lines 15792-15804) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 15804-15805) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 15805-15811) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15812-15813) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 15814-15838) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 15838-15841) - [-----------------------------------------------------------------------------------](misc/-----------------------------------------------------------------------------------.md) (lines 15842-15844) - [Enable agentic filtering](misc/enable_agentic_filtering.md) (lines 15844-15851) - [Query for Jordan Mitchell's experience and skills with filters in query so that Agent can automatically pick them up](misc/query_for_jordan_mitchell_s_experience_and_skills_with_filters_in_query_so_that_agent_can_automatically_pick_them_up.md) (lines 15851-15881) - [Filtering with Pdf-Url](misc/filtering_with_pdf-url.md) (lines 15881-15911) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 15911-15912) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 15912-15918) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15919-15920) - [This metadata can include cuisine type, source, region, or any other attributes](misc/this_metadata_can_include_cuisine_type_source_region_or_any_other_attributes.md) (lines 15921-15945) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 15945-15948) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 15949-15951) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 15951-15952) - [Query for Thai recipes](misc/query_for_thai_recipes.md) (lines 15960-15966) - [# Option 2: Filters on the run/print_response](misc/option_2_filters_on_the_run_print_response.md) (lines 15966-15967) - [agent = Agent(](misc/agent_agent.md) (lines 15967-15968) - [)](misc/.md) (lines 15970-15972) - [# Query for Cape Malay recipes](misc/query_for_cape_malay_recipes.md) (lines 15972-15973) - [agent.print_response(](misc/agent_print_response.md) (lines 15973-15974) - ["Tell me how to make Cape Malay Curry",](misc/tell_me_how_to_make_cape_malay_curry.md) (lines 15974-15975) - [markdown=True,](misc/markdown_true.md) (lines 15976-15977) - [)](misc/.md) (lines 15977-16003) - [Filtering on load with Pdf-Url](misc/filtering_on_load_with_pdf-url.md) (lines 16003-16015) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 16015-16016) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 16016-16022) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 16023-16024) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 16025-16027) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 16044-16045) - [# # Option 2: Filters on the run/print_response](misc/option_2_filters_on_the_run_print_response.md) (lines 16058-16059) - [agent = Agent(](misc/agent_agent.md) (lines 16059-16060) - [)](misc/.md) (lines 16062-16063) - [agent.print_response(](misc/agent_print_response.md) (lines 16063-16064) - ["Tell me how to make Cape Malay Curry",](misc/tell_me_how_to_make_cape_malay_curry.md) (lines 16064-16065) - [markdown=True,](misc/markdown_true.md) (lines 16066-16067) - [)](misc/.md) (lines 16067-16093) - [Agentic filtering with Text](misc/agentic_filtering_with_text.md) (lines 16093-16111) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 16111-16116) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 16116-16117) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 16117-16123) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 16124-16125) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 16126-16174) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 16174-16177) - [-----------------------------------------------------------------------------------](misc/-----------------------------------------------------------------------------------.md) (lines 16178-16180) - [Enable agentic filtering](misc/enable_agentic_filtering.md) (lines 16180-16187) - [Query for Jordan Mitchell's experience and skills with filters in query so that Agent can automatically pick them up](misc/query_for_jordan_mitchell_s_experience_and_skills_with_filters_in_query_so_that_agent_can_automatically_pick_them_up.md) (lines 16187-16217) - [Filtering with Text](misc/filtering_with_text.md) (lines 16217-16250) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 16250-16255) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 16255-16256) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 16256-16262) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 16263-16264) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 16265-16313) - [Load all documents into the vector database](misc/load_all_documents_into_the_vector_database.md) (lines 16313-16316) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 16317-16319) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 16319-16320) - [Query for Jordan Mitchell's experience and skills](misc/query_for_jordan_mitchell_s_experience_and_skills.md) (lines 16327-16333) - [# Option 2: Filters on the run/print_response](misc/option_2_filters_on_the_run_print_response.md) (lines 16333-16334) - [agent = Agent(](misc/agent_agent.md) (lines 16334-16335) - [)](misc/.md) (lines 16337-16339) - [# Query for Taylor Brooks as a candidate](misc/query_for_taylor_brooks_as_a_candidate.md) (lines 16339-16340) - [agent.print_response(](misc/agent_print_response.md) (lines 16340-16341) - ["Tell me about Taylor Brooks as a candidate",](misc/tell_me_about_taylor_brooks_as_a_candidate.md) (lines 16341-16342) - [markdown=True,](misc/markdown_true.md) (lines 16343-16344) - [)](misc/.md) (lines 16344-16370) - [Filtering on load with Text](misc/filtering_on_load_with_text.md) (lines 16370-16386) - [Download all sample CVs and get their paths](misc/download_all_sample_cvs_and_get_their_paths.md) (lines 16386-16391) - [Initialize LanceDB](misc/initialize_lancedb.md) (lines 16391-16392) - [By default, it stores data in /tmp/lancedb](misc/by_default_it_stores_data_in_tmp_lancedb.md) (lines 16392-16398) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 16399-16400) - [This metadata can include user IDs, document types, dates, or any other attributes](misc/this_metadata_can_include_user_ids_document_types_dates_or_any_other_attributes.md) (lines 16401-16403) - [Load second document with user_2 metadata](misc/load_second_document_with_user_2_metadata.md) (lines 16415-16421) - [Load second document with user_3 metadata](misc/load_second_document_with_user_3_metadata.md) (lines 16421-16427) - [Load second document with user_4 metadata](misc/load_second_document_with_user_4_metadata.md) (lines 16427-16433) - [Load second document with user_5 metadata](misc/load_second_document_with_user_5_metadata.md) (lines 16433-16439) - [------------------------------------------------------------------------------](misc/------------------------------------------------------------------------------.md) (lines 16440-16441) - [Uncomment the example you want to run](misc/uncomment_the_example_you_want_to_run.md) (lines 16441-16443) - [Option 1: Filters on the Agent](misc/option_1_filters_on_the_agent.md) (lines 16443-16444) - [agent = Agent(](misc/agent_agent.md) (lines 16455-16456) - [)](misc/.md) (lines 16458-16459) - [agent.print_response(](misc/agent_print_response.md) (lines 16459-16460) - ["Tell me about Jordan Mitchell's experience and skills",](misc/tell_me_about_jordan_mitchell_s_experience_and_skills.md) (lines 16460-16461) - [markdown=True,](misc/markdown_true.md) (lines 16462-16463) - [)](misc/.md) (lines 16463-16489) - [-*- Create a run](misc/-_-_create_a_run.md) (lines 16510-16512) - [-*- Ask a follow up question that continues the conversation](misc/-_-_ask_a_follow_up_question_that_continues_the_conversation.md) (lines 16520-16522) - [Add memories for Jane Doe](misc/add_memories_for_jane_doe.md) (lines 16578-16593) - [Run 1](misc/run_1.md) (lines 16663-16669) - [Run this the 2nd time](misc/run_this_the_2nd_time.md) (lines 16669-16670) - [user_id=john_doe_id,](misc/user_id_john_doe_id.md) (lines 16672-16673) - [)](misc/.md) (lines 16673-16707) - [Reset for this example](misc/reset_for_this_example.md) (lines 16726-16764) - [Send a history of messages and add memories](misc/send_a_history_of_messages_and_add_memories.md) (lines 16768-16827) - [Reset for this example](misc/reset_for_this_example.md) (lines 16847-16912) - [No need to set the model, it gets set by the agent to the agent's model](misc/no_need_to_set_the_model_it_gets_set_by_the_agent_to_the_agent_s_model.md) (lines 16936-16939) - [-*- Print the chat history](misc/-_-_print_the_chat_history.md) (lines 16965-16980) - [-*- Print the chat history](misc/-_-_print_the_chat_history.md) (lines 16980-16984) - [You can also get the user memories from the agent](misc/you_can_also_get_the_user_memories_from_the_agent.md) (lines 16984-17021) - [No need to set the model, it gets set by the agent to the agent's model](misc/no_need_to_set_the_model_it_gets_set_by_the_agent_to_the_agent_s_model.md) (lines 17043-17046) - [Agent with Session Summaries](misc/agent_with_session_summaries.md) (lines 17130-17157) - [No session and user ID is specified, so they are generated automatically](misc/no_session_and_user_id_is_specified_so_they_are_generated_automatically.md) (lines 17160-17186) - [Now lets do a new session with a different user](misc/now_lets_do_a_new_session_with_a_different_user.md) (lines 17186-17209) - [No need to set the model, it gets set by the agent to the agent's model](misc/no_need_to_set_the_model_it_gets_set_by_the_agent_to_the_agent_s_model.md) (lines 17267-17270) - [You can also override the entire `system_message` for the session summarizer](misc/you_can_also_override_the_entire_system_message_for_the_session_summarizer.md) (lines 17373-17387) - [Multi-User Multi-Session Chat](misc/multi-user_multi-session_chat.md) (lines 17441-17473) - [Multi-User Multi-Session Chat Concurrent](misc/multi-user_multi-session_chat_concurrent.md) (lines 17592-17622) - [Alternatively, you can create/update user memories but not add them to the Agent](misc/alternatively_you_can_create_update_user_memories_but_not_add_them_to_the_agent.md) (lines 17788-17789) - [agent = Agent(](misc/agent_agent.md) (lines 17789-17790) - [model=Gemini(id="gemini-2.0-flash-exp"),](misc/model_gemini_id_gemini-2_0-flash-exp.md) (lines 17790-17791) - [enable_user_memories=True,](misc/enable_user_memories_true.md) (lines 17792-17793) - [)](misc/.md) (lines 17794-17824) - [Session Summary References](misc/session_summary_references.md) (lines 17824-17845) - [Reset for this example](misc/reset_for_this_example.md) (lines 17845-17861) - [This will create a new session summary](misc/this_will_create_a_new_session_summary.md) (lines 17861-17867) - [You can use existing session summaries from session storage without creating or updating any new ones.](misc/you_can_use_existing_session_summaries_from_session_storage_without_creating_or_updating_any_new_ones.md) (lines 17867-17904) - [Get MongoDB connection string from environment](misc/get_mongodb_connection_string_from_environment.md) (lines 17924-17925) - [Format: mongodb://username:password@localhost:27017/](misc/format_mongodb_username_password_localhost_27017.md) (lines 17925-17929) - [This will create the collection if it doesn't exist](misc/this_will_create_the_collection_if_it_doesn_t_exist.md) (lines 17939-17942) - [This will clear any existing memories](misc/this_will_clear_any_existing_memories.md) (lines 18107-18110) - [Session and user identifiers](misc/session_and_user_identifiers.md) (lines 18110-18114) - [First interaction - introducing personal information](misc/first_interaction_-_introducing_personal_information.md) (lines 18123-18131) - [Display the memories stored in Redis](misc/display_the_memories_stored_in_redis.md) (lines 18139-18183) - [This will create the table if it doesn't exist](misc/this_will_create_the_table_if_it_doesn_t_exist.md) (lines 18210-18213) - [Session and user identifiers](misc/session_and_user_identifiers.md) (lines 18213-18217) - [First interaction - introducing personal information](misc/first_interaction_-_introducing_personal_information.md) (lines 18229-18237) - [Display the memories stored in SQLite](misc/display_the_memories_stored_in_sqlite.md) (lines 18245-18283) - [Comment out the following line after running the script once](misc/comment_out_the_following_line_after_running_the_script_once.md) (lines 18304-18354) - [Audio Input Output](misc/audio_input_output.md) (lines 18354-18368) - [Fetch the audio file and convert it to a base64 encoded string](misc/fetch_the_audio_file_and_convert_it_to_a_base64_encoded_string.md) (lines 18368-18425) - [Multi-turn Audio Agent](misc/multi-turn_audio_agent.md) (lines 18425-18491) - [Audio Sentiment Analysis Agent](misc/audio_sentiment_analysis_agent.md) (lines 18491-18553) - [Audio Configuration](misc/audio_configuration.md) (lines 18567-18585) - [Audio to text Agent](misc/audio_to_text_agent.md) (lines 18640-18702) - [Blog to Podcast Agent](misc/blog_to_podcast_agent.md) (lines 18702-18800) - [Generate Images with Intermediate Steps](misc/generate_images_with_intermediate_steps.md) (lines 18800-18869) - [Generate Music using Models Lab](misc/generate_music_using_models_lab.md) (lines 18869-18952) - [Generate Video using Models Lab](misc/generate_video_using_models_lab.md) (lines 18952-19013) - [Generate Video using Replicate](misc/generate_video_using_replicate.md) (lines 19013-19079) - [Image to Audio Agent](misc/image_to_audio_agent.md) (lines 19079-19156) - [Image to Image Agent](misc/image_to_image_agent.md) (lines 19156-19222) - [Image to Text Agent](misc/image_to_text_agent.md) (lines 19222-19287) - [Video Caption Agent](misc/video_caption_agent.md) (lines 19287-19360) - [Video to Shorts Agent](misc/video_to_shorts_agent.md) (lines 19360-19399) - [Upload and process video](misc/upload_and_process_video.md) (lines 19399-19405) - [Multimodal Query for Video Analysis](misc/multimodal_query_for_video_analysis.md) (lines 19405-19434) - [Generate Video Analysis](misc/generate_video_analysis.md) (lines 19434-19437) - [Create output directory](misc/create_output_directory.md) (lines 19437-19441) - [Extract and cut video segments](misc/extract_and_cut_video_segments.md) (lines 19441-19491) - [Process segments](misc/process_segments.md) (lines 19491-19494) - [Print results](misc/print_results.md) (lines 19494-19546) - [Arize Phoenix via OpenInference](misc/arize_phoenix_via_openinference.md) (lines 19546-19566) - [Set environment variables for Arize Phoenix](misc/set_environment_variables_for_arize_phoenix.md) (lines 19566-19570) - [Configure the Phoenix tracer](misc/configure_the_phoenix_tracer.md) (lines 19570-19576) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 19576-19585) - [Use the agent](misc/use_the_agent.md) (lines 19585-19618) - [Arize Phoenix via OpenInference (Local Collector)](misc/arize_phoenix_via_openinference_local_collector.md) (lines 19618-19637) - [Set the local collector endpoint for Arize Phoenix](misc/set_the_local_collector_endpoint_for_arize_phoenix.md) (lines 19637-19640) - [Configure the Phoenix tracer](misc/configure_the_phoenix_tracer.md) (lines 19640-19646) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 19646-19655) - [Use the agent](misc/use_the_agent.md) (lines 19655-19690) - [Langfuse via OpenInference](misc/langfuse_via_openinference.md) (lines 19690-19720) - [os.environ["OTEL_EXPORTER_OTLP_ENDPOINT"]="https://cloud.langfuse.com/api/public/otel" # 🇪🇺 EU data region](misc/os_environ_otel_exporter_otlp_endpoint_https_cloud_langfuse_com_api_public_otel_eu_data_region.md) (lines 19720-19721) - [os.environ["OTEL_EXPORTER_OTLP_ENDPOINT"]="http://localhost:3000/api/public/otel" # 🏠 Local deployment (>= v3.22.0)](misc/os_environ_otel_exporter_otlp_endpoint_http_localhost_3000_api_public_otel_local_deployment_v3_22_0.md) (lines 19721-19729) - [Start instrumenting agno](misc/start_instrumenting_agno.md) (lines 19729-19780) - [Langfuse via OpenLIT](misc/langfuse_via_openlit.md) (lines 19780-19806) - [os.environ["OTEL_EXPORTER_OTLP_ENDPOINT"]="https://cloud.langfuse.com/api/public/otel" # 🇪🇺 EU data region](misc/os_environ_otel_exporter_otlp_endpoint_https_cloud_langfuse_com_api_public_otel_eu_data_region.md) (lines 19806-19807) - [os.environ["OTEL_EXPORTER_OTLP_ENDPOINT"]="http://localhost:3000/api/public/otel" # 🏠 Local deployment (>= v3.22.0)](misc/os_environ_otel_exporter_otlp_endpoint_http_localhost_3000_api_public_otel_local_deployment_v3_22_0.md) (lines 19807-19818) - [Sets the global default tracer provider](misc/sets_the_global_default_tracer_provider.md) (lines 19818-19823) - [Creates a tracer from the global tracer provider](misc/creates_a_tracer_from_the_global_tracer_provider.md) (lines 19823-19828) - [Initialize OpenLIT instrumentation. The disable_batch flag is set to true to process traces immediately.](misc/initialize_openlit_instrumentation_the_disable_batch_flag_is_set_to_true_to_process_traces_immediately.md) (lines 19828-19878) - [LangSmith](misc/langsmith.md) (lines 19878-19901) - [Set the endpoint and headers for LangSmith](misc/set_the_endpoint_and_headers_for_langsmith.md) (lines 19901-19908) - [Configure the tracer provider](misc/configure_the_tracer_provider.md) (lines 19908-19915) - [Start instrumenting agno](misc/start_instrumenting_agno.md) (lines 19915-19918) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 19918-19927) - [Use the agent](misc/use_the_agent.md) (lines 19927-19967) - [Langtrace](misc/langtrace.md) (lines 19967-19985) - [Initialize Langtrace](misc/initialize_langtrace.md) (lines 19985-19988) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 19988-19997) - [Use the agent](misc/use_the_agent.md) (lines 19997-20034) - [Weave](misc/weave.md) (lines 20034-20050) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 20050-20053) - [Initialize Weave with your project name](misc/initialize_weave_with_your_project_name.md) (lines 20053-20056) - [Define a function to run the agent, decorated with weave.op()](misc/define_a_function_to_run_the_agent_decorated_with_weave_op.md) (lines 20056-20061) - [Use the function to log a model call](misc/use_the_function_to_log_a_model_call.md) (lines 20061-20102) - [Fetch the audio file and convert it to a base64 encoded string](misc/fetch_the_audio_file_and_convert_it_to_a_base64_encoded_string.md) (lines 20117-20136) - [Datetime Instructions](misc/datetime_instructions.md) (lines 20276-20330) - [Image Input High Fidelity](misc/image_input_high_fidelity.md) (lines 20330-20391) - [Input as Dictionary](misc/input_as_dictionary.md) (lines 20391-20452) - [Input as List](misc/input_as_list.md) (lines 20452-20510) - [Input as Message](misc/input_as_message.md) (lines 20510-20571) - [Input as Messages List](misc/input_as_messages_list.md) (lines 20571-20634) - [Instructions](misc/instructions.md) (lines 20634-20681) - [Instructions via Function](misc/instructions_via_function.md) (lines 20681-20743) - [Intermediate Steps](misc/intermediate_steps.md) (lines 20743-20805) - [Location Instructions](misc/location_instructions.md) (lines 20805-20859) - [Response as Variable](misc/response_as_variable.md) (lines 20859-20893) - [run_response_strem: Iterator[RunResponse] = agent.run("What is the stock price of NVDA", stream=True)](misc/run_response_strem_iterator_runresponse_agent_run_what_is_the_stock_price_of_nvda_stream_true.md) (lines 20893-20894) - [for response in run_response_strem:](misc/for_response_in_run_response_strem.md) (lines 20894-20895) - [pprint(response)](misc/pprint_response.md) (lines 20895-20929) - [Success Criteria](misc/success_criteria.md) (lines 20929-20997) - [Agentic RAG with Agent UI](misc/agentic_rag_with_agent_ui.md) (lines 20997-21014) - [Agentic RAG with LanceDB](misc/agentic_rag_with_lancedb.md) (lines 21102-21116) - [Agentic RAG with PgVector](misc/agentic_rag_with_pgvector.md) (lines 21175-21190) - [Agentic RAG with Reranking](misc/agentic_rag_with_reranking.md) (lines 21263-21278) - [RAG with LanceDB and SQLite](misc/rag_with_lancedb_and_sqlite.md) (lines 21339-21354) - [Define the database URL where the vector database will be stored](misc/define_the_database_url_where_the_vector_database_will_be_stored.md) (lines 21354-21357) - [Configure the language model](misc/configure_the_language_model.md) (lines 21357-21360) - [Create Ollama embedder](misc/create_ollama_embedder.md) (lines 21360-21363) - [Create the vector database](misc/create_the_vector_database.md) (lines 21363-21370) - [Set up SQL storage for the agent's data](misc/set_up_sql_storage_for_the_agent_s_data.md) (lines 21379-21383) - [Use the agent to generate and print a response to a query, formatted in Markdown](misc/use_the_agent_to_generate_and_print_a_response_to_a_query_formatted_in_markdown.md) (lines 21394-21430) - [Traditional RAG with LanceDB](misc/traditional_rag_with_lancedb.md) (lines 21430-21444) - [Traditional RAG with PgVector](misc/traditional_rag_with_pgvector.md) (lines 21504-21519) - [Basic Reasoning Agent](misc/basic_reasoning_agent.md) (lines 21592-21651) - [Capture Reasoning Content](misc/capture_reasoning_content.md) (lines 21651-21668) - [Create agent with reasoning=True (default model COT)](misc/create_agent_with_reasoning_true_default_model_cot.md) (lines 21668-21675) - [Print the reasoning_content](misc/print_the_reasoning_content.md) (lines 21679-21689) - [Create agent with a specific reasoning_model](misc/create_agent_with_a_specific_reasoning_model.md) (lines 21689-21696) - [Print the reasoning_content](misc/print_the_reasoning_content.md) (lines 21702-21712) - [Non-Reasoning Model Agent](misc/non-reasoning_model_agent.md) (lines 21802-21831) - [It uses the default model of the Agent](misc/it_uses_the_default_model_of_the_agent.md) (lines 21831-21877) - [Azure AI Foundry](misc/azure_ai_foundry.md) (lines 21877-21942) - [Azure OpenAI o1](misc/azure_openai_o1.md) (lines 21942-21994) - [Azure OpenAI o3](misc/azure_openai_o3.md) (lines 21994-22057) - [Azure OpenAI GPT 4.1](misc/azure_openai_gpt_4_1.md) (lines 22057-22111) - [DeepSeek Reasoner](misc/deepseek_reasoner.md) (lines 22111-22173) - [Groq DeepSeek R1](misc/groq_deepseek_r1.md) (lines 22173-22225) - [Groq Claude + DeepSeek R1](misc/groq_claude_deepseek_r1.md) (lines 22225-22278) - [Ollama DeepSeek R1](misc/ollama_deepseek_r1.md) (lines 22278-22334) - [OpenAI o1 pro](misc/openai_o1_pro.md) (lines 22334-22385) - [OpenAI o3-mini](misc/openai_o3-mini.md) (lines 22385-22446) - [OpenAI o3-mini with reasoning effort](misc/openai_o3-mini_with_reasoning_effort.md) (lines 22446-22508) - [xAI Grok 3 Mini](misc/xai_grok_3_mini.md) (lines 22508-22569) - [Example usage with a complex reasoning problem](misc/example_usage_with_a_complex_reasoning_problem.md) (lines 23115-23157) - [Basic State Management](misc/basic_state_management.md) (lines 23227-23247) - [Create an Agent that maintains state](misc/create_an_agent_that_maintains_state.md) (lines 23247-23260) - [Example usage](misc/example_usage.md) (lines 23260-23297) - [State in Instructions](misc/state_in_instructions.md) (lines 23297-23355) - [Persistant State with Storage](misc/persistant_state_with_storage.md) (lines 23355-23371) - [Define a tool that adds an item to the shopping list](misc/define_a_tool_that_adds_an_item_to_the_shopping_list.md) (lines 23371-23397) - [Example usage](misc/example_usage.md) (lines 23397-23435) - [Multi User State](misc/multi_user_state.md) (lines 23435-23451) - [Organized by user ID and session ID](misc/organized_by_user_id_and_session_id.md) (lines 23452-23491) - [Create an Agent that maintains state](misc/create_an_agent_that_maintains_state.md) (lines 23491-23509) - [Example usage](misc/example_usage.md) (lines 23509-23541) - [What is on Mark Smith's shopping list?](misc/what_is_on_mark_smith_s_shopping_list.md) (lines 23541-23549) - [New session, so new shopping list](misc/new_session_so_new_shopping_list.md) (lines 23549-23591) - [State Management Across Multiple Runs](misc/state_management_across_multiple_runs.md) (lines 23591-23607) - [Create a Shopping List Manager Agent that maintains state](misc/create_a_shopping_list_manager_agent_that_maintains_state.md) (lines 23640-23658) - [Example usage](misc/example_usage.md) (lines 23658-23711) - [Team with Shared State](misc/team_with_shared_state.md) (lines 23711-23726) - [Example usage](misc/example_usage.md) (lines 23806-23860) - [DynamoDB Agent Storage](misc/dynamodb_agent_storage.md) (lines 23860-23874) - [AWS Credentials](misc/aws_credentials.md) (lines 23874-23889) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 23889-23902) - [JSON Agent Storage](misc/json_agent_storage.md) (lines 23902-23936) - [Mongo Agent Storage](misc/mongo_agent_storage.md) (lines 23936-23952) - [Create a storage backend using the Mongo database](misc/create_a_storage_backend_using_the_mongo_database.md) (lines 23952-23959) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 23959-23972) - [MySQL Agent Storage](misc/mysql_agent_storage.md) (lines 23972-24001) - [Create a storage backend using the Postgres database](misc/create_a_storage_backend_using_the_postgres_database.md) (lines 24001-24009) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 24009-24022) - [Postgres Agent Storage](misc/postgres_agent_storage.md) (lines 24022-24052) - [Create a storage backend using the Postgres database](misc/create_a_storage_backend_using_the_postgres_database.md) (lines 24052-24060) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 24060-24073) - [Redis Agent Storage](misc/redis_agent_storage.md) (lines 24073-24095) - [Initialize Redis storage with default local connection](misc/initialize_redis_storage_with_default_local_connection.md) (lines 24095-24105) - [Create agent with Redis storage](misc/create_agent_with_redis_storage.md) (lines 24105-24120) - [Singlestore Agent Storage](misc/singlestore_agent_storage.md) (lines 24120-24139) - [SingleStore Configuration](misc/singlestore_configuration.md) (lines 24139-24147) - [SingleStore DB URL](misc/singlestore_db_url.md) (lines 24147-24152) - [Create a database engine](misc/create_a_database_engine.md) (lines 24152-24155) - [Create a storage backend using the Singlestore database](misc/create_a_storage_backend_using_the_singlestore_database.md) (lines 24155-24165) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 24165-24178) - [Sqlite Agent Storage](misc/sqlite_agent_storage.md) (lines 24178-24192) - [Create a storage backend using the Sqlite database](misc/create_a_storage_backend_using_the_sqlite_database.md) (lines 24192-24200) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 24200-24213) - [YAML Agent Storage](misc/yaml_agent_storage.md) (lines 24213-24246) - [DynamoDB Team Storage](misc/dynamodb_team_storage.md) (lines 24246-24325) - [JSON Team Storage](misc/json_team_storage.md) (lines 24325-24402) - [Mongo Team Storage](misc/mongo_team_storage.md) (lines 24402-24428) - [MongoDB connection settings](misc/mongodb_connection_settings.md) (lines 24428-24486) - [MySQL Team Storage](misc/mysql_team_storage.md) (lines 24486-24515) - [Create a storage backend using the MySQL database](misc/create_a_storage_backend_using_the_mysql_database.md) (lines 24515-24523) - [Add storage to the Team](misc/add_storage_to_the_team.md) (lines 24523-24536) - [Postgres Team Storage](misc/postgres_team_storage.md) (lines 24536-24634) - [Redis Team Storage](misc/redis_team_storage.md) (lines 24634-24666) - [Initialize Redis storage with default local connection](misc/initialize_redis_storage_with_default_local_connection.md) (lines 24666-24729) - [Singlestore Team Storage](misc/singlestore_team_storage.md) (lines 24729-24759) - [Configure SingleStore DB connection](misc/configure_singlestore_db_connection.md) (lines 24759-24768) - [Download the certificate if SSL_CERT is not provided](misc/download_the_certificate_if_ssl_cert_is_not_provided.md) (lines 24768-24778) - [SingleStore DB URL](misc/singlestore_db_url.md) (lines 24778-24785) - [Create a DB engine](misc/create_a_db_engine.md) (lines 24785-24846) - [Sqlite Team Storage](misc/sqlite_team_storage.md) (lines 24846-24927) - [YAML Team Storage](misc/yaml_team_storage.md) (lines 24927-25004) - [DynamoDB Workflow Storage](misc/dynamodb_workflow_storage.md) (lines 25004-25118) - [JSON Workflow Storage](misc/json_workflow_storage.md) (lines 25118-25227) - [MongoDB Workflow Storage](misc/mongodb_workflow_storage.md) (lines 25227-25350) - [MySQL Workflow Storage](misc/mysql_workflow_storage.md) (lines 25350-25379) - [Create a storage backend using the MySQL database](misc/create_a_storage_backend_using_the_mysql_database.md) (lines 25379-25400) - [Postgres Workflow Storage](misc/postgres_workflow_storage.md) (lines 25400-25529) - [Redis Workflow Storage](misc/redis_workflow_storage.md) (lines 25529-25563) - [Initialize Redis storage with default local connection](misc/initialize_redis_storage_with_default_local_connection.md) (lines 25563-25660) - [Singlestore Workflow Storage](misc/singlestore_workflow_storage.md) (lines 25660-25804) - [SQLite Workflow Storage](misc/sqlite_workflow_storage.md) (lines 25804-25915) - [YAML Workflow Storage](misc/yaml_workflow_storage.md) (lines 25915-26024) - [null](misc/null.md) (lines 26213-26277) - [Initialize the Agent](misc/initialize_the_agent.md) (lines 26399-26407) - [Interact with the Agent so that it can learn about the user](misc/interact_with_the_agent_so_that_it_can_learn_about_the_user.md) (lines 26407-26412) - [Allow the memories to sync with Zep database](misc/allow_the_memories_to_sync_with_zep_database.md) (lines 26412-26415) - [Ask the Agent about the user](misc/ask_the_agent_about_the_user.md) (lines 26418-26454) - [Calculator](misc/calculator.md) (lines 26536-26597) - [Airbnb MCP agent](misc/airbnb_mcp_agent.md) (lines 26916-26963) - [GitHub MCP agent](misc/github_mcp_agent.md) (lines 26963-27028) - [Example usage](misc/example_usage.md) (lines 27028-27038) - [Notion MCP agent](misc/notion_mcp_agent.md) (lines 27064-27142) - [Pipedream Auth](misc/pipedream_auth.md) (lines 27142-27211) - [Pipedream Google Calendar](misc/pipedream_google_calendar.md) (lines 27211-27266) - [Pipedream LinkedIn](misc/pipedream_linkedin.md) (lines 27266-27321) - [Pipedream Slack](misc/pipedream_slack.md) (lines 27321-27381) - [Stagehand MCP agent](misc/stagehand_mcp_agent.md) (lines 27381-27406) - [Navigate to the stagehand directory](misc/navigate_to_the_stagehand_directory.md) (lines 27406-27409) - [Install dependencies and build](misc/install_dependencies_and_build.md) (lines 27409-27505) - [Stripe MCP agent](misc/stripe_mcp_agent.md) (lines 27505-27587) - [Supabase MCP agent](misc/supabase_mcp_agent.md) (lines 27614-27696) - [Meeting Summary Agent](misc/meeting_summary_agent.md) (lines 27709-27791) - [Recipe RAG Image Agent](misc/recipe_rag_image_agent.md) (lines 27791-27876) - [Using 'schedule' instead of deprecated 'schedule_interval'](misc/using_schedule_instead_of_deprecated_schedule_interval.md) (lines 27908-27963) - [Configure email settings](misc/configure_email_settings.md) (lines 28079-28084) - [Create an agent that can research and send personalized email updates](misc/create_an_agent_that_can_research_and_send_personalized_email_updates.md) (lines 28084-28175) - [E2B Code Execution](misc/e2b_code_execution.md) (lines 28560-28597) - [Example: Generate Fibonacci numbers](misc/example_generate_fibonacci_numbers.md) (lines 28597-28602) - [Example: Data visualization](misc/example_data_visualization.md) (lines 28602-28607) - [Example: Run a web server](misc/example_run_a_web_server.md) (lines 28607-28612) - [Example: Sandbox management](misc/example_sandbox_management.md) (lines 28612-28615) - [Example: File operations](misc/example_file_operations.md) (lines 28615-28664) - [Example 1: Financial Statements](misc/example_1_financial_statements.md) (lines 28755-28762) - [Example 2: Balance Sheet Analysis](misc/example_2_balance_sheet_analysis.md) (lines 28762-28802) - [Example 1: Business Search](misc/example_1_business_search.md) (lines 28996-29004) - [Example 2: Directions](misc/example_2_directions.md) (lines 29004-29013) - [Example 3: Address Validation and Geocoding](misc/example_3_address_validation_and_geocoding.md) (lines 29013-29022) - [Example 4: Distance Matrix](misc/example_4_distance_matrix.md) (lines 29022-29032) - [Example 5: Location Analysis](misc/example_5_location_analysis.md) (lines 29032-29046) - [Example 6: Multi-mode Transit Comparison](misc/example_6_multi-mode_transit_comparison.md) (lines 29046-29093) - [Example 1: Create a task](misc/example_1_create_a_task.md) (lines 29550-29555) - [Example 2: Delete a task](misc/example_2_delete_a_task.md) (lines 29555-29562) - [Example 3: Get all tasks](misc/example_3_get_all_tasks.md) (lines 29562-29599) - [DuckDuckGo Search](misc/duckduckgo_search.md) (lines 29958-29972) - [We will search DDG but limit the site to Politifact](misc/we_will_search_ddg_but_limit_the_site_to_politifact.md) (lines 29972-30012) - [List all spaces in Webex](misc/list_all_spaces_in_webex.md) (lines 30822-30825) - [Send a message to a Space in Webex](misc/send_a_message_to_a_space_in_webex.md) (lines 30825-30873) - [Example: Send a template message](misc/example_send_a_template_message.md) (lines 30891-30892) - [Note: Replace '''hello_world''' with your actual template name](misc/note_replace_hello_world_with_your_actual_template_name.md) (lines 30892-30937) - [Example 1: Scrape a webpage as Markdown](misc/example_1_scrape_a_webpage_as_markdown.md) (lines 31021-31059) - [User Confirmation Required](misc/user_confirmation_required.md) (lines 31410-31524) - [Async User Confirmation](misc/async_user_confirmation.md) (lines 31524-31644) - [User Input Required](misc/user_input_required.md) (lines 31768-31784) - [You can either specify the user_input_fields leave empty for all fields to be provided by the user](misc/you_can_either_specify_the_user_input_fields_leave_empty_for_all_fields_to_be_provided_by_the_user.md) (lines 31784-31868) - [Async User Input](misc/async_user_input.md) (lines 31868-31974) - [Dynamic User Input (Agentic)](misc/dynamic_user_input_agentic.md) (lines 32078-32141) - [We use a while loop to continue the running until the agent is satisfied with the user input](misc/we_use_a_while_loop_to_continue_the_running_until_the_agent_is_satisfied_with_the_user_input.md) (lines 32141-32210) - [External Tool Execution](misc/external_tool_execution.md) (lines 32210-32301) - [Async External Tool Execution](misc/async_external_tool_execution.md) (lines 32301-32397) - [We have to create a tool with the correct name, arguments and docstring for the agent to know what to call.](misc/we_have_to_create_a_tool_with_the_correct_name_arguments_and_docstring_for_the_agent_to_know_what_to_call.md) (lines 32412-32502) - [Azure Cosmos DB MongoDB vCore Integration](misc/azure_cosmos_db_mongodb_vcore_integration.md) (lines 32502-32515) - [Azure Cosmos DB MongoDB connection string](misc/azure_cosmos_db_mongodb_connection_string.md) (lines 32515-32532) - [Comment out after first run](misc/comment_out_after_first_run.md) (lines 32532-32535) - [Create and use the agent](misc/create_and_use_the_agent.md) (lines 32535-32565) - [Cassandra Integration](misc/cassandra_integration.md) (lines 32565-32646) - [ChromaDB Integration](misc/chromadb_integration.md) (lines 32646-32658) - [Initialize ChromaDB](misc/initialize_chromadb.md) (lines 32658-32661) - [Create and use the agent](misc/create_and_use_the_agent.md) (lines 32669-32699) - [Clickhouse Integration](misc/clickhouse_integration.md) (lines 32699-32775) - [Couchbase Integration](misc/couchbase_integration.md) (lines 32775-32792) - [Couchbase connection settings](misc/couchbase_connection_settings.md) (lines 32792-32797) - [Create cluster options with authentication](misc/create_cluster_options_with_authentication.md) (lines 32797-32823) - [Wait for the vector index to sync with KV](misc/wait_for_the_vector_index_to_sync_with_kv.md) (lines 32823-32881) - [LanceDB Integration](misc/lancedb_integration.md) (lines 32881-32934) - [Milvus Integration](misc/milvus_integration.md) (lines 32934-32986) - [MongoDB Integration](misc/mongodb_integration.md) (lines 32986-33040) - [PgVector Integration](misc/pgvector_integration.md) (lines 33040-33105) - [Pinecone Integration](misc/pinecone_integration.md) (lines 33105-33178) - [Qdrant Integration](misc/qdrant_integration.md) (lines 33178-33238) - [SingleStore Integration](misc/singlestore_integration.md) (lines 33238-33325) - [Weaviate Integration](misc/weaviate_integration.md) (lines 33325-33345) - [Create and use the agent](misc/create_and_use_the_agent.md) (lines 33352-33387) - [Accuracy with Given Answer](misc/accuracy_with_given_answer.md) (lines 33387-33415) - [Accuracy with Teams](misc/accuracy_with_teams.md) (lines 33415-33430) - [Setup a team with two members](misc/setup_a_team_with_two_members.md) (lines 33430-33456) - [Evaluate the accuracy of the Team's responses](misc/evaluate_the_accuracy_of_the_team_s_responses.md) (lines 33456-33470) - [Simple Accuracy](misc/simple_accuracy.md) (lines 33502-33535) - [Performance on Agent Instantiation](misc/performance_on_agent_instantiation.md) (lines 33535-33568) - [Performance on Agent Instantiation with Tool](misc/performance_on_agent_instantiation_with_tool.md) (lines 33568-33609) - [Performance on Agent Response](misc/performance_on_agent_response.md) (lines 33609-33643) - [Performance with Teams](misc/performance_with_teams.md) (lines 33643-33672) - [Performance on Agent with Storage](misc/performance_on_agent_with_storage.md) (lines 33672-33709) - [Reliability with Single Tool](misc/reliability_with_single_tool.md) (lines 33709-33747) - [Reliability with Teams](misc/reliability_with_teams.md) (lines 33785-33838) - [Example usage](misc/example_usage.md) (lines 33907-33933) - [Agent Session](misc/agent_session.md) (lines 33933-34061) - [Agent State](misc/agent_state.md) (lines 34061-34085) - [Define a tool that increments our counter and returns the new value](misc/define_a_tool_that_increments_our_counter_and_returns_the_new_value.md) (lines 34085-34092) - [Create a State Manager Agent that maintains state](misc/create_a_state_manager_agent_that_maintains_state.md) (lines 34092-34119) - [Example usage](misc/example_usage.md) (lines 34119-34125) - [Agent Team](misc/agent_team.md) (lines 34157-34273) - [Example usage with diverse queries](misc/example_usage_with_diverse_queries.md) (lines 34273-34285) - [Agent with Storage](misc/agent_with_storage.md) (lines 34435-34599) - [Create a News Reporter Agent with a fun personality](misc/create_a_news_reporter_agent_with_a_fun_personality.md) (lines 34623-34655) - [Example usage](misc/example_usage.md) (lines 34655-34660) - [Audio Agent](misc/audio_agent.md) (lines 34690-34715) - [Create an AI Voice Interaction Agent](misc/create_an_ai_voice_interaction_agent.md) (lines 34715-34739) - [Fetch the audio file and convert it to a base64 encoded string](misc/fetch_the_audio_file_and_convert_it_to_a_base64_encoded_string.md) (lines 34739-34744) - [Process the audio and get a response](misc/process_the_audio_and_get_a_response.md) (lines 34744-34750) - [Save the audio response if available](misc/save_the_audio_response_if_available.md) (lines 34750-34756) - [More example interactions to try:](misc/more_example_interactions_to_try.md) (lines 34756-34793) - [Basic Agent](misc/basic_agent.md) (lines 34793-34815) - [Create our News Reporter with a fun personality](misc/create_our_news_reporter_with_a_fun_personality.md) (lines 34815-34834) - [Example usage](misc/example_usage.md) (lines 34834-34839) - [Create a Tech News Reporter Agent with a Silicon Valley personality](misc/create_a_tech_news_reporter_agent_with_a_silicon_valley_personality.md) (lines 34923-34944) - [Example questions to try:](misc/example_questions_to_try.md) (lines 34944-34945) - [- "What are the trending tech discussions on HN right now?"](misc/-_what_are_the_trending_tech_discussions_on_hn_right_now.md) (lines 34945-34946) - [- "Summarize the top 5 stories on Hacker News"](misc/-_summarize_the_top_5_stories_on_hacker_news.md) (lines 34946-34947) - [- "What's the most upvoted story today?"](misc/-_what_s_the_most_upvoted_story_today.md) (lines 34947-34970) - [Human in the Loop](misc/human_in_the_loop.md) (lines 34970-35005) - [This is the console instance used by the print_response method](misc/this_is_the_console_instance_used_by_the_print_response_method.md) (lines 35005-35006) - [We can use this to stop and restart the live display and ask for user confirmation](misc/we_can_use_this_to_stop_and_restart_the_live_display_and_ask_for_user_confirmation.md) (lines 35006-35061) - [Initialize the agent with a tech-savvy personality and clear instructions](misc/initialize_the_agent_with_a_tech-savvy_personality_and_clear_instructions.md) (lines 35061-35081) - [Example questions to try:](misc/example_questions_to_try.md) (lines 35081-35082) - [- "What are the top 3 HN stories right now?"](misc/-_what_are_the_top_3_hn_stories_right_now.md) (lines 35082-35083) - [- "Show me the most recent story from Hacker News"](misc/-_show_me_the_most_recent_story_from_hacker_news.md) (lines 35083-35084) - [- "Get the top 5 stories (you can try accepting and declining the confirmation)"](misc/-_get_the_top_5_stories_you_can_try_accepting_and_declining_the_confirmation.md) (lines 35084-35109) - [Image Agent](misc/image_agent.md) (lines 35109-35182) - [Example usage with a famous landmark](misc/example_usage_with_a_famous_landmark.md) (lines 35182-35193) - [More examples to try:](misc/more_examples_to_try.md) (lines 35193-35208) - [To get the response in a variable:](misc/to_get_the_response_in_a_variable.md) (lines 35208-35209) - [from rich.pretty import pprint](misc/from_rich_pretty_import_pprint.md) (lines 35209-35210) - [response = agent.run(](misc/response_agent_run.md) (lines 35210-35211) - ["Analyze this landmark's architecture and recent news.",](misc/analyze_this_landmark_s_architecture_and_recent_news.md) (lines 35211-35212) - [images=[Image(url="YOUR_IMAGE_URL")],](misc/images_image_url_your_image_url.md) (lines 35212-35213) - [)](misc/.md) (lines 35213-35214) - [pprint(response.content)](misc/pprint_response_content.md) (lines 35214-35236) - [Image Generation](misc/image_generation.md) (lines 35236-35261) - [Create an Creative AI Artist Agent](misc/create_an_creative_ai_artist_agent.md) (lines 35261-35284) - [Example usage](misc/example_usage.md) (lines 35284-35289) - [Retrieve and display generated images](misc/retrieve_and_display_generated_images.md) (lines 35289-35296) - [Introduction](misc/introduction.md) (lines 35326-35439) - [Research Agent](misc/research_agent.md) (lines 35439-35541) - [Example usage](misc/example_usage.md) (lines 35541-35548) - [Research Workflow](misc/research_workflow.md) (lines 35578-35960) - [Run the workflow if the script is executed directly](misc/run_the_workflow_if_the_script_is_executed_directly.md) (lines 35960-36028) - [Retry Functions](misc/retry_functions.md) (lines 36028-36092) - [Structured Output](misc/structured_output.md) (lines 36092-36148) - [Agent that uses JSON mode](misc/agent_that_uses_json_mode.md) (lines 36148-36187) - [Agent that uses structured outputs](misc/agent_that_uses_structured_outputs.md) (lines 36187-36225) - [Example usage with different locations](misc/example_usage_with_different_locations.md) (lines 36225-36229) - [More examples to try:](misc/more_examples_to_try.md) (lines 36229-36239) - [To get the response in a variable:](misc/to_get_the_response_in_a_variable.md) (lines 36239-36240) - [from rich.pretty import pprint](misc/from_rich_pretty_import_pprint.md) (lines 36240-36242) - [json_mode_response: RunResponse = json_mode_agent.run("New York")](misc/json_mode_response_runresponse_json_mode_agent_run_new_york.md) (lines 36242-36243) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 36243-36244) - [structured_output_response: RunResponse = structured_output_agent.run("New York")](misc/structured_output_response_runresponse_structured_output_agent_run_new_york.md) (lines 36244-36245) - [pprint(structured_output_response.content)](misc/pprint_structured_output_response_content.md) (lines 36245-36267) - [User Memories](misc/user_memories.md) (lines 36267-36470) - [Video Generation](misc/video_generation.md) (lines 36470-36495) - [Create a Creative AI Video Director Agent](misc/create_a_creative_ai_video_director_agent.md) (lines 36495-36519) - [Example usage](misc/example_usage.md) (lines 36519-36524) - [Retrieve and display generated videos](misc/retrieve_and_display_generated_videos.md) (lines 36524-36530) - [Examples Gallery](misc/examples_gallery.md) (lines 36566-36721) - [Basic Agent](misc/basic_agent.md) (lines 36721-36734) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 36734-36735) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 36735-36736) - [print(run.content)](misc/print_run_content.md) (lines 36736-36738) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 36738-36773) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 36787-36788) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 36788-36789) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 36789-36790) - [print(chunk.content)](misc/print_chunk_content.md) (lines 36790-36792) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 36792-36827) - [Code Execution Tool](misc/code_execution_tool.md) (lines 36827-36861) - [File Upload](misc/file_upload.md) (lines 36861-36928) - [Download the file using the download_file function](misc/download_the_file_using_the_download_file_function.md) (lines 36928-36933) - [Initialize Anthropic client](misc/initialize_anthropic_client.md) (lines 36933-36936) - [Upload the file to Anthropic](misc/upload_the_file_to_anthropic.md) (lines 36936-36957) - [Image Input Bytes Content](misc/image_input_bytes_content.md) (lines 36957-36985) - [Read the image file content as bytes](misc/read_the_image_file_content_as_bytes.md) (lines 36985-37028) - [Image Input URL](misc/image_input_url.md) (lines 37028-37089) - [PDF Input Bytes Agent](misc/pdf_input_bytes_agent.md) (lines 37170-37186) - [Download the file using the download_file function](misc/download_the_file_using_the_download_file_function.md) (lines 37186-37240) - [PDF Input Local Agent](misc/pdf_input_local_agent.md) (lines 37240-37256) - [Download the file using the download_file function](misc/download_the_file_using_the_download_file_function.md) (lines 37256-37310) - [PDF Input URL Agent](misc/pdf_input_url_agent.md) (lines 37310-37367) - [Prompt Caching](misc/prompt_caching.md) (lines 37367-37421) - [Load an example large system message from S3. A large prompt like this would benefit from caching.](misc/load_an_example_large_system_message_from_s3_a_large_prompt_like_this_would_benefit_from_caching.md) (lines 37421-37442) - [First run - this will create the cache](misc/first_run_-_this_will_create_the_cache.md) (lines 37442-37448) - [Second run - this will use the cached system prompt](misc/second_run_-_this_will_use_the_cached_system_prompt.md) (lines 37448-37456) - [Agent with Storage](misc/agent_with_storage.md) (lines 37456-37526) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 37526-37567) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 37567-37603) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 37671-37672) - [run_response: AsyncIterator[RunResponse] = await agent.arun("Share a 2 sentence horror story", stream=True)](misc/run_response_asynciterator_runresponse_await_agent_arun_share_a_2_sentence_horror_story_stream_true.md) (lines 37672-37673) - [async for chunk in run_response:](misc/async_for_chunk_in_run_response.md) (lines 37673-37674) - [print(chunk.content)](misc/print_chunk_content.md) (lines 37674-37676) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 37676-37713) - [Basic Agent](misc/basic_agent.md) (lines 37767-37782) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 37782-37783) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 37783-37784) - [print(run.content)](misc/print_run_content.md) (lines 37784-37786) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 37786-37823) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 37839-37840) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 37840-37841) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 37841-37842) - [print(chunk.content)](misc/print_chunk_content.md) (lines 37842-37844) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 37844-37881) - [Agent with Image Input](misc/agent_with_image_input.md) (lines 37881-37905) - [Read the image file content as bytes](misc/read_the_image_file_content_as_bytes.md) (lines 37905-37954) - [Agent with Storage](misc/agent_with_storage.md) (lines 38030-38102) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 38102-38143) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 38143-38144) - [movie_agent: RunResponse = movie_agent.run("New York")](misc/movie_agent_runresponse_movie_agent_run_new_york.md) (lines 38144-38145) - [pprint(movie_agent.content)](misc/pprint_movie_agent_content.md) (lines 38145-38183) - [Basic Agent](misc/basic_agent.md) (lines 38237-38252) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 38252-38253) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 38253-38254) - [print(run.content)](misc/print_run_content.md) (lines 38254-38256) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 38256-38293) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 38309-38310) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 38310-38311) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 38311-38312) - [print(chunk.content)](misc/print_chunk_content.md) (lines 38312-38314) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 38314-38351) - [Agent with Storage](misc/agent_with_storage.md) (lines 38427-38499) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 38499-38540) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 38540-38541) - [movie_agent: RunResponse = movie_agent.run("New York")](misc/movie_agent_runresponse_movie_agent_run_new_york.md) (lines 38541-38542) - [pprint(movie_agent.content)](misc/pprint_movie_agent_content.md) (lines 38542-38580) - [Basic Agent](misc/basic_agent.md) (lines 38634-38647) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 38647-38648) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 38648-38649) - [print(run.content)](misc/print_run_content.md) (lines 38649-38651) - [Print the response on the terminal](misc/print_the_response_on_the_terminal.md) (lines 38651-38687) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 38702-38703) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 38703-38704) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 38704-38705) - [print(chunk.content)](misc/print_chunk_content.md) (lines 38705-38707) - [Print the response on the terminal](misc/print_the_response_on_the_terminal.md) (lines 38707-38743) - [Agent with Storage](misc/agent_with_storage.md) (lines 38824-38895) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 38895-38937) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 38937-38938) - [run: RunResponse = agent.run("New York")](misc/run_runresponse_agent_run_new_york.md) (lines 38938-38939) - [pprint(run.content)](misc/pprint_run_content.md) (lines 38939-38976) - [Basic Agent](misc/basic_agent.md) (lines 39030-39043) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 39043-39044) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 39044-39045) - [print(run.content)](misc/print_run_content.md) (lines 39045-39047) - [Print the response on the terminal](misc/print_the_response_on_the_terminal.md) (lines 39047-39084) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 39099-39100) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 39100-39101) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 39101-39102) - [print(chunk.content)](misc/print_chunk_content.md) (lines 39102-39104) - [Print the response on the terminal](misc/print_the_response_on_the_terminal.md) (lines 39104-39141) - [Agent with Storage](misc/agent_with_storage.md) (lines 39223-39295) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 39295-39337) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 39337-39338) - [run: RunResponse = agent.run("New York")](misc/run_runresponse_agent_run_new_york.md) (lines 39338-39339) - [pprint(run.content)](misc/pprint_run_content.md) (lines 39339-39377) - [Basic Agent](misc/basic_agent.md) (lines 39432-39484) - [Agent with Storage](misc/agent_with_storage.md) (lines 39606-39667) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 39667-39700) - [Agent that uses a JSON schema output](misc/agent_that_uses_a_json_schema_output.md) (lines 39700-39741) - [Basic Agent](misc/basic_agent.md) (lines 39795-39813) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 39813-39848) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 39919-39954) - [Basic Agent](misc/basic_agent.md) (lines 39954-39967) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 39967-39968) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 39968-39969) - [print(run.content)](misc/print_run_content.md) (lines 39969-39971) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 39971-40006) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40019-40020) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 40020-40021) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 40021-40022) - [print(chunk.content)](misc/print_chunk_content.md) (lines 40022-40024) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 40024-40059) - [Image Agent](misc/image_agent.md) (lines 40059-40118) - [Agent with Storage](misc/agent_with_storage.md) (lines 40192-40262) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 40262-40304) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40304-40305) - [json_mode_response: RunResponse = json_mode_agent.run("New York")](misc/json_mode_response_runresponse_json_mode_agent_run_new_york.md) (lines 40305-40306) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 40306-40342) - [Basic Agent](misc/basic_agent.md) (lines 40395-40411) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40411-40412) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 40412-40413) - [print(run.content)](misc/print_run_content.md) (lines 40413-40415) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 40415-40450) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40468-40469) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 40469-40470) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 40470-40471) - [print(chunk.content)](misc/print_chunk_content.md) (lines 40471-40473) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 40473-40508) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 40508-40549) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40549-40550) - [json_mode_response: RunResponse = json_mode_agent.run("New York")](misc/json_mode_response_runresponse_json_mode_agent_run_new_york.md) (lines 40550-40551) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 40551-40587) - [Basic Agent](misc/basic_agent.md) (lines 40640-40653) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40653-40654) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 40654-40655) - [print(run.content)](misc/print_run_content.md) (lines 40655-40657) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 40657-40692) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40705-40706) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 40706-40707) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 40707-40708) - [print(chunk.content)](misc/print_chunk_content.md) (lines 40708-40710) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 40710-40745) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 40745-40786) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40786-40787) - [json_mode_response: RunResponse = json_mode_agent.run("New York")](misc/json_mode_response_runresponse_json_mode_agent_run_new_york.md) (lines 40787-40788) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 40788-40824) - [Basic Agent](misc/basic_agent.md) (lines 40877-40893) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40893-40894) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 40894-40895) - [print(run.content)](misc/print_run_content.md) (lines 40895-40897) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 40897-40932) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 40949-40950) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 40950-40951) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 40951-40952) - [print(chunk.content)](misc/print_chunk_content.md) (lines 40952-40954) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 40954-40989) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 40989-41024) - [Agent that uses JSON mode](misc/agent_that_uses_json_mode.md) (lines 41024-41031) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 41031-41032) - [response: RunResponse = agent.run("New York")](misc/response_runresponse_agent_run_new_york.md) (lines 41032-41033) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 41033-41069) - [Audio Input (Bytes Content)](misc/audio_input_bytes_content.md) (lines 41121-41141) - [Download the audio file from the URL as bytes](misc/download_the_audio_file_from_the_url_as_bytes.md) (lines 41141-41182) - [Audio Input (Upload the file)](misc/audio_input_upload_the_file.md) (lines 41182-41202) - [Please download a sample audio file to test this Agent and upload using:](misc/please_download_a_sample_audio_file_to_test_this_agent_and_upload_using.md) (lines 41202-41261) - [Audio Input (Local file)](misc/audio_input_local_file.md) (lines 41261-41279) - [Please download a sample audio file to test this Agent and upload using:](misc/please_download_a_sample_audio_file_to_test_this_agent_and_upload_using.md) (lines 41279-41320) - [Basic Agent](misc/basic_agent.md) (lines 41320-41333) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 41333-41334) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 41334-41335) - [print(run.content)](misc/print_run_content.md) (lines 41335-41337) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 41337-41372) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 41386-41387) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 41387-41388) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 41388-41389) - [print(chunk.content)](misc/print_chunk_content.md) (lines 41389-41391) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 41391-41426) - [Flash Thinking Agent](misc/flash_thinking_agent.md) (lines 41426-41479) - [Image Agent](misc/image_agent.md) (lines 41479-41540) - [Agent with PDF Input (Local file)](misc/agent_with_pdf_input_local_file.md) (lines 41619-41635) - [Download the file using the download_file function](misc/download_the_file_using_the_download_file_function.md) (lines 41635-41684) - [Agent with PDF Input (URL)](misc/agent_with_pdf_input_url.md) (lines 41684-41738) - [Agent with Storage](misc/agent_with_storage.md) (lines 41738-41808) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 41808-41883) - [Video Input (Bytes Content)](misc/video_input_bytes_content.md) (lines 41935-41955) - [Download the video file from the URL as bytes](misc/download_the_video_file_from_the_url_as_bytes.md) (lines 41955-41996) - [Video Input (File Upload)](misc/video_input_file_upload.md) (lines 41996-42018) - [Please download a sample video file to test this Agent](misc/please_download_a_sample_video_file_to_test_this_agent.md) (lines 42018-42019) - [Run: `wget https://storage.googleapis.com/generativeai-downloads/images/GreatRedSpot.mp4` to download a sample video](misc/run_wget_https_storage_googleapis_com_generativeai-downloads_images_greatredspot_mp4_to_download_a_sample_video.md) (lines 42019-42029) - [Upload the video file if it doesn't exist](misc/upload_the_video_file_if_it_doesn_t_exist.md) (lines 42029-42086) - [Video Input (Local File Upload)](misc/video_input_local_file_upload.md) (lines 42086-42105) - [Get sample videos from https://www.pexels.com/search/videos/sample/](misc/get_sample_videos_from_https_www_pexels_com_search_videos_sample.md) (lines 42105-42142) - [Basic Agent](misc/basic_agent.md) (lines 42142-42155) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 42155-42156) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 42156-42157) - [print(run.content)](misc/print_run_content.md) (lines 42157-42159) - [Print the response on the terminal](misc/print_the_response_on_the_terminal.md) (lines 42159-42194) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 42208-42209) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 42209-42210) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 42210-42211) - [print(chunk.content)](misc/print_chunk_content.md) (lines 42211-42213) - [Print the response on the terminal](misc/print_the_response_on_the_terminal.md) (lines 42213-42248) - [Image Agent](misc/image_agent.md) (lines 42248-42302) - [Agent with Storage](misc/agent_with_storage.md) (lines 42376-42446) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 42446-42521) - [Basic Agent](misc/basic_agent.md) (lines 42581-42633) - [Llama Essay Writer](misc/llama_essay_writer.md) (lines 42686-42741) - [Async Basic Agent](misc/async_basic_agent.md) (lines 42741-42756) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 42756-42757) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 42757-42758) - [print(run.content)](misc/print_run_content.md) (lines 42758-42760) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 42760-42798) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 42815-42816) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 42816-42817) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 42817-42818) - [print(chunk.content)](misc/print_chunk_content.md) (lines 42818-42820) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 42820-42858) - [Agent with Async Tool Usage](misc/agent_with_async_tool_usage.md) (lines 42858-42914) - [Basic Agent](misc/basic_agent.md) (lines 42914-42927) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 42927-42928) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 42928-42929) - [print(run.content)](misc/print_run_content.md) (lines 42929-42931) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 42931-42969) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 42983-42984) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 42984-42985) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 42985-42986) - [print(chunk.content)](misc/print_chunk_content.md) (lines 42986-42988) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 42988-43026) - [Image Agent](misc/image_agent.md) (lines 43026-43049) - [Read the image file content as bytes](misc/read_the_image_file_content_as_bytes.md) (lines 43049-43102) - [RAG Agent](misc/rag_agent.md) (lines 43102-43175) - [Agent with Storage](misc/agent_with_storage.md) (lines 43175-43240) - [Agent with Structured Output](misc/agent_with_structured_output.md) (lines 43240-43281) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 43281-43282) - [movie_agent: RunResponse = movie_agent.run("New York")](misc/movie_agent_runresponse_movie_agent_run_new_york.md) (lines 43282-43283) - [pprint(movie_agent.content)](misc/pprint_movie_agent_content.md) (lines 43283-43322) - [Basic Usage](misc/basic_usage.md) (lines 43375-43394) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 43394-43395) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 43395-43396) - [print(run.content)](misc/print_run_content.md) (lines 43396-43398) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 43398-43435) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 43453-43454) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 43454-43455) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 43455-43456) - [print(chunk.content)](misc/print_chunk_content.md) (lines 43456-43458) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 43458-43494) - [agent.print_response("Share the NVDA stock price and analyst recommendations", stream=True)](misc/agent_print_response_share_the_nvda_stock_price_and_analyst_recommendations_stream_true.md) (lines 43523-43559) - [Basic Agent](misc/basic_agent.md) (lines 43559-43612) - [Agent with Storage](misc/agent_with_storage.md) (lines 43725-43738) - [Create a storage backend using the Sqlite database](misc/create_a_storage_backend_using_the_sqlite_database.md) (lines 43738-43746) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 43746-43789) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 43789-43865) - [Ask a question that would likely trigger tool use](misc/ask_a_question_that_would_likely_trigger_tool_use.md) (lines 43886-43921) - [Basic Agent](misc/basic_agent.md) (lines 43921-43940) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 43940-43941) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 43941-43942) - [print(run.content)](misc/print_run_content.md) (lines 43942-43944) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 43944-43979) - [Basic Agent](misc/basic_agent.md) (lines 44092-44105) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 44105-44106) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 44106-44107) - [print(run.content)](misc/print_run_content.md) (lines 44107-44109) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 44109-44139) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 44153-44154) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 44154-44155) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 44155-44156) - [print(chunk.content)](misc/print_chunk_content.md) (lines 44156-44158) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 44158-44188) - [Image Agent](misc/image_agent.md) (lines 44188-44244) - [Agent with Storage](misc/agent_with_storage.md) (lines 44322-44391) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 44391-44426) - [Agent that returns a structured output](misc/agent_that_returns_a_structured_output.md) (lines 44426-44433) - [Run the agent synchronously](misc/run_the_agent_synchronously.md) (lines 44433-44467) - [null](misc/null.md) (lines 44518-44572) - [null](misc/null.md) (lines 44572-44631) - [null](misc/null.md) (lines 44631-44688) - [null](misc/null.md) (lines 44688-44755) - [null](misc/null.md) (lines 44755-44812) - [Basic Agent](misc/basic_agent.md) (lines 44812-44862) - [null](misc/null.md) (lines 44862-44912) - [null](misc/null.md) (lines 44912-44967) - [null](misc/null.md) (lines 44967-45028) - [null](misc/null.md) (lines 45028-45081) - [Basic Agent](misc/basic_agent.md) (lines 45081-45102) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 45102-45103) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 45103-45104) - [print(run.content)](misc/print_run_content.md) (lines 45104-45106) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 45106-45141) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 45162-45163) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 45163-45164) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 45164-45165) - [print(chunk.content)](misc/print_chunk_content.md) (lines 45165-45167) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 45167-45202) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 45202-45253) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 45253-45254) - [json_mode_response: RunResponse = json_mode_agent.run("New York")](misc/json_mode_response_runresponse_json_mode_agent_run_new_york.md) (lines 45254-45255) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 45255-45291) - [Basic Agent](misc/basic_agent.md) (lines 45346-45364) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 45364-45399) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 45417-45452) - [Agent with Storage](misc/agent_with_storage.md) (lines 45526-45598) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 45598-45633) - [Agent that uses a structured output](misc/agent_that_uses_a_structured_output.md) (lines 45633-45675) - [Basic Agent](misc/basic_agent.md) (lines 45728-45741) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 45741-45742) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 45742-45743) - [print(run.content)](misc/print_run_content.md) (lines 45743-45745) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 45745-45780) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 45793-45794) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 45794-45795) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 45795-45796) - [print(chunk.content)](misc/print_chunk_content.md) (lines 45796-45798) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 45798-45833) - [Basic Agent](misc/basic_agent.md) (lines 45886-45899) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 45899-45900) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 45900-45901) - [print(run.content)](misc/print_run_content.md) (lines 45901-45903) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 45903-45940) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 45954-45955) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 45955-45956) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 45956-45957) - [print(chunk.content)](misc/print_chunk_content.md) (lines 45957-45959) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 45959-45996) - [Image Agent](misc/image_agent.md) (lines 45996-46055) - [Set Ollama Client](misc/set_ollama_client.md) (lines 46136-46155) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 46155-46192) - [Agent with Storage](misc/agent_with_storage.md) (lines 46192-46264) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 46264-46299) - [Agent that returns a structured output](misc/agent_that_returns_a_structured_output.md) (lines 46299-46306) - [Run the agent synchronously](misc/run_the_agent_synchronously.md) (lines 46306-46310) - [Run the agent asynchronously](misc/run_the_agent_asynchronously.md) (lines 46310-46351) - [Audio Input Agent](misc/audio_input_agent.md) (lines 46405-46418) - [Fetch the audio file and convert it to a base64 encoded string](misc/fetch_the_audio_file_and_convert_it_to_a_base64_encoded_string.md) (lines 46418-46424) - [Provide the agent with the audio file and get result as text](misc/provide_the_agent_with_the_audio_file_and_get_result_as_text.md) (lines 46424-46465) - [Audio Output Agent](misc/audio_output_agent.md) (lines 46465-46478) - [Provide the agent with the audio file and audio configuration and get result as text + audio](misc/provide_the_agent_with_the_audio_file_and_audio_configuration_and_get_result_as_text_audio.md) (lines 46478-46489) - [Save the response audio to a file](misc/save_the_response_audio_to_a_file.md) (lines 46489-46527) - [Basic Agent](misc/basic_agent.md) (lines 46527-46540) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 46540-46541) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 46541-46542) - [print(run.content)](misc/print_run_content.md) (lines 46542-46544) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 46544-46581) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 46595-46596) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 46596-46597) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 46597-46598) - [print(chunk.content)](misc/print_chunk_content.md) (lines 46598-46600) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 46600-46635) - [Generate Images](misc/generate_images.md) (lines 46635-46696) - [Image Agent](misc/image_agent.md) (lines 46696-46757) - [Agent with Reasoning Effort](misc/agent_with_reasoning_effort.md) (lines 46832-46885) - [Agent with Storage](misc/agent_with_storage.md) (lines 46885-46955) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 46955-46990) - [Agent that uses JSON mode](misc/agent_that_uses_json_mode.md) (lines 46990-46998) - [Agent that uses structured outputs](misc/agent_that_uses_structured_outputs.md) (lines 46998-47006) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 47006-47007) - [json_mode_response: RunResponse = json_mode_agent.run("New York")](misc/json_mode_response_runresponse_json_mode_agent_run_new_york.md) (lines 47007-47008) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 47008-47009) - [structured_output_response: RunResponse = structured_output_agent.run("New York")](misc/structured_output_response_runresponse_structured_output_agent_run_new_york.md) (lines 47009-47010) - [pprint(structured_output_response.content)](misc/pprint_structured_output_response_content.md) (lines 47010-47047) - [Basic Agent](misc/basic_agent.md) (lines 47099-47112) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 47112-47113) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 47113-47114) - [print(run.content)](misc/print_run_content.md) (lines 47114-47116) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 47116-47153) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 47167-47168) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 47168-47169) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 47169-47170) - [print(chunk.content)](misc/print_chunk_content.md) (lines 47170-47172) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 47172-47207) - [Image Agent](misc/image_agent.md) (lines 47207-47268) - [Image Agent (Bytes Content)](misc/image_agent_bytes_content.md) (lines 47268-47297) - [Read the image file content as bytes](misc/read_the_image_file_content_as_bytes.md) (lines 47297-47340) - [Agent with PDF Input (Local File)](misc/agent_with_pdf_input_local_file.md) (lines 47414-47431) - [Download the file using the download_file function](misc/download_the_file_using_the_download_file_function.md) (lines 47431-47481) - [Agent with PDF Input (URL)](misc/agent_with_pdf_input_url.md) (lines 47481-47539) - [Agent with Storage](misc/agent_with_storage.md) (lines 47539-47609) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 47609-47644) - [Agent that uses JSON mode](misc/agent_that_uses_json_mode.md) (lines 47644-47652) - [Agent that uses structured outputs](misc/agent_that_uses_structured_outputs.md) (lines 47652-47660) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 47660-47661) - [json_mode_response: RunResponse = json_mode_agent.run("New York")](misc/json_mode_response_runresponse_json_mode_agent_run_new_york.md) (lines 47661-47662) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 47662-47663) - [structured_output_response: RunResponse = structured_output_agent.run("New York")](misc/structured_output_response_runresponse_structured_output_agent_run_new_york.md) (lines 47663-47664) - [pprint(structured_output_response.content)](misc/pprint_structured_output_response_content.md) (lines 47664-47701) - [Basic Agent](misc/basic_agent.md) (lines 47753-47766) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 47766-47767) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 47767-47768) - [print(run.content)](misc/print_run_content.md) (lines 47768-47770) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 47770-47805) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 47818-47819) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 47819-47820) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 47820-47821) - [print(chunk.content)](misc/print_chunk_content.md) (lines 47821-47823) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 47823-47858) - [Basic Agent](misc/basic_agent.md) (lines 47858-47873) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 47873-47874) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 47874-47875) - [print(run.content)](misc/print_run_content.md) (lines 47875-47877) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 47877-47912) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 47928-47929) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 47929-47930) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 47930-47931) - [print(chunk.content)](misc/print_chunk_content.md) (lines 47931-47933) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 47933-47968) - [Agent with Structured Outputs](misc/agent_with_structured_outputs.md) (lines 47968-48003) - [Agent that uses JSON mode](misc/agent_that_uses_json_mode.md) (lines 48003-48010) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 48010-48011) - [json_mode_response: RunResponse = json_mode_agent.run("New York")](misc/json_mode_response_runresponse_json_mode_agent_run_new_york.md) (lines 48011-48012) - [pprint(json_mode_response.content)](misc/pprint_json_mode_response_content.md) (lines 48012-48048) - [Basic Agent](misc/basic_agent.md) (lines 48100-48113) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 48113-48114) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 48114-48115) - [print(run.content)](misc/print_run_content.md) (lines 48115-48117) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 48117-48152) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 48166-48167) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 48167-48168) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 48168-48169) - [print(chunk.content)](misc/print_chunk_content.md) (lines 48169-48171) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 48171-48206) - [Async Agent](misc/async_agent.md) (lines 48258-48305) - [Code Generation](misc/code_generation.md) (lines 48455-48505) - [Share personal details; the agent should remember them.](misc/share_personal_details_the_agent_should_remember_them.md) (lines 48534-48553) - [Ask the agent to recall](misc/ask_the_agent_to_recall.md) (lines 48553-48597) - [Agent with Storage](misc/agent_with_storage.md) (lines 48597-48663) - [Structured Output](misc/structured_output.md) (lines 48663-48738) - [Basic Agent](misc/basic_agent.md) (lines 48841-48854) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 48854-48855) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 48855-48856) - [print(run.content)](misc/print_run_content.md) (lines 48856-48858) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 48858-48887) - [Async Basic Agent](misc/async_basic_agent.md) (lines 48887-48902) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 48902-48903) - [run: RunResponse = agent.run("Share a 2 sentence horror story")](misc/run_runresponse_agent_run_share_a_2_sentence_horror_story.md) (lines 48903-48904) - [print(run.content)](misc/print_run_content.md) (lines 48904-48906) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 48906-48935) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 48951-48952) - [run_response: Iterator[RunResponseEvent] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponseevent_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 48952-48953) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 48953-48954) - [print(chunk.content)](misc/print_chunk_content.md) (lines 48954-48956) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 48956-48985) - [Get the response in a variable](misc/get_the_response_in_a_variable.md) (lines 48999-49000) - [run_response: Iterator[RunResponse] = agent.run("Share a 2 sentence horror story", stream=True)](misc/run_response_iterator_runresponse_agent_run_share_a_2_sentence_horror_story_stream_true.md) (lines 49000-49001) - [for chunk in run_response:](misc/for_chunk_in_run_response.md) (lines 49001-49002) - [print(chunk.content)](misc/print_chunk_content.md) (lines 49002-49004) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 49004-49033) - [Finance Agent](misc/finance_agent.md) (lines 49033-49113) - [Example usage with detailed market analysis request](misc/example_usage_with_detailed_market_analysis_request.md) (lines 49113-49146) - [Image Agent](misc/image_agent.md) (lines 49146-49201) - [Image Agent with Bytes](misc/image_agent_with_bytes.md) (lines 49201-49230) - [Read the image file content as bytes](misc/read_the_image_file_content_as_bytes.md) (lines 49230-49267) - [Live Search Agent](misc/live_search_agent.md) (lines 49325-49375) - [Reasoning Agent](misc/reasoning_agent.md) (lines 49427-49490) - [Structured Output Agent](misc/structured_output_agent.md) (lines 49490-49526) - [Agent that returns a structured output](misc/agent_that_returns_a_structured_output.md) (lines 49526-49561) - [Agentic RAG](misc/agentic_rag.md) (lines 49607-49718) - [Sage: Answer Engine](misc/sage_answer_engine.md) (lines 49718-49830) - [Chess Battle](misc/chess_battle.md) (lines 49830-49911) - [Game Generator](misc/game_generator.md) (lines 49911-50031) - [Run the workflow if the script is executed directly](misc/run_the_workflow_if_the_script_is_executed_directly.md) (lines 50031-50083) - [GeoBuddy](misc/geobuddy.md) (lines 50083-50163) - [SQL Agent](misc/sql_agent.md) (lines 50163-50298) - [Discussion Team](misc/discussion_team.md) (lines 50298-50436) - [Autonomous Startup Team](misc/autonomous_startup_team.md) (lines 50436-50705) - [HackerNews Team](misc/hackernews_team.md) (lines 50705-50801) - [News Agency Team](misc/news_agency_team.md) (lines 50801-50904) - [AI Support Team](misc/ai_support_team.md) (lines 50904-51015) - [Add in the query and the agent redirects it to the appropriate agent](misc/add_in_the_query_and_the_agent_redirects_it_to_the_appropriate_agent.md) (lines 51015-51020) - [customer_support_team.print_response(](misc/customer_support_team_print_response.md) (lines 51020-51021) - ["[Feature Request] Support json schemas in Gemini client in addition to pydantic base model",](misc/feature_request_support_json_schemas_in_gemini_client_in_addition_to_pydantic_base_model.md) (lines 51021-51022) - [stream=True,](misc/stream_true.md) (lines 51022-51023) - [)](misc/.md) (lines 51023-51024) - [customer_support_team.print_response(](misc/customer_support_team_print_response.md) (lines 51024-51025) - ["[Feature Request] Can you please update me on the above feature",](misc/feature_request_can_you_please_update_me_on_the_above_feature.md) (lines 51025-51026) - [stream=True,](misc/stream_true.md) (lines 51026-51027) - [)](misc/.md) (lines 51027-51028) - [customer_support_team.print_response(](misc/customer_support_team_print_response.md) (lines 51028-51029) - [stream=True,](misc/stream_true.md) (lines 51030-51031) - [)](misc/.md) (lines 51031-51061) - [Multi Language Team](misc/multi_language_team.md) (lines 51061-51155) - [Ask "How are you?" in all supported languages](misc/ask_how_are_you_in_all_supported_languages.md) (lines 51155-51156) - [multi_language_team.print_response(](misc/multi_language_team_print_response.md) (lines 51156-51157) - ["How are you?", stream=True # English](misc/how_are_you_stream_true_english.md) (lines 51157-51158) - [)](misc/.md) (lines 51158-51160) - [multi_language_team.print_response(](misc/multi_language_team_print_response.md) (lines 51160-51161) - ["你好吗？", stream=True # Chinese](misc/你好吗_stream_true_chinese.md) (lines 51161-51162) - [)](misc/.md) (lines 51162-51164) - [multi_language_team.print_response(](misc/multi_language_team_print_response.md) (lines 51164-51165) - ["お元気ですか?", stream=True # Japanese](misc/お元気ですか_stream_true_japanese.md) (lines 51165-51166) - [)](misc/.md) (lines 51166-51173) - [multi_language_team.print_response(](misc/multi_language_team_print_response.md) (lines 51173-51174) - ["Wie geht es Ihnen?", stream=True # German](misc/wie_geht_es_ihnen_stream_true_german.md) (lines 51174-51175) - [)](misc/.md) (lines 51175-51178) - [multi_language_team.print_response(](misc/multi_language_team_print_response.md) (lines 51178-51179) - ["Come stai?", stream=True # Italian](misc/come_stai_stream_true_italian.md) (lines 51179-51180) - [)](misc/.md) (lines 51180-51211) - [Team Session State](misc/team_session_state.md) (lines 51211-51227) - [Shopping management team - new layer for handling all shopping list modifications](misc/shopping_management_team_-_new_layer_for_handling_all_shopping_list_modifications.md) (lines 51279-51339) - [Create meal planning subteam](misc/create_meal_planning_subteam.md) (lines 51339-51414) - [Example usage](misc/example_usage.md) (lines 51414-51429) - [Try the meal planning feature](misc/try_the_meal_planning_feature.md) (lines 51429-51469) - [Scenario Testing](misc/scenario_testing.md) (lines 51469-51486) - [Configure Scenario defaults (model for user simulator and judge)](misc/configure_scenario_defaults_model_for_user_simulator_and_judge.md) (lines 51486-51562) - [Agentic Deep Researcher](misc/agentic_deep_researcher.md) (lines 51562-51748) - [Async Hacker News Reporter](misc/async_hacker_news_reporter.md) (lines 51748-51902) - [Blog Post Generator](misc/blog_post_generator.md) (lines 51902-52289) - [Run the workflow if the script is executed directly](misc/run_the_workflow_if_the_script_is_executed_directly.md) (lines 52289-52360) - [Content Creator](misc/content_creator.md) (lines 52360-52383) - [Define the enums](misc/define_the_enums.md) (lines 52383-52392) - [Planner Agents Configuration](misc/planner_agents_configuration.md) (lines 52392-52426) - [Planner Tasks Configuration](misc/planner_tasks_configuration.md) (lines 52426-52651) - [Load environment variables](misc/load_environment_variables.md) (lines 52651-52655) - [Define Pydantic models to structure responses](misc/define_pydantic_models_to_structure_responses.md) (lines 52655-52855) - [Investment Report Generator](misc/investment_report_generator.md) (lines 52855-53050) - [Run the workflow if the script is executed directly](misc/run_the_workflow_if_the_script_is_executed_directly.md) (lines 53050-53112) - [Personalized Email Generator](misc/personalized_email_generator.md) (lines 53112-53182) - [Demo mode](misc/demo_mode.md) (lines 53182-53183) - [- set to True to print email to console](misc/-_set_to_true_to_print_email_to_console.md) (lines 53183-53184) - [- set to False to send to yourself](misc/-_set_to_false_to_send_to_yourself.md) (lines 53184-53188) - [Example leads - Replace with your actual targets](misc/example_leads_-_replace_with_your_actual_targets.md) (lines 53188-53199) - [Updated sender details for an AI analytics company](misc/updated_sender_details_for_an_ai_analytics_company.md) (lines 53199-53601) - [Product Manager](misc/product_manager.md) (lines 53601-53768) - [Create the workflow](misc/create_the_workflow.md) (lines 53768-53786) - [Run workflow](misc/run_workflow.md) (lines 53786-53845) - [Startup Idea Validator](misc/startup_idea_validator.md) (lines 53845-54097) - [Run the workflow if the script is executed directly](misc/run_the_workflow_if_the_script_is_executed_directly.md) (lines 54097-54143) - [Team Workflow](misc/team_workflow.md) (lines 54143-54282) - [Function instead of steps](misc/function_instead_of_steps.md) (lines 54282-54305) - [Define agents](misc/define_agents.md) (lines 54305-54319) - [Define research team for complex analysis](misc/define_research_team_for_complex_analysis.md) (lines 54319-54369) - [Create and use workflow](misc/create_and_use_workflow.md) (lines 54369-54393) - [Sequence of functions and agents](misc/sequence_of_functions_and_agents.md) (lines 54393-54417) - [Define agents](misc/define_agents.md) (lines 54417-54471) - [Define research team for complex analysis](misc/define_research_team_for_complex_analysis.md) (lines 54471-54480) - [Create and use workflow](misc/create_and_use_workflow.md) (lines 54480-54510) - [Sequence of steps](misc/sequence_of_steps.md) (lines 54510-54534) - [Define agents](misc/define_agents.md) (lines 54534-54548) - [Define research team for complex analysis](misc/define_research_team_for_complex_analysis.md) (lines 54548-54565) - [Define steps](misc/define_steps.md) (lines 54565-54576) - [Create and use workflow](misc/create_and_use_workflow.md) (lines 54576-54601) - [Step with function](misc/step_with_function.md) (lines 54601-54624) - [Define agents](misc/define_agents.md) (lines 54624-54639) - [Define research team for complex analysis](misc/define_research_team_for_complex_analysis.md) (lines 54639-54710) - [Define steps using different executor types](misc/define_steps_using_different_executor_types.md) (lines 54710-54723) - [Define and use examples](misc/define_and_use_examples.md) (lines 54723-54750) - [Workflow using steps](misc/workflow_using_steps.md) (lines 54750-54771) - [Define agents for different tasks](misc/define_agents_for_different_tasks.md) (lines 54771-54791) - [Define individual steps](misc/define_individual_steps.md) (lines 54791-54810) - [Create a Steps sequence that chains these above steps together](misc/create_a_steps_sequence_that_chains_these_above_steps_together.md) (lines 54810-54817) - [Create and use workflow](misc/create_and_use_workflow.md) (lines 54817-54836) - [Workflow using Steps with Nested Pattern](misc/workflow_using_steps_with_nested_pattern.md) (lines 54836-54857) - [Define agents for different tasks](misc/define_agents_for_different_tasks.md) (lines 54857-54897) - [Define individual steps](misc/define_individual_steps.md) (lines 54897-54905) - [Condition evaluator function](misc/condition_evaluator_function.md) (lines 54905-54922) - [Define parallel research steps](misc/define_parallel_research_steps.md) (lines 54922-54935) - [Define content preparation step](misc/define_content_preparation_step.md) (lines 54935-54954) - [Create a Steps sequence with a Condition containing Parallel steps](misc/create_a_steps_sequence_with_a_condition_containing_parallel_steps.md) (lines 54954-54980) - [Create and use workflow](misc/create_and_use_workflow.md) (lines 54980-54997) - [Condition and Parallel Steps Workflow](misc/condition_and_parallel_steps_workflow.md) (lines 54997-55024) - [=== AGENTS ===](misc/agents.md) (lines 55024-55048) - [=== RESEARCH STEPS ===](misc/research_steps.md) (lines 55048-55080) - [=== CONDITION EVALUATORS ===](misc/condition_evaluators.md) (lines 55080-55174) - [Condition steps workflow](misc/condition_steps_workflow.md) (lines 55174-55194) - [=== BASIC AGENTS ===](misc/basic_agents.md) (lines 55194-55217) - [=== CONDITION EVALUATOR ===](misc/condition_evaluator.md) (lines 55217-55244) - [=== WORKFLOW STEPS ===](misc/workflow_steps.md) (lines 55244-55257) - [Conditional fact-checking step](misc/conditional_fact-checking_step.md) (lines 55257-55270) - [=== BASIC LINEAR WORKFLOW ===](misc/basic_linear_workflow.md) (lines 55270-55309) - [Condition with list of steps](misc/condition_with_list_of_steps.md) (lines 55309-55332) - [=== AGENTS ===](misc/agents.md) (lines 55332-55350) - [Additional agents for multi-step condition](misc/additional_agents_for_multi-step_condition.md) (lines 55350-55361) - [=== RESEARCH STEPS ===](misc/research_steps.md) (lines 55361-55374) - [=== MULTI-STEP CONDITION STEPS ===](misc/multi-step_condition_steps.md) (lines 55374-55393) - [=== FINAL STEPS ===](misc/final_steps.md) (lines 55393-55401) - [=== CONDITION EVALUATORS ===](misc/condition_evaluators.md) (lines 55401-55476) - [Loop Steps Workflow](misc/loop_steps_workflow.md) (lines 55476-55498) - [Create agents for research](misc/create_agents_for_research.md) (lines 55498-55514) - [Create research steps](misc/create_research_steps.md) (lines 55514-55534) - [End condition function](misc/end_condition_function.md) (lines 55534-55556) - [Create workflow with loop](misc/create_workflow_with_loop.md) (lines 55556-55585) - [Loop with Parallel Steps Workflow](misc/loop_with_parallel_steps_workflow.md) (lines 55585-55607) - [Create agents for research](misc/create_agents_for_research.md) (lines 55607-55630) - [Create research steps](misc/create_research_steps.md) (lines 55630-55643) - [Create analysis steps](misc/create_analysis_steps.md) (lines 55643-55663) - [End condition function](misc/end_condition_function.md) (lines 55663-55689) - [Create workflow with loop containing parallel steps](misc/create_workflow_with_loop_containing_parallel_steps.md) (lines 55689-55728) - [Parallel Steps Workflow](misc/parallel_steps_workflow.md) (lines 55728-55748) - [Create agents](misc/create_agents.md) (lines 55748-55753) - [Create individual steps](misc/create_individual_steps.md) (lines 55753-55759) - [Create workflow with direct execution](misc/create_workflow_with_direct_execution.md) (lines 55759-55779) - [Conditional Branching Workflow](misc/conditional_branching_workflow.md) (lines 55779-55803) - [Define the research agents](misc/define_the_research_agents.md) (lines 55803-55821) - [Create the research steps](misc/create_the_research_steps.md) (lines 55821-55841) - [Now returns Step(s) to execute](misc/now_returns_step_s_to_execute.md) (lines 55841-55904) - [Router with Loop Steps Workflow](misc/router_with_loop_steps_workflow.md) (lines 55904-55928) - [Define the research agents](misc/define_the_research_agents.md) (lines 55928-55946) - [Create the research steps](misc/create_the_research_steps.md) (lines 55946-55965) - [End condition function for the loop](misc/end_condition_function_for_the_loop.md) (lines 55965-55988) - [Create a Loop step for deep tech research](misc/create_a_loop_step_for_deep_tech_research.md) (lines 55988-55997) - [Router function that selects between simple web research or deep tech research loop](misc/router_function_that_selects_between_simple_web_research_or_deep_tech_research_loop.md) (lines 55997-56065) - [Selector for Image Video Generation Pipelines](misc/selector_for_image_video_generation_pipelines.md) (lines 56065-56099) - [Define the structured message data](misc/define_the_structured_message_data.md) (lines 56099-56109) - [Define specialized agents for different media types](misc/define_specialized_agents_for_different_media_types.md) (lines 56109-56162) - [Define steps for image pipeline](misc/define_steps_for_image_pipeline.md) (lines 56162-56175) - [Define steps for video pipeline](misc/define_steps_for_video_pipeline.md) (lines 56175-56188) - [Define the two distinct pipelines](misc/define_the_two_distinct_pipelines.md) (lines 56188-56231) - [Usage examples](misc/usage_examples.md) (lines 56231-56278) - [Access Multiple Previous Steps Output](misc/access_multiple_previous_steps_output.md) (lines 56278-56303) - [Define the research agents](misc/define_the_research_agents.md) (lines 56303-56321) - [Create the research steps](misc/create_the_research_steps.md) (lines 56321-56334) - [Custom function step that has access to ALL previous step outputs](misc/custom_function_step_that_has_access_to_all_previous_step_outputs.md) (lines 56334-56380) - [Final reasoning step using reasoning agent](misc/final_reasoning_step_using_reasoning_agent.md) (lines 56380-56407) - [Background Execution Poll](misc/background_execution_poll.md) (lines 56407-56427) - [Define agents](misc/define_agents.md) (lines 56427-56441) - [Define research team for complex analysis](misc/define_research_team_for_complex_analysis.md) (lines 56441-56458) - [Define steps](misc/define_steps.md) (lines 56458-56536) - [Early Stop a Workflow](misc/early_stop_a_workflow.md) (lines 56536-56556) - [Create agents with more specific validation criteria](misc/create_agents_with_more_specific_validation_criteria.md) (lines 56556-56605) - [Create workflow with conditional early termination](misc/create_workflow_with_conditional_early_termination.md) (lines 56605-56638) - [Step with Function using Additional Data](misc/step_with_function_using_additional_data.md) (lines 56638-56661) - [Define agents](misc/define_agents.md) (lines 56661-56676) - [Define research team for complex analysis](misc/define_research_team_for_complex_analysis.md) (lines 56676-56766) - [Define steps using different executor types](misc/define_steps_using_different_executor_types.md) (lines 56766-56779) - [Define and use examples](misc/define_and_use_examples.md) (lines 56779-56813) - [Store Events and Events to Skip in a Workflow](misc/store_events_and_events_to_skip_in_a_workflow.md) (lines 56813-56846) - [Define agents for different tasks](misc/define_agents_for_different_tasks.md) (lines 56846-56932) - [Print stored events in a nice format](misc/print_stored_events_in_a_nice_format.md) (lines 56932-56935) - [------------------------------------------------------------------------------------------------ #](misc/------------------------------------------------------------------------------------------------.md) (lines 56935-56936) - [------------------------------------------------------------------------------------------------ #](misc/------------------------------------------------------------------------------------------------.md) (lines 56936-56938) - [Example 2: Parallel Primitive with Event Storage](misc/example_2_parallel_primitive_with_event_storage.md) (lines 56938-56980) - [null](misc/null.md) (lines 56980-57008) - [Define structured models for each step](misc/define_structured_models_for_each_step.md) (lines 57008-57056) - [Define agents with response models](misc/define_agents_with_response_models.md) (lines 57056-57097) - [Define steps](misc/define_steps.md) (lines 57097-57113) - [Create workflow](misc/create_workflow.md) (lines 57113-57137) - [Blog Post Generator](misc/blog_post_generator.md) (lines 57137-57187) - [--- Response Models ---](misc/---_response_models_---.md) (lines 57187-57212) - [--- Agents ---](misc/---_agents_---.md) (lines 57212-57341) - [--- Helper Functions ---](misc/---_helper_functions_---.md) (lines 57341-57503) - [--- Main Execution Function ---](misc/---_main_execution_function_---.md) (lines 57503-57593) - [--- Workflow Definition ---](misc/---_workflow_definition_---.md) (lines 57593-57665) - [Employee Recruiter](misc/employee_recruiter.md) (lines 57665-57736) - [--- Response models ---](misc/---_response_models_---.md) (lines 57736-57756) - [--- PDF utility ---](misc/---_pdf_utility_---.md) (lines 57756-57768) - [--- Agents ---](misc/---_agents_---.md) (lines 57801-57850) - [--- Execution function ---](misc/---_execution_function_---.md) (lines 57850-57991) - [--- Workflow definition ---](misc/---_workflow_definition_---.md) (lines 57991-58049) - [Startup Idea Validator](misc/startup_idea_validator.md) (lines 58049-58095) - [--- Response models ---](misc/---_response_models_---.md) (lines 58095-58136) - [--- Agents ---](misc/---_agents_---.md) (lines 58136-58202) - [--- Execution function ---](misc/---_execution_function_---.md) (lines 58202-58408) - [--- Workflow definition ---](misc/---_workflow_definition_---.md) (lines 58408-58460) - [When to use a Workflow vs a Team in Agno](misc/when_to_use_a_workflow_vs_a_team_in_agno.md) (lines 58460-58502) - [Command line authentication](misc/command_line_authentication.md) (lines 58502-58530) - [Connecting to Tableplus](misc/connecting_to_tableplus.md) (lines 58530-58574) - [Could Not Connect To Docker](misc/could_not_connect_to_docker.md) (lines 58574-58634) - [Setting Environment Variables](misc/setting_environment_variables.md) (lines 58634-58755) - [OpenAI Key Request While Using Other Models](misc/openai_key_request_while_using_other_models.md) (lines 58838-58863) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 58863-58880) - [Embed sentence in database](misc/embed_sentence_in_database.md) (lines 58880-58883) - [Print the embeddings and their dimensions](misc/print_the_embeddings_and_their_dimensions.md) (lines 58883-58887) - [Playground Connection Issues](misc/playground_connection_issues.md) (lines 58901-58971) - [Structured outputs](misc/structured_outputs.md) (lines 58971-59051) - [Tokens-per-minute rate limiting](misc/tokens-per-minute_rate_limiting.md) (lines 59051-59081) - [null](misc/null.md) (lines 59081-59086) - [null](misc/null.md) (lines 59192-59197) - [null](misc/null.md) (lines 59277-59282) - [Authenticate with Agno Platform](misc/authenticate_with_agno_platform.md) (lines 59407-59445) - [Contributing to Agno](misc/contributing_to_agno.md) (lines 59445-59491) - [Install & Setup](misc/install_setup.md) (lines 59491-59559) - [Migrate from Phidata to Agno](misc/migrate_from_phidata_to_agno.md) (lines 59559-59752) - [General commands](misc/general_commands.md) (lines 59752-59762) - [Workspace Management](misc/workspace_management.md) (lines 59762-59803) - [What is Agno?](misc/what_is_agno.md) (lines 59803-59839) - [Getting Started](misc/getting_started.md) (lines 59839-59859) - [Why Agno?](misc/why_agno.md) (lines 59859-59874) - [Dive deeper](misc/dive_deeper.md) (lines 59874-59929) - [You can also use `https://docs.agno.com/llms-full.txt` for the full documentation](misc/you_can_also_use_https_docs_agno_com_llms-full_txt_for_the_full_documentation.md) (lines 60039-60051) - [Community & Support](misc/community_support.md) (lines 60168-60192) - [Monitoring & Debugging](misc/monitoring_debugging.md) (lines 60192-60197) - [Monitoring](misc/monitoring.md) (lines 60197-60281) - [Multi Agent Systems](misc/multi_agent_systems.md) (lines 60281-60462) - [Playground](misc/playground.md) (lines 60462-60594) - [Implementing a Custom Retriever](misc/implementing_a_custom_retriever.md) (lines 60991-61014) - [---------------------------------------------------------](misc/---------------------------------------------------------.md) (lines 61014-61015) - [Define the embedder](misc/define_the_embedder.md) (lines 61016-61018) - [Initialize vector database connection](misc/initialize_vector_database_connection.md) (lines 61018-61020) - [---------------------------------------------------------](misc/---------------------------------------------------------.md) (lines 61029-61031) - [Define the custom retriever](misc/define_the_custom_retriever.md) (lines 61031-61032) - [This is the function that the agent will use to retrieve documents](misc/this_is_the_function_that_the_agent_will_use_to_retrieve_documents.md) (lines 61032-61096) - [---------------------------------------------------------](misc/---------------------------------------------------------.md) (lines 61096-61097) - [---------------------------------------------------------](misc/---------------------------------------------------------.md) (lines 61104-61106) - [Define the custom async retriever](misc/define_the_custom_async_retriever.md) (lines 61106-61168) - [Hybrid Search- Combining Keyword and Vector Search](misc/hybrid_search-_combining_keyword_and_vector_search.md) (lines 61332-61390) - [Database URL](misc/database_url.md) (lines 61390-61393) - [Initialize hybrid vector DB](misc/initialize_hybrid_vector_db.md) (lines 61393-61400) - [Load the data into the DB (first-time setup)](misc/load_the_data_into_the_db_first-time_setup.md) (lines 61406-61409) - [Run a hybrid search query](misc/run_a_hybrid_search_query.md) (lines 61409-61419) - [-*- Get the vectordb](misc/-_-_get_the_vectordb.md) (lines 61689-61691) - [-*- Create a retriever from the vector store](misc/-_-_create_a_retriever_from_the_vector_store.md) (lines 61691-61694) - [Use the agent to ask a question and print a response.](misc/use_the_agent_to_ask_a_question_and_print_a_response.md) (lines 61850-61868) - [Agentic Search](misc/agentic_search.md) (lines 62348-62395) - [User Memories](misc/user_memories.md) (lines 62891-62916) - [UserId for the memories](misc/userid_for_the_memories.md) (lines 62916-62918) - [Initialize storage](misc/initialize_storage.md) (lines 62927-62930) - [Initialize Agent](misc/initialize_agent.md) (lines 62930-62985) - [No need to set the model, it gets set to the model of the agent](misc/no_need_to_set_the_model_it_gets_set_to_the_model_of_the_agent.md) (lines 62985-62988) - [Print the memories for the user](misc/print_the_memories_for_the_user.md) (lines 63010-63015) - [Send a message to the agent that removes all memories for the user](misc/send_a_message_to_the_agent_that_removes_all_memories_for_the_user.md) (lines 63015-63039) - [Send a history of messages and add memories](misc/send_a_history_of_messages_and_add_memories.md) (lines 63139-63182) - [Search for memories related to the query](misc/search_for_memories_related_to_the_query.md) (lines 63223-63242) - [Retrieve memories (these will be loaded from the database)](misc/retrieve_memories_these_will_be_loaded_from_the_database.md) (lines 63316-63329) - [Add user memories](misc/add_user_memories.md) (lines 63338-63348) - [Create agent storage](misc/create_agent_storage.md) (lines 63370-63376) - [Clear all memories](misc/clear_all_memories.md) (lines 63400-63409) - [AI/ML API](misc/ai_ml_api.md) (lines 63409-63455) - [Anthropic Claude](misc/anthropic_claude.md) (lines 63455-63530) - [AWS Bedrock](misc/aws_bedrock.md) (lines 63530-63593) - [AWS Claude](misc/aws_claude.md) (lines 63593-63653) - [Azure AI Foundry](misc/azure_ai_foundry.md) (lines 63653-63712) - [Azure OpenAI](misc/azure_openai.md) (lines 63712-63776) - [Cerebras](misc/cerebras.md) (lines 63776-63811) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 63811-63853) - [Cerebras OpenAI](misc/cerebras_openai.md) (lines 63853-63884) - [Print the response in the terminal](misc/print_the_response_in_the_terminal.md) (lines 63884-63905) - [Cohere](misc/cohere.md) (lines 63905-63967) - [Models Compatibility](misc/models_compatibility.md) (lines 63967-63975) - [DeepInfra](misc/deepinfra.md) (lines 63975-64035) - [DeepSeek](misc/deepseek.md) (lines 64035-64090) - [Fireworks](misc/fireworks.md) (lines 64090-64144) - [Gemini](misc/gemini.md) (lines 64144-64294) - [Groq](misc/groq.md) (lines 64294-64354) - [HuggingFace](misc/huggingface.md) (lines 64354-64411) - [IBM WatsonX](misc/ibm_watsonx.md) (lines 64411-64474) - [What are Models?](misc/what_are_models.md) (lines 64474-64538) - [LangDB](misc/langdb.md) (lines 64538-64591) - [LiteLLM](misc/litellm.md) (lines 64591-64608) - [Install required packages](misc/install_required_packages.md) (lines 64608-64629) - [Create an agent with GPT-4o](misc/create_an_agent_with_gpt-4o.md) (lines 64629-64638) - [Get a response](misc/get_a_response.md) (lines 64638-64682) - [LiteLLM OpenAI](misc/litellm_openai.md) (lines 64682-64744) - [LM Studio](misc/lm_studio.md) (lines 64744-64794) - [Meta](misc/meta.md) (lines 64794-64858) - [Mistral](misc/mistral.md) (lines 64858-64925) - [Nebius](misc/nebius.md) (lines 64925-64978) - [Nvidia](misc/nvidia.md) (lines 64978-65031) - [Ollama](misc/ollama.md) (lines 65031-65093) - [OpenAI](misc/openai.md) (lines 65093-65159) - [OpenAI Like](misc/openai_like.md) (lines 65159-65194) - [OpenAI Responses](misc/openai_responses.md) (lines 65194-65251) - [OpenRouter](misc/openrouter.md) (lines 65251-65304) - [Perplexity](misc/perplexity.md) (lines 65304-65353) - [Sambanova](misc/sambanova.md) (lines 65353-65398) - [Together](misc/together.md) (lines 65398-65452) - [Vercel v0](misc/vercel_v0.md) (lines 65452-65506) - [vLLM](misc/vllm.md) (lines 65506-65589) - [xAI](misc/xai.md) (lines 65589-65666) - [AgentOps](misc/agentops.md) (lines 65666-65700) - [Initialize AgentOps](misc/initialize_agentops.md) (lines 65700-65703) - [Create and run an agent](misc/create_and_run_an_agent.md) (lines 65703-65707) - [Print the response](misc/print_the_response.md) (lines 65707-65720) - [Arize](misc/arize.md) (lines 65720-65767) - [Set environment variables for Arize Phoenix](misc/set_environment_variables_for_arize_phoenix.md) (lines 65767-65771) - [Configure the Phoenix tracer](misc/configure_the_phoenix_tracer.md) (lines 65771-65777) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 65777-65786) - [Use the agent](misc/use_the_agent.md) (lines 65786-65812) - [Set the local collector endpoint](misc/set_the_local_collector_endpoint.md) (lines 65812-65815) - [Configure the Phoenix tracer](misc/configure_the_phoenix_tracer.md) (lines 65815-65821) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 65821-65830) - [Use the agent](misc/use_the_agent.md) (lines 65830-65842) - [Atla](misc/atla.md) (lines 65842-65877) - [Step 1: Configure Atla](misc/step_1_configure_atla.md) (lines 65877-65880) - [Step 2: Create your Agno agent](misc/step_2_create_your_agno_agent.md) (lines 65880-65889) - [Step 3: Instrument and execute](misc/step_3_instrument_and_execute.md) (lines 65889-65902) - [OpenTelemetry](misc/opentelemetry.md) (lines 65902-65932) - [LangDB](misc/langdb.md) (lines 65932-65981) - [Initialize LangDB tracing - must be called before creating agents](misc/initialize_langdb_tracing_-_must_be_called_before_creating_agents.md) (lines 65981-65988) - [Create agent with LangDB model (uses environment variables automatically)](misc/create_agent_with_langdb_model_uses_environment_variables_automatically.md) (lines 65988-65996) - [Use the agent](misc/use_the_agent.md) (lines 65996-66015) - [Research Agent](misc/research_agent.md) (lines 66015-66023) - [Financial Analysis Agent](misc/financial_analysis_agent.md) (lines 66023-66031) - [Coordinated Team](misc/coordinated_team.md) (lines 66031-66044) - [Execute team workflow](misc/execute_team_workflow.md) (lines 66044-66079) - [Langfuse](misc/langfuse.md) (lines 66079-66131) - [Set environment variables for Langfuse](misc/set_environment_variables_for_langfuse.md) (lines 66131-66138) - [Configure the tracer provider](misc/configure_the_tracer_provider.md) (lines 66138-66143) - [Start instrumenting agno](misc/start_instrumenting_agno.md) (lines 66143-66146) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 66146-66155) - [Use the agent](misc/use_the_agent.md) (lines 66155-66175) - [Set environment variables for Langfuse](misc/set_environment_variables_for_langfuse.md) (lines 66175-66182) - [Configure the tracer provider](misc/configure_the_tracer_provider.md) (lines 66182-66187) - [Initialize OpenLIT instrumentation](misc/initialize_openlit_instrumentation.md) (lines 66187-66191) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 66191-66199) - [Use the agent](misc/use_the_agent.md) (lines 66199-66214) - [LangSmith](misc/langsmith.md) (lines 66214-66265) - [Set the endpoint and headers for LangSmith](misc/set_the_endpoint_and_headers_for_langsmith.md) (lines 66265-66272) - [Configure the tracer provider](misc/configure_the_tracer_provider.md) (lines 66272-66279) - [Start instrumenting agno](misc/start_instrumenting_agno.md) (lines 66279-66282) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 66282-66291) - [Use the agent](misc/use_the_agent.md) (lines 66291-66303) - [Langtrace](misc/langtrace.md) (lines 66303-66346) - [Initialize Langtrace](misc/initialize_langtrace.md) (lines 66346-66349) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 66349-66358) - [Use the agent](misc/use_the_agent.md) (lines 66358-66370) - [LangWatch](misc/langwatch.md) (lines 66370-66408) - [Initialize LangWatch and instrument Agno](misc/initialize_langwatch_and_instrument_agno.md) (lines 66408-66433) - [Weave](misc/weave.md) (lines 66433-66476) - [Initialize Weave with your project name](misc/initialize_weave_with_your_project_name.md) (lines 66476-66479) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 66479-66482) - [Define a function to run the agent, decorated with weave.op()](misc/define_a_function_to_run_the_agent_decorated_with_weave_op.md) (lines 66482-66487) - [Use the function to log a model call](misc/use_the_function_to_log_a_model_call.md) (lines 66487-66516) - [Set the endpoint and headers for Weave](misc/set_the_endpoint_and_headers_for_weave.md) (lines 66516-66521) - [Configure authentication](misc/configure_authentication.md) (lines 66521-66530) - [Configure the tracer provider](misc/configure_the_tracer_provider.md) (lines 66530-66537) - [Start instrumenting agno](misc/start_instrumenting_agno.md) (lines 66537-66540) - [Create and configure the agent](misc/create_and_configure_the_agent.md) (lines 66540-66549) - [Use the agent](misc/use_the_agent.md) (lines 66549-66563) - [What is Reasoning?](misc/what_is_reasoning.md) (lines 66563-66692) - [Reasoning Agents](misc/reasoning_agents.md) (lines 66692-66858) - [Reasoning Models](misc/reasoning_models.md) (lines 66858-66967) - [Agent](misc/agent.md) (lines 67096-67104) - [RunResponse](misc/runresponse.md) (lines 67104-67266) - [AgentSession](misc/agentsession.md) (lines 67266-67274) - [Arxiv Reader](misc/arxiv_reader.md) (lines 67330-67340) - [Reader](misc/reader.md) (lines 67340-67350) - [CSV Reader](misc/csv_reader.md) (lines 67350-67360) - [CSV URL Reader](misc/csv_url_reader.md) (lines 67360-67370) - [Docx Reader](misc/docx_reader.md) (lines 67370-67380) - [FireCrawl Reader](misc/firecrawl_reader.md) (lines 67380-67390) - [JSON Reader](misc/json_reader.md) (lines 67390-67400) - [PDF Reader](misc/pdf_reader.md) (lines 67400-67410) - [PDF Image Reader](misc/pdf_image_reader.md) (lines 67410-67420) - [PDF Image URL Reader](misc/pdf_image_url_reader.md) (lines 67420-67430) - [PDF URL Reader](misc/pdf_url_reader.md) (lines 67430-67440) - [Text Reader](misc/text_reader.md) (lines 67440-67450) - [Website Reader](misc/website_reader.md) (lines 67450-67460) - [YouTube Reader](misc/youtube_reader.md) (lines 67460-67470) - [Azure OpenAI](misc/azure_openai.md) (lines 67470-67480) - [Cohere](misc/cohere.md) (lines 67480-67490) - [FastEmbed](misc/fastembed.md) (lines 67490-67500) - [Fireworks](misc/fireworks.md) (lines 67500-67510) - [Gemini](misc/gemini.md) (lines 67510-67520) - [Hugging Face](misc/hugging_face.md) (lines 67520-67530) - [Mistral](misc/mistral.md) (lines 67530-67540) - [Ollama](misc/ollama.md) (lines 67540-67550) - [OpenAI](misc/openai.md) (lines 67550-67560) - [Sentence Transformer](misc/sentence_transformer.md) (lines 67560-67570) - [Together](misc/together.md) (lines 67570-67580) - [VoyageAI](misc/voyageai.md) (lines 67580-67590) - [AI/ML API](misc/ai_ml_api.md) (lines 67794-67808) - [Claude](misc/claude.md) (lines 67808-67818) - [Azure AI Foundry](misc/azure_ai_foundry.md) (lines 67818-67828) - [Azure OpenAI](misc/azure_openai.md) (lines 67828-67838) - [AWS Bedrock](misc/aws_bedrock.md) (lines 67838-67848) - [AWS Bedrock Claude](misc/aws_bedrock_claude.md) (lines 67848-67858) - [Cohere](misc/cohere.md) (lines 67858-67868) - [DeepInfra](misc/deepinfra.md) (lines 67868-67878) - [DeepSeek](misc/deepseek.md) (lines 67878-67888) - [Fireworks](misc/fireworks.md) (lines 67888-67898) - [Gemini](misc/gemini.md) (lines 67898-67908) - [Groq](misc/groq.md) (lines 67908-67918) - [HuggingFace](misc/huggingface.md) (lines 67918-67928) - [IBM WatsonX](misc/ibm_watsonx.md) (lines 67928-67938) - [InternLM](misc/internlm.md) (lines 67938-67948) - [Meta](misc/meta.md) (lines 67948-67958) - [Mistral](misc/mistral.md) (lines 67958-67968) - [Model](misc/model.md) (lines 67968-67978) - [Nebius](misc/nebius.md) (lines 67978-67988) - [Nvidia](misc/nvidia.md) (lines 67988-67998) - [Ollama](misc/ollama.md) (lines 67998-68008) - [OpenAI](misc/openai.md) (lines 68018-68028) - [OpenAI Like](misc/openai_like.md) (lines 68028-68038) - [OpenRouter](misc/openrouter.md) (lines 68038-68048) - [Perplexity](misc/perplexity.md) (lines 68048-68058) - [Sambanova](misc/sambanova.md) (lines 68058-68068) - [Together](misc/together.md) (lines 68068-68078) - [Vercel v0](misc/vercel_v0.md) (lines 68078-68088) - [xAI](misc/xai.md) (lines 68088-68098) - [Cohere Reranker](misc/cohere_reranker.md) (lines 68098-68106) - [DynamoDB](misc/dynamodb.md) (lines 68106-68116) - [JSON](misc/json.md) (lines 68116-68126) - [MongoDB](misc/mongodb.md) (lines 68126-68136) - [MySQL](misc/mysql.md) (lines 68136-68146) - [PostgreSQL](misc/postgresql.md) (lines 68146-68156) - [SingleStore](misc/singlestore.md) (lines 68156-68166) - [SQLite](misc/sqlite.md) (lines 68166-68176) - [YAML](misc/yaml.md) (lines 68176-68186) - [Team Session](misc/team_session.md) (lines 68186-68194) - [Team](misc/team.md) (lines 68194-68202) - [TeamRunResponse](misc/teamrunresponse.md) (lines 68202-68391) - [Cassandra](misc/cassandra.md) (lines 68391-68399) - [ChromaDb](misc/chromadb.md) (lines 68399-68407) - [Clickhouse](misc/clickhouse.md) (lines 68407-68415) - [Couchbase](misc/couchbase.md) (lines 68415-68423) - [LanceDb](misc/lancedb.md) (lines 68423-68431) - [Milvus](misc/milvus.md) (lines 68431-68439) - [MongoDb](misc/mongodb.md) (lines 68439-68447) - [PgVector](misc/pgvector.md) (lines 68447-68455) - [Pinecone](misc/pinecone.md) (lines 68455-68463) - [Qdrant](misc/qdrant.md) (lines 68463-68471) - [SingleStore](misc/singlestore.md) (lines 68471-68479) - [SurrealDB](misc/surrealdb.md) (lines 68479-68487) - [Weaviate](misc/weaviate.md) (lines 68487-68495) - [MongoDB Workflow Storage](misc/mongodb_workflow_storage.md) (lines 68495-68503) - [Postgres Workflow Storage](misc/postgres_workflow_storage.md) (lines 68503-68511) - [SQLite Workflow Storage](misc/sqlite_workflow_storage.md) (lines 68511-68519) - [Workflow](misc/workflow.md) (lines 68519-68527) - [Conditional Steps](misc/conditional_steps.md) (lines 68527-68535) - [Loop Steps](misc/loop_steps.md) (lines 68535-68543) - [Parallel Steps](misc/parallel_steps.md) (lines 68543-68551) - [Router Steps](misc/router_steps.md) (lines 68551-68559) - [StepInput](misc/stepinput.md) (lines 68559-68567) - [StepOutput](misc/stepoutput.md) (lines 68567-68575) - [Steps](misc/steps.md) (lines 68575-68583) - [Workflows](misc/workflows.md) (lines 68583-68591) - [WorkflowRunResponse](misc/workflowrunresponse.md) (lines 68591-68675) - [DynamoDB Storage](misc/dynamodb_storage.md) (lines 68675-68689) - [AWS Credentials](misc/aws_credentials.md) (lines 68689-68704) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 68704-68717) - [What is Storage?](misc/what_is_storage.md) (lines 68717-68838) - [JSON Storage](misc/json_storage.md) (lines 68838-68947) - [Mongo Storage](misc/mongo_storage.md) (lines 68947-68963) - [Create a storage backend using the Mongo database](misc/create_a_storage_backend_using_the_mongo_database.md) (lines 68963-68970) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 68970-68983) - [MySQL Storage](misc/mysql_storage.md) (lines 68983-69012) - [Create a storage backend using the Postgres database](misc/create_a_storage_backend_using_the_postgres_database.md) (lines 69012-69020) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 69020-69033) - [Postgres Storage](misc/postgres_storage.md) (lines 69033-69063) - [Create a storage backend using the Postgres database](misc/create_a_storage_backend_using_the_postgres_database.md) (lines 69063-69071) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 69071-69084) - [Redis Storage](misc/redis_storage.md) (lines 69084-69106) - [Initialize Redis storage with default local connection](misc/initialize_redis_storage_with_default_local_connection.md) (lines 69106-69113) - [Create agent with Redis storage](misc/create_agent_with_redis_storage.md) (lines 69113-69124) - [Verify storage contents](misc/verify_storage_contents.md) (lines 69124-69145) - [Singlestore Storage](misc/singlestore_storage.md) (lines 69145-69164) - [SingleStore Configuration](misc/singlestore_configuration.md) (lines 69164-69172) - [SingleStore DB URL](misc/singlestore_db_url.md) (lines 69172-69177) - [Create a database engine](misc/create_a_database_engine.md) (lines 69177-69180) - [Create a storage backend using the Singlestore database](misc/create_a_storage_backend_using_the_singlestore_database.md) (lines 69180-69190) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 69190-69203) - [Sqlite Storage](misc/sqlite_storage.md) (lines 69203-69217) - [Create a storage backend using the Sqlite database](misc/create_a_storage_backend_using_the_sqlite_database.md) (lines 69217-69225) - [Add storage to the Agent](misc/add_storage_to_the_agent.md) (lines 69225-69238) - [YAML Storage](misc/yaml_storage.md) (lines 69238-69271) - [Collaborate](misc/collaborate.md) (lines 69271-69429) - [Coordinate](misc/coordinate.md) (lines 69429-69545) - [What are Teams?](misc/what_are_teams.md) (lines 69545-69611) - [The team will remember previous interactions](misc/the_team_will_remember_previous_interactions.md) (lines 69611-69623) - [Setup paths](misc/setup_paths.md) (lines 69653-69658) - [Create individual specialized agents](misc/create_individual_specialized_agents.md) (lines 69804-69818) - [Create a team with these agents](misc/create_a_team_with_these_agents.md) (lines 69818-69828) - [Run the team with a task](misc/run_the_team_with_a_task.md) (lines 69828-69930) - [Create team members](misc/create_team_members.md) (lines 69972-69980) - [Create the team](misc/create_the_team.md) (lines 69980-69988) - [Run the team](misc/run_the_team.md) (lines 69988-69994) - [Route](misc/route.md) (lines 70131-70333) - [This should route to the stock_searcher](misc/this_should_route_to_the_stock_searcher.md) (lines 70333-70339) - [Running your Team](misc/running_your_team.md) (lines 70339-70362) - [Synchronous execution](misc/synchronous_execution.md) (lines 70362-70365) - [Asynchronous execution](misc/asynchronous_execution.md) (lines 70365-70368) - [Stream with intermediate steps](misc/stream_with_intermediate_steps.md) (lines 70384-70504) - [Define agents](misc/define_agents.md) (lines 70504-70530) - [Team State](misc/team_state.md) (lines 70530-70598) - [Create an agent that manages the shopping list](misc/create_an_agent_that_manages_the_shopping_list.md) (lines 70620-70629) - [Create a team with both shared and private state](misc/create_a_team_with_both_shared_and_private_state.md) (lines 70652-70672) - [Example usage](misc/example_usage.md) (lines 70672-70727) - [Structured Output](misc/structured_output.md) (lines 70727-70762) - [Create specialized agents](misc/create_specialized_agents.md) (lines 70762-70790) - [Create team with structured output](misc/create_team_with_structured_output.md) (lines 70790-70852) - [Create research agents](misc/create_research_agents.md) (lines 70852-70867) - [Stream the team response](misc/stream_the_team_response.md) (lines 70878-70895) - [Scenario Testing](misc/scenario_testing.md) (lines 70895-70912) - [Configure Scenario defaults (model for user simulator and judge)](misc/configure_scenario_defaults_model_for_user_simulator_and_judge.md) (lines 70912-70988) - [Team Example](misc/team_example.md) (lines 71125-71190) - [Tool Result Caching](misc/tool_result_caching.md) (lines 71190-71227) - [Writing your own Toolkit](misc/writing_your_own_toolkit.md) (lines 71227-71283) - [Exceptions](misc/exceptions.md) (lines 71283-71333) - [Human in the loop](misc/human_in_the_loop.md) (lines 71333-71380) - [This is the console instance used by the print_response method](misc/this_is_the_console_instance_used_by_the_print_response_method.md) (lines 71380-71381) - [We can use this to stop and restart the live display and ask for user confirmation](misc/we_can_use_this_to_stop_and_restart_the_live_display_and_ask_for_user_confirmation.md) (lines 71381-71448) - [Initialize the agent with a tech-savvy personality and clear instructions](misc/initialize_the_agent_with_a_tech-savvy_personality_and_clear_instructions.md) (lines 71448-71461) - [Hooks](misc/hooks.md) (lines 71461-71648) - [Advanced MCP Usage](misc/advanced_mcp_usage.md) (lines 71720-71767) - [Example usage](misc/example_usage.md) (lines 71767-71874) - [Example usage](misc/example_usage.md) (lines 71874-71881) - [Example usage](misc/example_usage.md) (lines 71992-72018) - [MCP server parameters setup](misc/mcp_server_parameters_setup.md) (lines 72018-72029) - [This is required to start the MCP connection correctly in the FastAPI lifecycle](misc/this_is_required_to_start_the_mcp_connection_correctly_in_the_fastapi_lifecycle.md) (lines 72029-72068) - [Setup the Playground app](misc/setup_the_playground_app.md) (lines 72068-72076) - [Initialize the Playground app with our lifespan logic](misc/initialize_the_playground_app_with_our_lifespan_logic.md) (lines 72076-72091) - [Before exiting](misc/before_exiting.md) (lines 72091-72115) - [SSE Transport](misc/sse_transport.md) (lines 72115-72260) - [Stdio Transport](misc/stdio_transport.md) (lines 72260-72343) - [Example usage](misc/example_usage.md) (lines 72343-72354) - [Streamable HTTP Transport](misc/streamable_http_transport.md) (lines 72354-72526) - [Tool Call Limit](misc/tool_call_limit.md) (lines 73046-73068) - [The first tool call will be performed. The second one will fail gracefully.](misc/the_first_tool_call_will_be_performed_the_second_one_will_fail_gracefully.md) (lines 73068-73082) - [CSV](misc/csv.md) (lines 73082-73151) - [DuckDb](misc/duckdb.md) (lines 73151-73225) - [Mem0](misc/mem0.md) (lines 73225-73260) - [Interact with the Agent to store memories](misc/interact_with_the_agent_to_store_memories.md) (lines 73282-73287) - [Query the stored memories](misc/query_the_stored_memories.md) (lines 73287-73351) - [Pandas](misc/pandas.md) (lines 73351-73362) - [Example: Create a dataframe with sample data and get the first 5 rows](misc/example_create_a_dataframe_with_sample_data_and_get_the_first_5_rows.md) (lines 73365-73398) - [Postgres](misc/postgres.md) (lines 73398-73435) - [Example: Ask the agent to run a SQL query](misc/example_ask_the_agent_to_run_a_sql_query.md) (lines 73447-73486) - [SQL](misc/sql.md) (lines 73486-73580) - [Zep](misc/zep.md) (lines 73580-73610) - [Initialize the Agent](misc/initialize_the_agent.md) (lines 73613-73621) - [Interact with the Agent so that it can learn about the user](misc/interact_with_the_agent_so_that_it_can_learn_about_the_user.md) (lines 73621-73626) - [Allow the memories to sync with Zep database](misc/allow_the_memories_to_sync_with_zep_database.md) (lines 73626-73629) - [Ask the Agent about the user](misc/ask_the_agent_about_the_user.md) (lines 73632-73669) - [Calculator](misc/calculator.md) (lines 73669-73736) - [Docker](misc/docker.md) (lines 73736-73869) - [File](misc/file.md) (lines 73869-73911) - [Python](misc/python.md) (lines 73911-73961) - [Shell](misc/shell.md) (lines 73961-73996) - [Sleep](misc/sleep.md) (lines 73996-74009) - [Create an Agent with the Sleep tool](misc/create_an_agent_with_the_sleep_tool.md) (lines 74009-74012) - [Example 1: Sleep for 2 seconds](misc/example_1_sleep_for_2_seconds.md) (lines 74012-74015) - [Example 2: Sleep for a longer duration](misc/example_2_sleep_for_a_longer_duration.md) (lines 74015-74037) - [Gemini](misc/gemini.md) (lines 74037-74147) - [Groq](misc/groq.md) (lines 74147-74277) - [OpenAI](misc/openai.md) (lines 74277-74431) - [Airflow](misc/airflow.md) (lines 74431-74461) - [Using 'schedule' instead of deprecated 'schedule_interval'](misc/using_schedule_instead_of_deprecated_schedule_interval.md) (lines 74461-74506) - [Apify](misc/apify.md) (lines 74506-74539) - [Use the agent to get website content](misc/use_the_agent_to_get_website_content.md) (lines 74551-74575) - [Search for information and process the results](misc/search_for_information_and_process_the_results.md) (lines 74575-74594) - [Ask the agent to process web content](misc/ask_the_agent_to_process_web_content.md) (lines 74594-74613) - [Find business information in a specific location](misc/find_business_information_in_a_specific_location.md) (lines 74613-74638) - [Get general information and local businesses](misc/get_general_information_and_local_businesses.md) (lines 74638-74669) - [AWS Lambda](misc/aws_lambda.md) (lines 74669-74692) - [Create an Agent with the AWSLambdaTool](misc/create_an_agent_with_the_awslambdatool.md) (lines 74692-74699) - [Example 1: List all Lambda functions](misc/example_1_list_all_lambda_functions.md) (lines 74699-74702) - [Example 2: Invoke a specific Lambda function](misc/example_2_invoke_a_specific_lambda_function.md) (lines 74702-74725) - [AWS SES](misc/aws_ses.md) (lines 74725-74744) - [OR set environment variables manually](misc/or_set_environment_variables_manually.md) (lines 74744-74766) - [Configure email settings](misc/configure_email_settings.md) (lines 74766-74819) - [Cal.com](misc/cal_com.md) (lines 74819-74894) - [Cartesia](misc/cartesia.md) (lines 74894-75006) - [Composio](misc/composio.md) (lines 75006-75055) - [Confluence](misc/confluence.md) (lines 75055-75074) - [or](misc/or.md) (lines 75074-75123) - [Custom API](misc/custom_api.md) (lines 75123-75184) - [Dalle](misc/dalle.md) (lines 75184-75211) - [Create an Agent with the DALL-E tool](misc/create_an_agent_with_the_dall-e_tool.md) (lines 75211-75214) - [Example 1: Generate a basic image with default settings](misc/example_1_generate_a_basic_image_with_default_settings.md) (lines 75214-75217) - [Example 2: Generate an image with custom settings](misc/example_2_generate_an_image_with_custom_settings.md) (lines 75217-75252) - [Daytona](misc/daytona.md) (lines 75252-75363) - [E2B](misc/e2b.md) (lines 75363-75422) - [Example: Generate Fibonacci numbers](misc/example_generate_fibonacci_numbers.md) (lines 75422-75501) - [Eleven Labs](misc/eleven_labs.md) (lines 75501-75530) - [Create an Agent with the ElevenLabs tool](misc/create_an_agent_with_the_elevenlabs_tool.md) (lines 75530-75564) - [Fal](misc/fal.md) (lines 75564-75630) - [Financial Datasets API](misc/financial_datasets_api.md) (lines 75630-75675) - [Get the most recent income statement for Apple](misc/get_the_most_recent_income_statement_for_apple.md) (lines 75675-75724) - [Giphy](misc/giphy.md) (lines 75724-75776) - [Github](misc/github.md) (lines 75776-75847) - [Google Maps](misc/google_maps.md) (lines 75847-75927) - [Google Sheets](misc/google_sheets.md) (lines 75927-76029) - [Google Calendar](misc/google_calendar.md) (lines 76029-76141) - [Jira](misc/jira.md) (lines 76141-76198) - [Linear](misc/linear.md) (lines 76198-76261) - [Lumalabs](misc/lumalabs.md) (lines 76261-76333) - [MLX Transcribe](misc/mlx_transcribe.md) (lines 76333-76374) - [Get audio files from storage/audio directory](misc/get_audio_files_from_storage_audio_directory.md) (lines 76374-76428) - [ModelsLabs](misc/modelslabs.md) (lines 76428-76455) - [Create an Agent with the ModelsLabs tool](misc/create_an_agent_with_the_modelslabs_tool.md) (lines 76455-76483) - [OpenBB](misc/openbb.md) (lines 76570-76584) - [Example usage showing stock analysis](misc/example_usage_showing_stock_analysis.md) (lines 76584-76589) - [Example showing market analysis](misc/example_showing_market_analysis.md) (lines 76589-76594) - [Example showing economic indicators](misc/example_showing_economic_indicators.md) (lines 76594-76624) - [OpenWeather](misc/openweather.md) (lines 76624-76647) - [Get current weather for a location](misc/get_current_weather_for_a_location.md) (lines 76657-76687) - [Replicate](misc/replicate.md) (lines 76687-76754) - [Resend](misc/resend.md) (lines 76754-76807) - [Todoist](misc/todoist.md) (lines 76807-76867) - [Example 1: Create a task](misc/example_1_create_a_task.md) (lines 76867-76872) - [Example 2: Delete a task](misc/example_2_delete_a_task.md) (lines 76872-76879) - [Example 3: Get all tasks](misc/example_3_get_all_tasks.md) (lines 76879-76908) - [Trello](misc/trello.md) (lines 76908-77005) - [Yfinance](misc/yfinance.md) (lines 77038-77104) - [Youtube](misc/youtube.md) (lines 77104-77157) - [Zendesk](misc/zendesk.md) (lines 77157-77210) - [Arxiv](misc/arxiv.md) (lines 77210-77258) - [BaiduSearch](misc/baidusearch.md) (lines 77258-77315) - [Brave Search](misc/brave_search.md) (lines 77315-77371) - [DuckDuckGo](misc/duckduckgo.md) (lines 77371-77421) - [Exa](misc/exa.md) (lines 77421-77514) - [Google Search](misc/google_search.md) (lines 77514-77574) - [Hacker News](misc/hacker_news.md) (lines 77574-77623) - [Linkup](misc/linkup.md) (lines 77623-77674) - [Pubmed](misc/pubmed.md) (lines 77674-77712) - [Searxng](misc/searxng.md) (lines 77712-77725) - [Initialize Searxng with your Searxng instance URL](misc/initialize_searxng_with_your_searxng_instance_url.md) (lines 77725-77734) - [Create an agent with Searxng](misc/create_an_agent_with_searxng.md) (lines 77734-77737) - [Example: Ask the agent to search using Searxng](misc/example_ask_the_agent_to_search_using_searxng.md) (lines 77737-77778) - [Serpapi](misc/serpapi.md) (lines 77778-77829) - [Serper](misc/serper.md) (lines 77829-77920) - [Tavily](misc/tavily.md) (lines 77920-77976) - [Wikipedia](misc/wikipedia.md) (lines 77976-78022) - [Discord](misc/discord.md) (lines 78022-78078) - [Email](misc/email.md) (lines 78078-78131) - [Gmail](misc/gmail.md) (lines 78131-78205) - [Slack](misc/slack.md) (lines 78205-78238) - [Example 1: Send a message to a Slack channel](misc/example_1_send_a_message_to_a_slack_channel.md) (lines 78238-78241) - [Example 2: List all channels in the Slack workspace](misc/example_2_list_all_channels_in_the_slack_workspace.md) (lines 78241-78244) - [Example 3: Get the message history of a specific channel by channel ID](misc/example_3_get_the_message_history_of_a_specific_channel_by_channel_id.md) (lines 78244-78272) - [Telegram](misc/telegram.md) (lines 78272-78297) - [How to get the token and chat_id:](misc/how_to_get_the_token_and_chat_id.md) (lines 78297-78298) - [1. Create a new bot with BotFather on Telegram. https://core.telegram.org/bots/features#creating-a-new-bot](misc/1_create_a_new_bot_with_botfather_on_telegram_https_core_telegram_org_bots_features_creating-a-new-bot.md) (lines 78298-78299) - [2. Get the token from BotFather.](misc/2_get_the_token_from_botfather.md) (lines 78299-78300) - [3. Send a message to the bot.](misc/3_send_a_message_to_the_bot.md) (lines 78300-78301) - [4. Get the chat_id by going to the URL:](misc/4_get_the_chat_id_by_going_to_the_url.md) (lines 78301-78302) - [https://api.telegram.org/bot//getUpdates](misc/https_api_telegram_org_bot_your-bot-token_getupdates.md) (lines 78302-78334) - [Twilio](misc/twilio.md) (lines 78334-78401) - [Webex](misc/webex.md) (lines 78401-78446) - [List all spaces in Webex](misc/list_all_spaces_in_webex.md) (lines 78446-78449) - [Send a message to a Space in Webex](misc/send_a_message_to_a_space_in_webex.md) (lines 78449-78476) - [WhatsApp](misc/whatsapp.md) (lines 78476-78531) - [Example: Send a template message](misc/example_send_a_template_message.md) (lines 78531-78532) - [Note: Replace '''hello_world''' with your actual template name](misc/note_replace_hello_world_with_your_actual_template_name.md) (lines 78532-78533) - [and +91 1234567890 with the recipient's WhatsApp ID](misc/and_91_1234567890_with_the_recipient_s_whatsapp_id.md) (lines 78533-78564) - [X (Twitter)](misc/x_twitter.md) (lines 78564-78609) - [Initialize the X toolkit](misc/initialize_the_x_toolkit.md) (lines 78609-78614) - [Create an agent equipped with X toolkit](misc/create_an_agent_equipped_with_x_toolkit.md) (lines 78614-78626) - [Search for posts](misc/search_for_posts.md) (lines 78626-78629) - [Create and post a tweet](misc/create_and_post_a_tweet.md) (lines 78629-78632) - [Get user timeline](misc/get_user_timeline.md) (lines 78632-78635) - [Reply to a post](misc/reply_to_a_post.md) (lines 78635-78641) - [Get information about a user](misc/get_information_about_a_user.md) (lines 78641-78644) - [Send a direct message](misc/send_a_direct_message.md) (lines 78644-78650) - [Get user profile](misc/get_user_profile.md) (lines 78650-78686) - [Zoom](misc/zoom.md) (lines 78686-78727) - [Create an agent with Zoom capabilities](misc/create_an_agent_with_zoom_capabilities.md) (lines 78734-78737) - [Schedule a meeting](misc/schedule_a_meeting.md) (lines 78737-78781) - [Toolkit Index](misc/toolkit_index.md) (lines 78781-79135) - [AgentQL](misc/agentql.md) (lines 79135-79196) - [BrightData](misc/brightdata.md) (lines 79196-79247) - [Example 1: Scrape a webpage as Markdown](misc/example_1_scrape_a_webpage_as_markdown.md) (lines 79247-79343) - [Browserbase](misc/browserbase.md) (lines 79343-79415) - [Crawl4AI](misc/crawl4ai.md) (lines 79415-79460) - [Firecrawl](misc/firecrawl.md) (lines 79460-79520) - [Jina Reader](misc/jina_reader.md) (lines 79520-79569) - [Newspaper](misc/newspaper.md) (lines 79569-79614) - [Newspaper4k](misc/newspaper4k.md) (lines 79614-79662) - [Oxylabs](misc/oxylabs.md) (lines 79662-79741) - [ScrapeGraph](misc/scrapegraph.md) (lines 79741-79809) - [Spider](misc/spider.md) (lines 79809-79857) - [Azure Cosmos DB MongoDB connection string](misc/azure_cosmos_db_mongodb_connection_string.md) (lines 79926-79943) - [Comment out after first run](misc/comment_out_after_first_run.md) (lines 79943-79946) - [Create and use the agent](misc/create_and_use_the_agent.md) (lines 79946-79963) - [(Optional) Set up your Cassandra DB](misc/optional_set_up_your_cassandra_db.md) (lines 79995-80013) - [Comment out after first run](misc/comment_out_after_first_run.md) (lines 80256-80317) - [Couchbase connection settings](misc/couchbase_connection_settings.md) (lines 80379-80384) - [Create cluster options with authentication](misc/create_cluster_options_with_authentication.md) (lines 80384-80409) - [Wait for the vector index to sync with KV](misc/wait_for_the_vector_index_to_sync_with_kv.md) (lines 80412-80415) - [Create and use the agent](misc/create_and_use_the_agent.md) (lines 80415-80502) - [What are Vector Databases?](misc/what_are_vector_databases.md) (lines 80502-80574) - [LanceDB Vector DB](misc/lancedb_vector_db.md) (lines 80597-80604) - [Create and use the agent](misc/create_and_use_the_agent.md) (lines 80730-80791) - [MongoDB Atlas connection string](misc/mongodb_atlas_connection_string.md) (lines 80813-80831) - [SurrealDB connection parameters](misc/surrealdb_connection_parameters.md) (lines 81388-81395) - [Create a client](misc/create_a_client.md) (lines 81395-81511) - [Create and use the agent](misc/create_and_use_the_agent.md) (lines 81564-81631) - [Advanced](misc/advanced.md) (lines 81631-81654) - [Define the workflow](misc/define_the_workflow.md) (lines 81654-81673) - [Instantiate the workflow](misc/instantiate_the_workflow.md) (lines 81673-81676) - [Run workflow and get the response as an iterator of RunResponse objects](misc/run_workflow_and_get_the_response_as_an_iterator_of_runresponse_objects.md) (lines 81676-81679) - [Print the response](misc/print_the_response.md) (lines 81679-81688) - [Define the workflow](misc/define_the_workflow.md) (lines 81688-81706) - [Instantiate the workflow](misc/instantiate_the_workflow.md) (lines 81706-81709) - [Run workflow and get the response as a RunResponse object](misc/run_workflow_and_get_the_response_as_a_runresponse_object.md) (lines 81709-81712) - [Print the response](misc/print_the_response.md) (lines 81712-81717) - [What are Workflows?](misc/what_are_workflows.md) (lines 81717-82165) - [Run the workflow if the script is executed directly](misc/run_the_workflow_if_the_script_is_executed_directly.md) (lines 82165-82262) - [Workflow State](misc/workflow_state.md) (lines 82262-82274) - [Create the workflow](misc/create_the_workflow.md) (lines 82274-82322) - [Advanced Concepts in Workflows](misc/advanced_concepts_in_workflows.md) (lines 82322-82435) - [Define agents](misc/define_agents.md) (lines 82435-82449) - [Define research team for complex analysis](misc/define_research_team_for_complex_analysis.md) (lines 82449-82466) - [Define steps](misc/define_steps.md) (lines 82466-82561) - [Secure deployment pipeline](misc/secure_deployment_pipeline.md) (lines 82561-82572) - [Test with vulnerable code - workflow stops at security gate](misc/test_with_vulnerable_code_-_workflow_stops_at_security_gate.md) (lines 82572-82621) - [Use in workflow](misc/use_in_workflow.md) (lines 82621-82678) - [Common events you might want to skip](misc/common_events_you_might_want_to_skip.md) (lines 82678-82706) - [store everything](misc/store_everything.md) (lines 82706-82713) - [store only important events](misc/store_only_important_events.md) (lines 82713-82725) - [No event storage](misc/no_event_storage.md) (lines 82725-82796) - [Define workflow with steps](misc/define_workflow_with_steps.md) (lines 82796-82805) - [Run workflow with additional_data](misc/run_workflow_with_additional_data.md) (lines 82805-82868) - [Define agents with response models](misc/define_agents_with_response_models.md) (lines 82868-82881) - [Steps reference these agents](misc/steps_reference_these_agents.md) (lines 82881-82933) - [Define agents](misc/define_agents.md) (lines 82933-82947) - [Define steps](misc/define_steps.md) (lines 82947-82958) - [Create workflow with media input](misc/create_workflow_with_media_input.md) (lines 82958-82970) - [Run workflow with image input](misc/run_workflow_with_image_input.md) (lines 82970-83006) - [Migrating to Workflows 2.0](misc/migrating_to_workflows_2_0.md) (lines 83006-83399) - [Run the workflow if the script is executed directly](misc/run_the_workflow_if_the_script_is_executed_directly.md) (lines 83399-83475) - [--- Response Models ---](misc/---_response_models_---.md) (lines 83475-83500) - [--- Agents ---](misc/---_agents_---.md) (lines 83500-83629) - [--- Helper Functions ---](misc/---_helper_functions_---.md) (lines 83629-83791) - [--- Main Execution Function ---](misc/---_main_execution_function_---.md) (lines 83791-83881) - [--- Workflow Definition ---](misc/---_workflow_definition_---.md) (lines 83881-83937) - [What are Workflows?](misc/what_are_workflows.md) (lines 83937-84037) - [Running your Workflow](misc/running_your_workflow.md) (lines 84037-84062) - [Define agents](misc/define_agents.md) (lines 84062-84076) - [Define research team for complex analysis](misc/define_research_team_for_complex_analysis.md) (lines 84076-84104) - [Create and use workflow](misc/create_and_use_workflow.md) (lines 84104-84137) - [=== BASIC AGENTS ===](misc/basic_agents.md) (lines 84137-84160) - [=== CONDITION EVALUATOR ===](misc/condition_evaluator.md) (lines 84160-84186) - [=== WORKFLOW STEPS ===](misc/workflow_steps.md) (lines 84186-84199) - [Conditional fact-checking step](misc/conditional_fact-checking_step.md) (lines 84199-84212) - [=== BASIC LINEAR WORKFLOW ===](misc/basic_linear_workflow.md) (lines 84212-84245) - [Define your agents/team](misc/define_your_agents_team.md) (lines 84245-84259) - [Create and use workflow](misc/create_and_use_workflow.md) (lines 84259-84289) - [=== BASIC AGENTS ===](misc/basic_agents.md) (lines 84289-84312) - [=== CONDITION EVALUATOR ===](misc/condition_evaluator.md) (lines 84312-84338) - [=== WORKFLOW STEPS ===](misc/workflow_steps.md) (lines 84338-84351) - [Conditional fact-checking step](misc/conditional_fact-checking_step.md) (lines 84351-84364) - [=== BASIC LINEAR WORKFLOW ===](misc/basic_linear_workflow.md) (lines 84364-84421) - [Define your workflow](misc/define_your_workflow.md) (lines 84421-84525) - [Workflows Development Guide](misc/workflows_development_guide.md) (lines 84525-84637) - [Named steps for better tracking](misc/named_steps_for_better_tracking.md) (lines 84637-84838) - [Create a reusable content creation sequence](misc/create_a_reusable_content_creation_sequence.md) (lines 84838-84849) - [Use the sequence in a workflow](misc/use_the_sequence_in_a_workflow.md) (lines 84849-84865) - [Define two completely different workflows as Steps](misc/define_two_completely_different_workflows_as_steps.md) (lines 84865-84898) - [Clean workflow with clear branching](misc/clean_workflow_with_clear_branching.md) (lines 84898-84912) - [Usage examples](misc/usage_examples.md) (lines 84912-84970) - [Complex workflow combining multiple patterns](misc/complex_workflow_combining_multiple_patterns.md) (lines 84970-85018) - [How does shared state work between workflows, teams and agents?](misc/how_does_shared_state_work_between_workflows_teams_and_agents.md) (lines 85018-85055) - [Create steps](misc/create_steps.md) (lines 85151-85164) - [Create workflow with workflow_session_state](misc/create_workflow_with_workflow_session_state.md) (lines 85164-85213) - [Running the Agent App on AWS](misc/running_the_agent_app_on_aws.md) (lines 85277-85312) - [Agent App: FastAPI, Streamlit and Postgres](misc/agent_app_fastapi_streamlit_and_postgres.md) (lines 85312-85346) - [Standardized Codebases for Agentic Systems](misc/standardized_codebases_for_agentic_systems.md) (lines 85346-85353) - [What are Workspaces?](misc/what_are_workspaces.md) (lines 85353-85365) - [Here's how they work](misc/here_s_how_they_work.md) (lines 85365-85383) - [How we build Agentic Systems](misc/how_we_build_agentic_systems.md) (lines 85383-85405) - [CI/CD](misc/ci_cd.md) (lines 85405-85554) - [Database Tables](misc/database_tables.md) (lines 85554-85675) - [-*- Build container environment](misc/-_-_build_container_environment.md) (lines 85675-85750) - [-*- Only include tables that are in the target_metadata](misc/-_-_only_include_tables_that_are_in_the_target_metadata.md) (lines 85750-85760) - [Development Application](misc/development_application.md) (lines 85760-85861) - [Use Custom Domain and HTTPS](misc/use_custom_domain_and_https.md) (lines 85861-85911) - [-*- Streamlit running on ECS](misc/-_-_streamlit_running_on_ecs.md) (lines 85911-85920) - [-*- FastAPI running on ECS](misc/-_-_fastapi_running_on_ecs.md) (lines 85920-85963) - [Environment variables](misc/environment_variables.md) (lines 85963-86017) - [Format & Validate](misc/format_validate.md) (lines 86017-86057) - [Create Git Repo](misc/create_git_repo.md) (lines 86057-86088) - [Install & Setup](misc/install_setup.md) (lines 86088-86181) - [Introduction](misc/introduction.md) (lines 86181-86525) - [Setup workspace for new users](misc/setup_workspace_for_new_users.md) (lines 86525-86641) - [Production Application](misc/production_application.md) (lines 86641-86791) - [Add Python Libraries](misc/add_python_libraries.md) (lines 86791-86906) - [Add Secrets](misc/add_secrets.md) (lines 86906-86938) - [-*- Secrets for production application](misc/-_-_secrets_for_production_application.md) (lines 86938-86947) - [-*- Secrets for production database](misc/-_-_secrets_for_production_database.md) (lines 86947-86978) - [SSH Access](misc/ssh_access.md) (lines 86978-87010) - [Workspace Settings](misc/workspace_settings.md) (lines 87010-87090) [V0_FILE]markdown:file="agno/knowledge/-_-_create_a_knowledge_base_from_the_vector_store.md" --- title: -*- Create a knowledge base from the vector store category: knowledge source_lines: 61694-61717 line_count: 23 --- # -*- Create a knowledge base from the vector store knowledge_base = LangChainKnowledgeBase(retriever=retriever) agent = Agent(knowledge_base=knowledge_base, add_references_to_prompt=True) conv.print_response("What did the president say about technology?") \`\`\` ## Params | Parameter | Type | Default | Description | | --------------- | -------------------- | ------- | ------------------------------------------------------------------------- | | `loader` | `Optional[Callable]` | `None` | LangChain loader. | | `vectorstore` | `Optional[Any]` | `None` | LangChain vector store used to create a retriever. | | `search_kwargs` | `Optional[dict]` | `None` | Search kwargs when creating a retriever using the langchain vector store. | | `retriever` | `Optional[Any]` | `None` | LangChain retriever. | `LangChainKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/langchain_kb.py) [V0_FILE]markdown:file="agno/knowledge/-_-_print_the_messages_in_the_memory.md" --- title: -*- Print the messages in the memory category: knowledge source_lines: 16522-16556 line_count: 34 --- # -*- Print the messages in the memory pprint( [ m.model_dump(include={"role", "content"}) for m in agent.get_messages_for_session() ] ) \`\`\` ## Usage \`\`\`bash pip install -U agno \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/memory/00_builtin_memory.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/memory/00_builtin_memory.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/add_a_memory_for_the_default_user.md" --- title: Add a memory for the default user category: knowledge source_lines: 16571-16578 line_count: 7 --- # Add a memory for the default user memory.add_user_memory( memory=UserMemory(memory="The user's name is John Doe", topics=["name"]), ) print("Memories:") pprint(memory.memories) [V0_FILE]markdown:file="agno/knowledge/add_a_user_memory_that_will_persist_across_restarts.md" --- title: Add a user memory that will persist across restarts category: knowledge source_lines: 63306-63316 line_count: 10 --- # Add a user memory that will persist across restarts user_id = "user@example.com" memory.add_user_memory( memory=UserMemory( memory="The user prefers dark mode in applications", topics=["preferences", "ui"] ), user_id=user_id ) [V0_FILE]markdown:file="agno/knowledge/add_information_to_the_knowledge_base.md" --- title: Add information to the knowledge base category: knowledge source_lines: 61444-61447 line_count: 3 --- # Add information to the knowledge base knowledge_base.load_text("The sky is blue") [V0_FILE]markdown:file="agno/knowledge/add_memory_references_false.md" --- title: add_memory_references=False, category: knowledge source_lines: 17793-17794 line_count: 1 --- # add_memory_references=False, [V0_FILE]markdown:file="agno/knowledge/add_the_knowledge_base_to_the_agent.md" --- title: Add the knowledge base to the Agent category: knowledge source_lines: 5884-5903 line_count: 19 --- # Add the knowledge base to the Agent agent = Agent(knowledge_base=knowledge_base) \`\`\` The following embedders are supported: * [OpenAI](/embedder/openai) * [Gemini](/embedder/gemini) * [Ollama](/embedder/ollama) * [Voyage AI](/embedder/voyageai) * [Azure OpenAI](/embedder/azure_openai) * [Mistral](/embedder/mistral) * [Fireworks](/embedder/fireworks) * [Together](/embedder/together) * [HuggingFace](/embedder/huggingface) * [Qdrant FastEmbed](/embedder/qdrant_fastembed) * [Jina](/embedder/jina) [V0_FILE]markdown:file="agno/knowledge/add_the_knowledge_base_to_the_agent_and.md" --- title: Add the knowledge base to the Agent and category: knowledge source_lines: 61447-61448 line_count: 1 --- # Add the knowledge base to the Agent and [V0_FILE]markdown:file="agno/knowledge/agent_with_knowledge.md" --- title: Agent with Knowledge category: knowledge source_lines: 47340-47414 line_count: 74 --- # Agent with Knowledge Source: https://docs.agno.com/examples/models/openai/responses/knowledge ## Code \`\`\`python cookbook/models/openai/responses/knowledge.py from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.models.openai import OpenAIResponses from agno.vectordb.pgvector import PgVector db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai" knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=PgVector(table_name="recipes", db_url=db_url), ) knowledge_base.load(recreate=True) # Comment out after first run agent = Agent( model=OpenAIResponses(id="gpt-4o"), knowledge=knowledge_base, show_tool_calls=True, ) agent.print_response("How to make Thai curry?", markdown=True) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai sqlalchemy pgvector pypdf agno \`\`\` \`\`\`bash docker run -d \ -e POSTGRES_DB=ai \ -e POSTGRES_USER=ai \ -e POSTGRES_PASSWORD=ai \ -e PGDATA=/var/lib/postgresql/data/pgdata \ -v pgvolume:/var/lib/postgresql/data \ -p 5532:5432 \ --name pgvector \ agnohq/pgvector:16 \`\`\` \`\`\`bash Mac python cookbook/models/openai/responses/knowledge.py \`\`\` \`\`\`bash Windows python cookbook/models/openai/responses/knowledge.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/agent_with_knowledge_base.md" --- title: Agent with Knowledge Base category: knowledge source_lines: 39537-39606 line_count: 69 --- # Agent with Knowledge Base Source: https://docs.agno.com/examples/models/cerebras/knowledge ## Code \`\`\`python cookbook/models/cerebras/basic_knowledge.py from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.models.cerebras import Cerebras from agno.vectordb.pgvector import PgVector db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai" knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=PgVector(table_name="recipes", db_url=db_url), ) knowledge_base.load(recreate=True) # Comment out after first run agent = Agent( model=Cerebras(id="llama-4-scout-17b-16e-instruct"), knowledge=knowledge_base, ) agent.print_response("How to make Thai curry?", markdown=True) \`\`\` ## Usage \`\`\`bash export CEREBRAS_API_KEY=xxx \`\`\` \`\`\`bash pip install -U agno cerebras-cloud-sdk \`\`\` Ensure your Postgres server is running and accessible at the connection string used in `db_url`. The first run will load and index the PDF. This may take a while. \`\`\`bash Mac python cookbook/models/cerebras/basic_knowledge.py \`\`\` \`\`\`bash Windows python cookbook/models/cerebras/basic_knowledge.py \`\`\` After the first run, comment out or remove `knowledge_base.load(recreate=True)` to avoid reloading the PDF each time. [V0_FILE]markdown:file="agno/knowledge/agent_with_memory.md" --- title: Agent with Memory category: knowledge source_lines: 48505-48534 line_count: 29 --- # Agent with Memory Source: https://docs.agno.com/examples/models/vllm/memory ## Code \`\`\`python cookbook/models/vllm/memory.py from agno.agent import Agent from agno.memory.v2.db.postgres import PostgresMemoryDb from agno.memory.v2.memory import Memory from agno.models.vllm import vLLM from agno.storage.postgres import PostgresStorage DB_URL = "postgresql+psycopg://ai:ai@localhost:5532/ai" agent = Agent( model=vLLM(id="microsoft/Phi-3-mini-128k-instruct"), memory=Memory( db=PostgresMemoryDb(table_name="agent_memory", db_url=DB_URL), ), enable_user_memories=True, enable_session_summaries=True, storage=PostgresStorage( table_name="personalized_agent_sessions", db_url=DB_URL, ), ) [V0_FILE]markdown:file="agno/knowledge/agent_with_user_memory.md" --- title: Agent With User Memory category: knowledge source_lines: 10576-10616 line_count: 40 --- # Agent With User Memory Source: https://docs.agno.com/examples/applications/whatsapp/agent_with_user_memory This example shows how to use memory with whatsapp app. ## Code \`\`\`python cookbook/apps/whatsapp/agent_with_user_memory.py from textwrap import dedent from agno.agent import Agent from agno.app.whatsapp.app import WhatsappAPI from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.manager import MemoryManager from agno.memory.v2.memory import Memory from agno.models.google import Gemini from agno.storage.sqlite import SqliteStorage from agno.tools.googlesearch import GoogleSearchTools agent_storage = SqliteStorage( table_name="agent_sessions", db_file="tmp/persistent_memory.db" ) memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") memory = Memory( db=memory_db, memory_manager=MemoryManager( memory_capture_instructions="""\ Collect User\'s name, Collect Information about user\'s passion and hobbies, Collect Information about the users likes and dislikes, Collect information about what the user is doing with their life right now """, model=Gemini(id="gemini-2.0-flash"), ), ) [V0_FILE]markdown:file="agno/knowledge/agentic_chunking.md" --- title: Agentic Chunking category: knowledge source_lines: 67274-67285 line_count: 11 --- # Agentic Chunking Source: https://docs.agno.com/reference/chunking/agentic Agentic chunking is an intelligent method of splitting documents into smaller chunks by using a model to determine natural breakpoints in the text. Rather than splitting text at fixed character counts, it analyzes the content to find semantically meaningful boundaries like paragraph breaks and topic transitions. [V0_FILE]markdown:file="agno/knowledge/agentic_knowledge_filters.md" --- title: Agentic Knowledge Filters category: knowledge source_lines: 59086-59192 line_count: 106 --- # Agentic Knowledge Filters Agentic filtering lets the Agent automatically extract filter criteria from your query text, making the experience more natural and interactive. ## Step 1: Attach Metadata There are two ways to attach metadata to your documents: 1. **Attach Metadata When Initializing the Knowledge Base** \`\`\`python knowledge_base = PDFKnowledgeBase( path=[ { "path": "path/to/cv1.pdf", "metadata": { "user_id": "jordan_mitchell", "document_type": "cv", "year": 2025, }, }, # ... more documents ... ], vector_db=vector_db, ) knowledge_base.load(recreate=True) \`\`\` 2. **Attach Metadata When Loading Documents One by One** \`\`\`python # Initialize the PDFKnowledgeBase knowledge_base = PDFKnowledgeBase( vector_db=vector_db, num_documents=5, ) # Load first document with user_1 metadata knowledge_base.load_document( path=path/to/cv1.pdf, metadata={"user_id": "jordan_mitchell", "document_type": "cv", "year": 2025}, recreate=True, # Set to True only for the first run, then set to False ) # Load second document with user_2 metadata knowledge_base.load_document( path=path/to/cv2.pdf, metadata={"user_id": "taylor_brooks", "document_type": "cv", "year": 2025}, ) \`\`\` *** ## How It Works When you enable agentic filtering (`enable_agentic_knowledge_filters=True`), the Agent analyzes your query and applies filters based on the metadata it detects. **Example:** \`\`\`python agent = Agent( knowledge=knowledge_base, search_knowledge=True, enable_agentic_knowledge_filters=True, ) agent.print_response( "Tell me about Jordan Mitchell's experience and skills with jordan_mitchell as user id and document type cv", markdown=True, ) \`\`\` In this example, the Agent will automatically use: * `user_id = "jordan_mitchell"` * `document_type = "cv"` *** ## 🌟 See Agentic Filters in Action! Experience how agentic filters automatically extract relevant metadata from your query. ![Agentic Filters in Action](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/agentic_filters.png) *The Agent intelligently narrows down results based on your query.* *** ## When to Use Agentic Filtering * When you want a more conversational, user-friendly experience. * When users may not know the exact filter syntax. ## Try It Out! * Enable `enable_agentic_knowledge_filters=True` on your Agent. * Ask questions naturally, including filter info in your query. * See how the Agent narrows down results automatically! *** ## Developer Resources * [Agentic filtering](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/filters/pdf/agentic_filtering.py) [V0_FILE]markdown:file="agno/knowledge/agentic_memory.md" --- title: Agentic Memory category: knowledge source_lines: 17021-17043 line_count: 22 --- # Agentic Memory Source: https://docs.agno.com/examples/concepts/memory/06-agentic-memory This example shows you how to use persistent memory with an Agent. During each run the Agent can create/update/delete user memories. To enable this, set `enable_agentic_memory=True` in the Agent config. ## Code \`\`\`python cookbook/agent_concepts/memory/07_agentic_memory.py from agno.agent.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory from agno.models.openai import OpenAIChat from rich.pretty import pprint memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") [V0_FILE]markdown:file="agno/knowledge/agentknowledge.md" --- title: AgentKnowledge category: knowledge source_lines: 67600-67614 line_count: 14 --- # AgentKnowledge Source: https://docs.agno.com/reference/knowledge/base AgentKnolwedge is the base class for all knowledge base classes in Agno. It provides common functionality and parameters that are inherited by all other knowledge base classes. ## Function Reference [V0_FILE]markdown:file="agno/knowledge/arxiv_knowledge_base.md" --- title: Arxiv Knowledge Base category: knowledge source_lines: 67590-67600 line_count: 10 --- # Arxiv Knowledge Base Source: https://docs.agno.com/reference/knowledge/arxiv ArxivKnowledge is a knowledge base class that allows you to load and query papers from the Arxiv API. [V0_FILE]markdown:file="agno/knowledge/ask_the_agent_about_the_knowledge_base.md" --- title: Ask the agent about the knowledge base category: knowledge source_lines: 13389-13432 line_count: 43 --- # Ask the agent about the knowledge base agent.print_response("Ask me about something from the knowledge base", markdown=True) \`\`\` ## Usage \`\`\`bash pip install -U sqlalchemy 'psycopg[binary]' pgvector python-docx agno \`\`\` \`\`\`bash docker run -d \ -e POSTGRES_DB=ai \ -e POSTGRES_USER=ai \ -e POSTGRES_PASSWORD=ai \ -e PGDATA=/var/lib/postgresql/data/pgdata \ -v pgvolume:/var/lib/postgresql/data \ -p 5532:5432 \ --name pgvector \ agnohq/pgvector:16 \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/knowledge/docx_kb.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/knowledge/docx_kb.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/azure_cosmos_db_mongodb_vcore_agent_knowledge.md" --- title: Azure Cosmos DB MongoDB vCore Agent Knowledge category: knowledge source_lines: 79903-79926 line_count: 23 --- # Azure Cosmos DB MongoDB vCore Agent Knowledge Source: https://docs.agno.com/vectordb/azure_cosmos_mongodb ## Setup Follow the instructions in the [Azure Cosmos DB Setup Guide](https://learn.microsoft.com/en-us/azure/cosmos-db/mongodb/vcore) to get the connection string. Install MongoDB packages: \`\`\`shell pip install "pymongo[srv]" \`\`\` ## Example \`\`\`python agent_with_knowledge.py import urllib.parse from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.mongodb import MongoDb [V0_FILE]markdown:file="agno/knowledge/built-in_memory.md" --- title: Built-in Memory category: knowledge source_lines: 16489-16510 line_count: 21 --- # Built-in Memory Source: https://docs.agno.com/examples/concepts/memory/00-built-in-memory ## Code \`\`\`python cookbook/agent_concepts/memory/00_builtin_memory.py from agno.agent import Agent from agno.models.openai import OpenAIChat from rich.pretty import pprint agent = Agent( model=OpenAIChat(id="gpt-4o"), # Set add_history_to_messages=true to add the previous chat history to the messages sent to the Model. add_history_to_messages=True, # Number of historical responses to add to the messages. num_history_responses=3, description="You are a helpful assistant that always responds in a polite, upbeat and positive manner.", ) [V0_FILE]markdown:file="agno/knowledge/cassandra_agent_knowledge.md" --- title: Cassandra Agent Knowledge category: knowledge source_lines: 79963-79995 line_count: 32 --- # Cassandra Agent Knowledge Source: https://docs.agno.com/vectordb/cassandra ## Setup Install cassandra packages \`\`\`shell pip install cassandra-driver \`\`\` Run cassandra \`\`\`shell docker run -d \ --name cassandra-db\ -p 9042:9042 \ cassandra:latest \`\`\` ## Example \`\`\`python agent_with_knowledge.py from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.cassandra import Cassandra from agno.embedder.mistral import MistralEmbedder from agno.models.mistral import MistralChat [V0_FILE]markdown:file="agno/knowledge/chromadb_agent_knowledge.md" --- title: ChromaDB Agent Knowledge category: knowledge source_lines: 80099-80207 line_count: 108 --- # ChromaDB Agent Knowledge Source: https://docs.agno.com/vectordb/chroma ## Setup \`\`\`shell pip install chromadb \`\`\` ## Example \`\`\`python agent_with_knowledge.py import typer from rich.prompt import Prompt from typing import Optional from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.chroma import ChromaDb knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=ChromaDb(collection="recipes"), ) def pdf_agent(user: str = "user"): run_id: Optional[str] = None agent = Agent( run_id=run_id, user_id=user, knowledge_base=knowledge_base, use_tools=True, show_tool_calls=True, debug_mode=True, ) if run_id is None: run_id = agent.run_id print(f"Started Run: {run_id}\n") else: print(f"Continuing Run: {run_id}\n") while True: message = Prompt.ask(f"[bold] :sunglasses: {user} [/bold]") if message in ("exit", "bye"): break agent.print_response(message) if __name__ == "__main__": # Comment out after first run knowledge_base.load(recreate=False) typer.run(pdf_agent) \`\`\`

ChromaDB also supports asynchronous operations, enabling concurrency and leading to better performance.

\`\`\`python async_chroma_db.py # install chromadb - `pip install chromadb` import asyncio from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.chroma import ChromaDb # Initialize ChromaDB vector_db = ChromaDb(collection="recipes", path="tmp/chromadb", persistent_client=True) # Create knowledge base knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) # Create and use the agent agent = Agent(knowledge=knowledge_base, show_tool_calls=True) if __name__ == "__main__": # Comment out after first run asyncio.run(knowledge_base.aload(recreate=False)) # Create and use the agent asyncio.run(agent.aprint_response("How to make Tom Kha Gai", markdown=True)) \`\`\` Use aload() and aprint\_response() methods with asyncio.run() for non-blocking operations in high-throughput applications.

## ChromaDb Params ## Developer Resources * View [Cookbook (Sync)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/chroma_db/chroma_db.py) * View [Cookbook (Async)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/chroma_db/async_chroma_db.py) [V0_FILE]markdown:file="agno/knowledge/clickhouse_agent_knowledge.md" --- title: Clickhouse Agent Knowledge category: knowledge source_lines: 80207-80256 line_count: 49 --- # Clickhouse Agent Knowledge Source: https://docs.agno.com/vectordb/clickhouse ## Setup \`\`\`shell docker run -d \ -e CLICKHOUSE_DB=ai \ -e CLICKHOUSE_USER=ai \ -e CLICKHOUSE_PASSWORD=ai \ -e CLICKHOUSE_DEFAULT_ACCESS_MANAGEMENT=1 \ -v clickhouse_data:/var/lib/clickhouse/ \ -v clickhouse_log:/var/log/clickhouse-server/ \ -p 8123:8123 \ -p 9000:9000 \ --ulimit nofile=262144:262144 \ --name clickhouse-server \ clickhouse/clickhouse-server \`\`\` ## Example \`\`\`python agent_with_knowledge.py from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.storage.sqlite import SqliteStorage from agno.vectordb.clickhouse import Clickhouse agent = Agent( storage=SqliteStorage(table_name="recipe_agent"), knowledge=PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=Clickhouse( table_name="recipe_documents", host="localhost", port=8123, username="ai", password="ai", ), ), # Show tool calls in the response show_tool_calls=True, # Enable the agent to search the knowledge base search_knowledge=True, # Enable the agent to read the chat history read_chat_history=True, ) [V0_FILE]markdown:file="agno/knowledge/combine_knowledge_bases.md" --- title: Combine knowledge bases category: knowledge source_lines: 13051-13065 line_count: 14 --- # Combine knowledge bases knowledge_base = CombinedKnowledgeBase( sources=[ csv_kb, pdf_url_kb, website_kb, local_pdf_kb, ], vector_db=PgVector( table_name="combined_documents", db_url=db_url, ), ) [V0_FILE]markdown:file="agno/knowledge/combined_knowledge_base.md" --- title: Combined Knowledge Base category: knowledge source_lines: 67614-67624 line_count: 10 --- # Combined Knowledge Base Source: https://docs.agno.com/reference/knowledge/combined CombinedKnowledge is a knowledge base class that allows you to load and query multiple knowledge bases at once. [V0_FILE]markdown:file="agno/knowledge/comment_out_after_the_knowledge_base_is_loaded.md" --- title: Comment out after the knowledge base is loaded category: knowledge source_lines: 34405-34435 line_count: 30 --- # Comment out after the knowledge base is loaded if agent.knowledge is not None: agent.knowledge.load() agent.print_response( "How do I make chicken and galangal in coconut milk soup", stream=True ) agent.print_response("What is the history of Thai curry?", stream=True) agent.print_response("What ingredients do I need for Pad Thai?", stream=True) \`\`\` ## Usage \`\`\`bash pip install openai lancedb tantivy pypdf duckduckgo-search agno \`\`\` \`\`\`bash python agent_with_knowledge.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/couchbase_agent_knowledge.md" --- title: Couchbase Agent Knowledge category: knowledge source_lines: 80317-80379 line_count: 62 --- # Couchbase Agent Knowledge Source: https://docs.agno.com/vectordb/couchbase ## Setup ### Local Setup (Docker) Run Couchbase locally using Docker: \`\`\`shell docker run -d --name couchbase-server \ -p 8091-8096:8091-8096 \ -p 11210:11210 \ -e COUCHBASE_ADMINISTRATOR_USERNAME=Administrator \ -e COUCHBASE_ADMINISTRATOR_PASSWORD=password \ couchbase:latest \`\`\` 1. Access the Couchbase UI at: [http://localhost:8091](http://localhost:8091) 2. Login with username: `Administrator` and password: `password` 3. Create a bucket named `recipe_bucket`, a scope `recipe_scope`, and a collection `recipes` ### Managed Setup (Capella) For a managed cluster, use [Couchbase Capella](https://cloud.couchbase.com/): * Follow Capella's UI to create a database, bucket, scope, and collection ### Environment Variables Set up your environment variables: \`\`\`shell export COUCHBASE_USER="Administrator" export COUCHBASE_PASSWORD="password" export COUCHBASE_CONNECTION_STRING="couchbase://localhost" export OPENAI_API_KEY="" \`\`\` For Capella, set `COUCHBASE_CONNECTION_STRING` to your Capella connection string. ### Install Dependencies \`\`\`shell pip install couchbase \`\`\` ## Example \`\`\`python agent_with_knowledge.py import os import time from agno.agent import Agent from agno.embedder.openai import OpenAIEmbedder from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.couchbase import CouchbaseSearch from couchbase.options import ClusterOptions, KnownConfigProfiles from couchbase.auth import PasswordAuthenticator from couchbase.management.search import SearchIndex [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_containing_information_from_a_url.md" --- title: Create a knowledge base containing information from a URL category: knowledge source_lines: 72559-72609 line_count: 50 --- # Create a knowledge base containing information from a URL agno_docs = UrlKnowledge( urls=["https://docs.agno.com/llms-full.txt"], # Use LanceDB as the vector database and store embeddings in the `agno_docs` table vector_db=LanceDb( uri="tmp/lancedb", table_name="agno_docs", search_type=SearchType.hybrid, embedder=OpenAIEmbedder(id="text-embedding-3-small"), ), ) knowledge_tools = KnowledgeTools( knowledge=agno_docs, think=True, search=True, analyze=True, add_few_shot=True, ) agent = Agent( model=OpenAIChat(id="gpt-4o"), tools=[knowledge_tools], show_tool_calls=True, markdown=True, ) if __name__ == "__main__": # Load the knowledge base, comment after first run agno_docs.load(recreate=True) agent.print_response("How do I build multi-agent teams with Agno?", stream=True) \`\`\` The toolkit comes with default instructions and few-shot examples to help the Agent use the tools effectively. Here is how you can configure them: \`\`\`python from agno.tools.knowledge import KnowledgeTools knowledge_tools = KnowledgeTools( knowledge=my_knowledge_base, think=True, # Enable the think tool search=True, # Enable the search tool analyze=True, # Enable the analyze tool add_instructions=True, # Add default instructions add_few_shot=True, # Add few-shot examples few_shot_examples=None, # Optional custom few-shot examples ) \`\`\` [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_for_the_agent.md" --- title: Create a knowledge base for the Agent category: knowledge source_lines: 61441-61444 line_count: 3 --- # Create a knowledge base for the Agent knowledge_base = AgentKnowledge(vector_db=...) [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_from_a_pdf_url_using_lancedb_for_vector_storage_and_ollamaembedder_for_embedding.md" --- title: Create a knowledge base from a PDF URL using LanceDb for vector storage and OllamaEmbedder for embedding category: knowledge source_lines: 21370-21376 line_count: 6 --- # Create a knowledge base from a PDF URL using LanceDb for vector storage and OllamaEmbedder for embedding knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_from_the_vector_store.md" --- title: Create a knowledge base from the vector store category: knowledge source_lines: 61844-61847 line_count: 3 --- # Create a knowledge base from the vector store knowledge_base = LlamaIndexKnowledgeBase(retriever=retriever) [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_loaded_with_documents_from_a_url.md" --- title: Create a knowledge base, loaded with documents from a URL category: knowledge source_lines: 62452-62493 line_count: 41 --- # Create a knowledge base, loaded with documents from a URL knowledge_base = UrlKnowledge( urls=["https://docs.agno.com/introduction/agents.md"], # Use LanceDB as the vector database, store embeddings in the `agno_docs` table vector_db=LanceDb( uri="tmp/lancedb", table_name="agno_docs", search_type=SearchType.hybrid, embedder=CohereEmbedder(id="embed-v4.0"), reranker=CohereReranker(model="rerank-v3.5"), ), ) agent = Agent( model=Claude(id="claude-3-7-sonnet-latest"), # Agentic RAG is enabled by default when `knowledge` is provided to the Agent. knowledge=knowledge_base, # search_knowledge=True gives the Agent the ability to search on demand # search_knowledge is True by default search_knowledge=True, tools=[ReasoningTools(add_instructions=True)], instructions=[ "Include sources in your response.", "Always search your knowledge before answering the question.", "Only include the output in your response. No other text.", ], markdown=True, ) if __name__ == "__main__": # Load the knowledge base, comment after first run # knowledge_base.load(recreate=True) agent.print_response( "What are Agents?", stream=True, show_full_reasoning=True, stream_intermediate_steps=True, ) \`\`\` [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_of_pdfs_from_urls.md" --- title: Create a knowledge base of PDFs from URLs category: knowledge source_lines: 21519-21530 line_count: 11 --- # Create a knowledge base of PDFs from URLs knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], # Use PgVector as the vector database and store embeddings in the `ai.recipes` table vector_db=PgVector( table_name="recipes", db_url=db_url, search_type=SearchType.hybrid, embedder=OpenAIEmbedder(id="text-embedding-3-small"), ), ) [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_with_the_arxiv_documents.md" --- title: Create a knowledge base with the ArXiv documents category: knowledge source_lines: 60659-60664 line_count: 5 --- # Create a knowledge base with the ArXiv documents knowledge_base = ArxivKnowledgeBase( queries=["Generative AI", "Machine Learning"], vector_db=vector_db ) [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_with_the_docx_files_from_the_data_docs_directory.md" --- title: Create a knowledge base with the DOCX files from the data/docs directory category: knowledge source_lines: 61290-61300 line_count: 10 --- # Create a knowledge base with the DOCX files from the data/docs directory knowledge_base = DocxKnowledgeBase( path=Path("tmp/docs"), vector_db=LanceDb( uri="tmp/lancedb", table_name="docx_reader", search_type=SearchType.hybrid, ), ) [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_with_the_loaded_documents.md" --- title: Create a knowledge base with the loaded documents category: knowledge source_lines: 13294-13303 line_count: 9 --- # Create a knowledge base with the loaded documents knowledge_base = DocumentKnowledgeBase( documents=documents, vector_db=PgVector( table_name="documents", db_url=db_url, ), ) [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_with_the_pdfs_from_the_data_pdfs_directory.md" --- title: Create a knowledge base with the PDFs from the data/pdfs directory category: knowledge source_lines: 62195-62202 line_count: 7 --- # Create a knowledge base with the PDFs from the data/pdfs directory knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, reader=PDFUrlReader(chunk=True), ) [V0_FILE]markdown:file="agno/knowledge/create_a_knowledge_base_with_the_seed_urls.md" --- title: Create a knowledge base with the seed URLs category: knowledge source_lines: 62665-62674 line_count: 9 --- # Create a knowledge base with the seed URLs knowledge_base = WebsiteKnowledgeBase( urls=["https://docs.agno.com/introduction"], # Number of links to follow from the seed URLs max_links=5, # Table name: ai.website_documents vector_db=vector_db, ) [V0_FILE]markdown:file="agno/knowledge/create_a_memory_instance_with_persistent_storage.md" --- title: Create a memory instance with persistent storage category: knowledge source_lines: 69702-69804 line_count: 102 --- # Create a memory instance with persistent storage memory_db = SqliteMemoryDb(table_name="memory", db_file="memory.db") memory = Memory(db=memory_db) team_with_memory = Team( name="Team with Memory", members=[agent1, agent2], memory=memory, enable_user_memories=True, ) team_with_memory.print_response("Hi! My name is John Doe.") team_with_memory.print_response("What is my name?") \`\`\` ## Session Summaries To enable session summaries, set `enable_session_summaries=True` on the `Team`. \`\`\`python from agno.team import Team from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory team_with_session_summaries = Team( name="Team with Memory", members=[agent1, agent2], enable_session_summaries=True, ) team_with_session_summaries.print_response("Hi! My name is John Doe and I live in New York City.") session_summary = team_with_session_summaries.get_session_summary() print("Session Summary: ", session_summary.summary) \`\`\` ## Examples ### Multi-Language Team Let's walk through a simple example where we use different models to answer questions in different languages. The team consists of three specialized agents and the team leader routes the user's question to the appropriate language agent. \`\`\`python multilanguage_team.py from agno.agent import Agent from agno.models.deepseek import DeepSeek from agno.models.mistral.mistral import MistralChat from agno.models.openai import OpenAIChat from agno.team.team import Team english_agent = Agent( name="English Agent", role="You only answer in English", model=OpenAIChat(id="gpt-4o"), ) chinese_agent = Agent( name="Chinese Agent", role="You only answer in Chinese", model=DeepSeek(id="deepseek-chat"), ) french_agent = Agent( name="French Agent", role="You can only answer in French", model=MistralChat(id="mistral-large-latest"), ) multi_language_team = Team( name="Multi Language Team", mode="route", model=OpenAIChat("gpt-4o"), members=[english_agent, chinese_agent, french_agent], show_tool_calls=True, markdown=True, description="You are a language router that directs questions to the appropriate language agent.", instructions=[ "Identify the language of the user's question and direct it to the appropriate language agent.", "If the user asks in a language whose agent is not a team member, respond in English with:", "'I can only answer in the following languages: English, Chinese, French. Please ask your question in one of these languages.'", "Always check the language of the user's input before routing to an agent.", "For unsupported languages like Italian, respond in English with the above message.", ], show_members_responses=True, ) if __name__ == "__main__": # Ask "How are you?" in all supported languages multi_language_team.print_response("Comment allez-vous?", stream=True) # French multi_language_team.print_response("How are you?", stream=True) # English multi_language_team.print_response("你好吗？", stream=True) # Chinese multi_language_team.print_response("Come stai?", stream=True) # Italian \`\`\` ### Content Team Let's walk through another example where we use two specialized agents to write a blog post. The team leader coordinates the agents to write a blog post. \`\`\`python content_team.py from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.team import Team from agno.tools.duckduckgo import DuckDuckGoTools [V0_FILE]markdown:file="agno/knowledge/create_a_postgresql_memory_database.md" --- title: Create a PostgreSQL memory database category: knowledge source_lines: 63329-63335 line_count: 6 --- # Create a PostgreSQL memory database memory_db = PostgresMemoryDb( table_name="user_memories", connection_string="postgresql://user:password@localhost:5432/mydb" ) [V0_FILE]markdown:file="agno/knowledge/create_a_recipe_expert_agent_with_knowledge_of_thai_recipes.md" --- title: Create a Recipe Expert Agent with knowledge of Thai recipes category: knowledge source_lines: 34343-34405 line_count: 62 --- # Create a Recipe Expert Agent with knowledge of Thai recipes agent = Agent( model=OpenAIChat(id="gpt-4o"), instructions=dedent("""\ You are a passionate and knowledgeable Thai cuisine expert! 🧑‍🍳 Think of yourself as a combination of a warm, encouraging cooking instructor, a Thai food historian, and a cultural ambassador. Follow these steps when answering questions: 1. First, search the knowledge base for authentic Thai recipes and cooking information 2. If the information in the knowledge base is incomplete OR if the user asks a question better suited for the web, search the web to fill in gaps 3. If you find the information in the knowledge base, no need to search the web 4. Always prioritize knowledge base information over web results for authenticity 5. If needed, supplement with web searches for: - Modern adaptations or ingredient substitutions - Cultural context and historical background - Additional cooking tips and troubleshooting Communication style: 1. Start each response with a relevant cooking emoji 2. Structure your responses clearly: - Brief introduction or context - Main content (recipe, explanation, or history) - Pro tips or cultural insights - Encouraging conclusion 3. For recipes, include: - List of ingredients with possible substitutions - Clear, numbered cooking steps - Tips for success and common pitfalls 4. Use friendly, encouraging language Special features: - Explain unfamiliar Thai ingredients and suggest alternatives - Share relevant cultural context and traditions - Provide tips for adapting recipes to different dietary needs - Include serving suggestions and accompaniments End each response with an uplifting sign-off like: - 'Happy cooking! ขอให้อร่อย (Enjoy your meal)!' - 'May your Thai cooking adventure bring joy!' - 'Enjoy your homemade Thai feast!' Remember: - Always verify recipe authenticity with the knowledge base - Clearly indicate when information comes from web sources - Be encouraging and supportive of home cooks at all skill levels\ """), knowledge=PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=LanceDb( uri="tmp/lancedb", table_name="recipe_knowledge", search_type=SearchType.hybrid, embedder=OpenAIEmbedder(id="text-embedding-3-small"), ), ), tools=[DuckDuckGoTools()], show_tool_calls=True, markdown=True, add_references=True, ) [V0_FILE]markdown:file="agno/knowledge/create_a_sqlite_database_for_memory.md" --- title: Create a SQLite database for memory category: knowledge source_lines: 63272-63278 line_count: 6 --- # Create a SQLite database for memory memory_db = SqliteMemoryDb( table_name="memories", # The table name to use db_file="path/to/memory.db" # The SQLite database file ) [V0_FILE]markdown:file="agno/knowledge/create_a_sqlite_memory_database.md" --- title: Create a SQLite memory database category: knowledge source_lines: 63297-63303 line_count: 6 --- # Create a SQLite memory database memory_db = SqliteMemoryDb( table_name="user_memories", db_file="tmp/memory.db" ) [V0_FILE]markdown:file="agno/knowledge/create_a_user_memory_manually.md" --- title: Create a user memory manually category: knowledge source_lines: 63052-63070 line_count: 18 --- # Create a user memory manually memory_id = memory.add_user_memory( memory=UserMemory( memory="The user's name is Jane Doe", topics=["personal", "name"] ), user_id="jane_doe@example.com" ) \`\`\` ### Updating a memory \`\`\`python from agno.memory.v2.memory import Memory from agno.memory.v2.schema import UserMemory memory = Memory() [V0_FILE]markdown:file="agno/knowledge/create_agent_with_both_memory_and_storage.md" --- title: Create agent with both memory and storage category: knowledge source_lines: 63376-63390 line_count: 14 --- # Create agent with both memory and storage agent = Agent( model=OpenAIChat(id="gpt-4o"), memory=memory, storage=agent_storage, enable_user_memories=True, ) \`\`\` ## Memory Management When using persistent storage, the Memory system offers several functions to manage stored memories: \`\`\`python [V0_FILE]markdown:file="agno/knowledge/create_agent_with_memory.md" --- title: Create agent with memory category: knowledge source_lines: 17942-18002 line_count: 60 --- # Create agent with memory agent = Agent( model=OpenAIChat(id="gpt-4o"), memory=memory, enable_user_memories=True, ) async def run_example(): # Use the agent with MongoDB-backed memory await agent.aprint_response( "My name is Jane Smith and I enjoy painting and photography.", user_id="jane@example.com", ) await agent.aprint_response( "What are my creative interests?", user_id="jane@example.com", ) # Display the memories stored in MongoDB memories = memory.get_user_memories(user_id="jane@example.com") print("Memories stored in MongoDB:") for i, m in enumerate(memories): print(f"{i}: {m.memory}") if __name__ == "__main__": asyncio.run(run_example()) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U agno openai pymongo \`\`\` \`\`\`bash Mac/Linux python cookbook/agent_concepts/memory/mongodb_memory.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/memory/mongodb_memory.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/create_agent_with_memory_and_redis_storage.md" --- title: Create agent with memory and Redis storage category: knowledge source_lines: 18114-18123 line_count: 9 --- # Create agent with memory and Redis storage agent = Agent( model=OpenAIChat(id="gpt-4o-mini"), memory=memory, storage=RedisStorage(prefix="agno_test", host="localhost", port=6379), enable_user_memories=True, enable_session_summaries=True, ) [V0_FILE]markdown:file="agno/knowledge/create_agent_with_memory_and_sqlite_storage.md" --- title: Create agent with memory and SQLite storage category: knowledge source_lines: 18217-18229 line_count: 12 --- # Create agent with memory and SQLite storage agent = Agent( model=OpenAIChat(id="gpt-4o-mini"), memory=memory, storage=SqliteStorage( table_name="agent_sessions", db_file="tmp/memory.db" ), enable_user_memories=True, enable_session_summaries=True, ) [V0_FILE]markdown:file="agno/knowledge/create_an_agent_with_the_knowledge_base.md" --- title: Create an agent with the knowledge base category: knowledge source_lines: 62674-62702 line_count: 28 --- # Create an agent with the knowledge base agent = Agent(knowledge=knowledge_base, search_knowledge=True, debug_mode=True) if __name__ == "__main__": # Comment out after first run asyncio.run(knowledge_base.aload(recreate=False)) # Create and use the agent asyncio.run(agent.aprint_response("How does agno work?", markdown=True)) \`\`\` ## Params | Parameter | Type | Default | Description | | ----------- | ------------------------- | ------- | ------------------------------------------------------------------------------------------------- | | `urls` | `List[str]` | `[]` | URLs to read | | `reader` | `Optional[WebsiteReader]` | `None` | A `WebsiteReader` that reads the urls and converts them into `Documents` for the vector database. | | `max_depth` | `int` | `3` | Maximum depth to crawl. | | `max_links` | `int` | `10` | Number of links to crawl. | `WebsiteKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Sync loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/website_kb.py) * View [Async loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/website_kb_async.py) [V0_FILE]markdown:file="agno/knowledge/create_an_agent_with_the_knowledge_base_and_the_retriever.md" --- title: Create an agent with the knowledge base and the retriever category: knowledge source_lines: 61753-61786 line_count: 33 --- # Create an agent with the knowledge base and the retriever agent = Agent( model=Claude(id="claude-3-7-sonnet-latest"), # Agentic RAG is enabled by default when `knowledge` is provided to the Agent. knowledge=knowledge_base, retriever=lightrag_retriever, # search_knowledge=True gives the Agent the ability to search on demand # search_knowledge is True by default search_knowledge=True, instructions=[ "Include sources in your response.", "Always search your knowledge before answering the question.", "Use the async_search method to search the knowledge base.", ], markdown=True, ) asyncio.run(agent.aprint_response("What are Agno Agents?")) \`\`\` ## Params | Parameter | Type | Default | Description | | --------------------- | ------------------ | ------- | -------------------------------------------------------------------- | | `lightrag_server_url` | `str` | - | URL to LightRAG server. | | `path` | `Union[str, Path]` | - | Path to documents. Can point to a single file or directory of files. | | `url` | `str` | - | URLs of the website to read. | ## Developer Resources * View [Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/agentic_search/lightrag) [V0_FILE]markdown:file="agno/knowledge/create_csv_knowledge_base.md" --- title: Create CSV knowledge base category: knowledge source_lines: 13014-13023 line_count: 9 --- # Create CSV knowledge base csv_kb = CSVKnowledgeBase( path=Path("data/csvs"), vector_db=PgVector( table_name="csv_documents", db_url=db_url, ), ) [V0_FILE]markdown:file="agno/knowledge/create_knowledge_base.md" --- title: Create knowledge base category: knowledge source_lines: 81557-81564 line_count: 7 --- # Create knowledge base knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) knowledge_base.load(recreate=False) # Comment out after first run [V0_FILE]markdown:file="agno/knowledge/create_local_pdf_knowledge_base.md" --- title: Create Local PDF knowledge base category: knowledge source_lines: 13042-13051 line_count: 9 --- # Create Local PDF knowledge base local_pdf_kb = PDFKnowledgeBase( path="data/pdfs", vector_db=PgVector( table_name="pdf_documents", db_url=db_url, ), ) [V0_FILE]markdown:file="agno/knowledge/create_memory_instance_with_mongodb_backend.md" --- title: Create memory instance with MongoDB backend category: knowledge source_lines: 17936-17939 line_count: 3 --- # Create memory instance with MongoDB backend memory = Memory(db=memory_db) [V0_FILE]markdown:file="agno/knowledge/create_memory_instance_with_redis_backend.md" --- title: Create memory instance with Redis backend category: knowledge source_lines: 18104-18107 line_count: 3 --- # Create memory instance with Redis backend memory = Memory(db=memory_db) [V0_FILE]markdown:file="agno/knowledge/create_memory_instance_with_sqlite_backend.md" --- title: Create memory instance with SQLite backend category: knowledge source_lines: 18207-18210 line_count: 3 --- # Create memory instance with SQLite backend memory = Memory(db=memory_db) [V0_FILE]markdown:file="agno/knowledge/create_memory_storage.md" --- title: Create memory storage category: knowledge source_lines: 63363-63370 line_count: 7 --- # Create memory storage memory_db = SqliteMemoryDb( table_name="memories", db_file="tmp/memory.db" ) memory = Memory(db=memory_db) [V0_FILE]markdown:file="agno/knowledge/create_mongodb_memory_database.md" --- title: Create MongoDB memory database category: knowledge source_lines: 17929-17936 line_count: 7 --- # Create MongoDB memory database memory_db = MongoMemoryDb( connection_string=mongo_url, database_name=database_name, collection_name="memories" # Collection name to use in the database ) [V0_FILE]markdown:file="agno/knowledge/create_pdf_url_knowledge_base.md" --- title: Create PDF URL knowledge base category: knowledge source_lines: 13023-13032 line_count: 9 --- # Create PDF URL knowledge base pdf_url_kb = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=PgVector( table_name="pdf_documents", db_url=db_url, ), ) [V0_FILE]markdown:file="agno/knowledge/create_redis_memory_database.md" --- title: Create Redis memory database category: knowledge source_lines: 18097-18104 line_count: 7 --- # Create Redis memory database memory_db = RedisMemoryDb( prefix="agno_memory", # Prefix for Redis keys to namespace the memories host="localhost", # Redis host address port=6379, # Redis port number ) [V0_FILE]markdown:file="agno/knowledge/create_sqlite_memory_database.md" --- title: Create SQLite memory database category: knowledge source_lines: 18201-18207 line_count: 6 --- # Create SQLite memory database memory_db = SqliteMemoryDb( table_name="agent_memories", # Table name to use in the database db_file="tmp/memory.db", # Path to SQLite database file ) [V0_FILE]markdown:file="agno/knowledge/create_website_knowledge_base.md" --- title: Create Website knowledge base category: knowledge source_lines: 13032-13042 line_count: 10 --- # Create Website knowledge base website_kb = WebsiteKnowledgeBase( urls=["https://docs.agno.com/introduction"], max_links=10, vector_db=PgVector( table_name="website_documents", db_url=db_url, ), ) [V0_FILE]markdown:file="agno/knowledge/csv_knowledge_base.md" --- title: CSV Knowledge Base category: knowledge source_lines: 67624-67634 line_count: 10 --- # CSV Knowledge Base Source: https://docs.agno.com/reference/knowledge/csv CSVKnowledge is a knowledge base class that allows you to load and query data from CSV files. [V0_FILE]markdown:file="agno/knowledge/csv_url_knowledge_base.md" --- title: CSV URL Knowledge Base category: knowledge source_lines: 67634-67644 line_count: 10 --- # CSV URL Knowledge Base Source: https://docs.agno.com/reference/knowledge/csv_url CSVUrlKnowledge is a knowledge base class that allows you to load and query data from CSV files stored in URLs. [V0_FILE]markdown:file="agno/knowledge/custom_configuration_with_late_chunking_for_long_documents.md" --- title: Custom configuration with late chunking for long documents category: knowledge source_lines: 5948-5955 line_count: 7 --- # Custom configuration with late chunking for long documents custom_embedder = JinaEmbedder( dimensions=1024, late_chunking=True, # Improved processing for long documents timeout=30.0, # Request timeout in seconds ) [V0_FILE]markdown:file="agno/knowledge/custom_memory.md" --- title: Custom Memory category: knowledge source_lines: 17341-17365 line_count: 24 --- # Custom Memory Source: https://docs.agno.com/examples/concepts/memory/09-custom-memory This example shows how you can configure the Memory Manager and Summarizer models individually. In this example, we use OpenRouter and LLama 3.3-70b-instruct for the memory manager and Claude 3.5 Sonnet for the summarizer, while using Gemini for the Agent. We also set custom system prompts for the memory manager and summarizer. ## Code \`\`\`python cookbook/agent_concepts/memory/10_custom_memory.py from agno.agent.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory, MemoryManager, SessionSummarizer from agno.models.anthropic.claude import Claude from agno.models.google.gemini import Gemini from agno.models.openrouter.openrouter import OpenRouter from rich.pretty import pprint memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") [V0_FILE]markdown:file="agno/knowledge/custom_memory_creation.md" --- title: Custom Memory Creation category: knowledge source_lines: 16707-16726 line_count: 19 --- # Custom Memory Creation Source: https://docs.agno.com/examples/concepts/memory/03-custom-memory-creation This example demonstrates how to create user memories with an Agent by providing either text or a list of messages. The Agent uses a custom memory manager to capture and store relevant details. ## Code \`\`\`python cookbook/agent_concepts/memory/04_custom_memory_creation.py from agno.memory.v2 import Memory from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.manager import MemoryManager from agno.models.anthropic.claude import Claude from agno.models.google import Gemini from agno.models.message import Message from rich.pretty import pprint memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") [V0_FILE]markdown:file="agno/knowledge/database_file_for_memory_and_storage.md" --- title: Database file for memory and storage category: knowledge source_lines: 62918-62921 line_count: 3 --- # Database file for memory and storage db_file = "tmp/agent.db" [V0_FILE]markdown:file="agno/knowledge/delete_a_memory.md" --- title: Delete a memory category: knowledge source_lines: 16593-16601 line_count: 8 --- # Delete a memory print("\nDeleting memory") memory.delete_user_memory(user_id=jane_doe_id, memory_id=memory_id_2) print("Memory deleted\n") memories = memory.get_user_memories(user_id=jane_doe_id) print("Memories:") pprint(memories) [V0_FILE]markdown:file="agno/knowledge/delete_a_specific_memory.md" --- title: Delete a specific memory category: knowledge source_lines: 63390-63393 line_count: 3 --- # Delete a specific memory memory.delete_user_memory(user_id="user@example.com", memory_id="memory_123") [V0_FILE]markdown:file="agno/knowledge/delete_a_user_memory.md" --- title: Delete a user memory category: knowledge source_lines: 63090-63139 line_count: 49 --- # Delete a user memory memory.delete_user_memory(user_id="jane_doe@example.com", memory_id=memory_id) \`\`\` ### Creating memories from user information \`\`\`python from agno.memory.v2 import Memory from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.models.google import Gemini memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") memory = Memory(model=Gemini(id="gemini-2.0-flash-exp"), db=memory_db) john_doe_id = "john_doe@example.com" memory.create_user_memories( message=""" I enjoy hiking in the mountains on weekends, reading science fiction novels before bed, cooking new recipes from different cultures, playing chess with friends, and attending live music concerts whenever possible. Photography has become a recent passion of mine, especially capturing landscapes and street scenes. I also like to meditate in the mornings and practice yoga to stay centered. """, user_id=john_doe_id, ) memories = memory.get_user_memories(user_id=john_doe_id) print("John Doe's memories:") for i, m in enumerate(memories): print(f"{i}: {m.memory} - {m.topics}") \`\`\` ### Creating memories from a conversation \`\`\`python from agno.memory.v2 import Memory from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.models.google import Gemini from agno.models.message import Message memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") memory = Memory(model=Gemini(id="gemini-2.0-flash-exp"), db=memory_db) jane_doe_id = "jane_doe@example.com" [V0_FILE]markdown:file="agno/knowledge/document_chunking.md" --- title: Document Chunking category: knowledge source_lines: 67285-67296 line_count: 11 --- # Document Chunking Source: https://docs.agno.com/reference/chunking/document Document chunking is a method of splitting documents into smaller chunks based on document structure like paragraphs and sections. It analyzes natural document boundaries rather than splitting at fixed character counts. This is useful when you want to process large documents while preserving semantic meaning and context. [V0_FILE]markdown:file="agno/knowledge/document_knowledge_base.md" --- title: Document Knowledge Base category: knowledge source_lines: 61168-61228 line_count: 60 --- # Document Knowledge Base Source: https://docs.agno.com/knowledge/document Learn how to use local documents in your knowledge base. The **DocumentKnowledgeBase** reads **local docs** files, converts them into vector embeddings and loads them to a vector database. ## Usage We are using a local PgVector database for this example. [Make sure it's running](https://docs.agno.com/vectordb/pgvector) \`\`\`shell pip install textract \`\`\` \`\`\`python from agno.knowledge.document import DocumentKnowledgeBase from agno.vectordb.pgvector import PgVector knowledge_base = DocumentKnowledgeBase( path="data/docs", # Table name: ai.documents vector_db=PgVector( table_name="documents", db_url="postgresql+psycopg://ai:ai@localhost:5532/ai", ), ) \`\`\` Then use the `knowledge_base` with an `Agent`: \`\`\`python from agno.agent import Agent from knowledge_base import knowledge_base agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) agent.knowledge.load(recreate=False) agent.print_response("Ask me about something from the knowledge base") \`\`\` ## Params | Parameter | Type | Default | Description | | ----------- | ---------------- | ------- | --------------------------------------------------------- | | `documents` | `List[Document]` | - | List of Document objects to be used as the knowledge base | `DocumentKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Sync loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/doc_kb.py) * View [Async loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/doc_kb_async.py) [V0_FILE]markdown:file="agno/knowledge/docx_knowledge_base.md" --- title: Docx Knowledge Base category: knowledge source_lines: 67644-67654 line_count: 10 --- # Docx Knowledge Base Source: https://docs.agno.com/reference/knowledge/docx DocxKnowledge is a knowledge base class that allows you to load and query data from Docx files. [V0_FILE]markdown:file="agno/knowledge/example_usage_with_a_pdf_knowledge_base.md" --- title: Example usage with a PDF knowledge base category: knowledge source_lines: 5547-5562 line_count: 15 --- # Example usage with a PDF knowledge base knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], reader=PDFUrlReader( chunk_size=2048 ), # Required because Cohere model has a fixed size of 2048 vector_db=PgVector( table_name="recipes", db_url="postgresql+psycopg://ai:ai@localhost:5532/ai", embedder=AwsBedrockEmbedder(), ), ) knowledge_base.load(recreate=False) \`\`\` [V0_FILE]markdown:file="agno/knowledge/fixed_size_chunking.md" --- title: Fixed Size Chunking category: knowledge source_lines: 67296-67307 line_count: 11 --- # Fixed Size Chunking Source: https://docs.agno.com/reference/chunking/fixed-size Fixed size chunking is a method of splitting documents into smaller chunks of a specified size, with optional overlap between chunks. This is useful when you want to process large documents in smaller, manageable pieces. [V0_FILE]markdown:file="agno/knowledge/get_the_most_recent_memory.md" --- title: Get the most recent memory category: knowledge source_lines: 63182-63190 line_count: 8 --- # Get the most recent memory memories = memory.search_user_memories( user_id=john_doe_id, limit=1, retrieval_method="last_n" ) print("John Doe's last_n memories:") for i, m in enumerate(memories): print(f"{i}: {m.memory}") [V0_FILE]markdown:file="agno/knowledge/get_the_oldest_memory.md" --- title: Get the oldest memory category: knowledge source_lines: 63190-63207 line_count: 17 --- # Get the oldest memory memories = memory.search_user_memories( user_id=john_doe_id, limit=1, retrieval_method="first_n" ) print("John Doe's first_n memories:") for i, m in enumerate(memories): print(f"{i}: {m.memory}") \`\`\` ### Agentic Memory Search Agentic search allows you to find memories based on meaning rather than exact keyword matches. This is particularly useful for retrieving contextually relevant information: \`\`\`python from agno.memory.v2.memory import Memory, UserMemory from agno.models.google.gemini import Gemini [V0_FILE]markdown:file="agno/knowledge/give_it_a_tool_to_search_the_knowledge_base_as_needed.md" --- title: give it a tool to search the knowledge base as needed category: knowledge source_lines: 61448-61509 line_count: 61 --- # give it a tool to search the knowledge base as needed agent = Agent(knowledge=knowledge_base, search_knowledge=True) \`\`\` We can give our agent access to the knowledge base in the following ways: * We can set `search_knowledge=True` to add a `search_knowledge_base()` tool to the Agent. `search_knowledge` is `True` **by default** if you add `knowledge` to an Agent. * We can set `add_references=True` to automatically add references from the knowledge base to the Agent's prompt. This is the traditional 2023 RAG approach. If you need complete control over the knowledge base search, you can pass your own `retriever` function with the following signature: \`\`\`python def retriever(agent: Agent, query: str, num_documents: Optional[int], **kwargs) -> Optional[list[dict]]: ... \`\`\` This function is called during `search_knowledge_base()` and is used by the Agent to retrieve references from the knowledge base. For more details check out the [Custom Retriever](../knowledge/custom_retriever) page. ## Vector Databases While any type of storage can act as a knowledge base, vector databases offer the best solution for retrieving relevant results from dense information quickly. Here's how vector databases are used with Agents: Break down the knowledge into smaller chunks to ensure our search query returns only relevant results. Convert the chunks into embedding vectors and store them in a vector database. When the user sends a message, we convert the input message into an embedding and "search" for nearest neighbors in the vector database. ## Loading the Knowledge Base Before you can use a knowledge base, it needs to be loaded with embeddings that will be used for retrieval. ### Asynchronous Loading Many vector databases support asynchronous operations, which can significantly improve performance when loading large knowledge bases. You can leverage this capability using the `aload()` method: \`\`\`python import asyncio from agno.agent import Agent from agno.knowledge.pdf import PDFKnowledgeBase, PDFReader from agno.vectordb.qdrant import Qdrant COLLECTION_NAME = "pdf-reader" vector_db = Qdrant(collection=COLLECTION_NAME, url="http://localhost:6333") [V0_FILE]markdown:file="agno/knowledge/image_agent_with_memory.md" --- title: Image Agent with Memory category: knowledge source_lines: 49267-49325 line_count: 58 --- # Image Agent with Memory Source: https://docs.agno.com/examples/models/xai/image_agent_with_memory ## Code \`\`\`python cookbook/models/xai/image_agent_with_memory.py from agno.agent import Agent from agno.media import Image from agno.models.xai import xAI from agno.tools.duckduckgo import DuckDuckGoTools agent = Agent( model=xAI(id="grok-2-vision-latest"), tools=[DuckDuckGoTools()], markdown=True, add_history_to_messages=True, num_history_responses=3, ) agent.print_response( "Tell me about this image and give me the latest news about it.", images=[ Image( url="https://upload.wikimedia.org/wikipedia/commons/0/0c/GoldenGateBridge-001.jpg" ) ], ) agent.print_response("Tell me where I can get more images?") \`\`\` ## Usage \`\`\`bash export XAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai duckduckgo-search agno \`\`\` \`\`\`bash python cookbook/models/xai/image_agent_with_memory.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/in-memory_database_to_store_user_shopping_lists.md" --- title: In-memory database to store user shopping lists category: knowledge source_lines: 23451-23452 line_count: 1 --- # In-memory database to store user shopping lists [V0_FILE]markdown:file="agno/knowledge/initialize_knowledge_base.md" --- title: Initialize knowledge base category: knowledge source_lines: 69658-69702 line_count: 44 --- # Initialize knowledge base agno_docs_knowledge = UrlKnowledge( urls=["https://docs.agno.com/llms-full.txt"], vector_db=LanceDb( uri=str(tmp_dir.joinpath("lancedb")), table_name="agno_docs", search_type=SearchType.hybrid, embedder=OpenAIEmbedder(id="text-embedding-3-small"), ), ) web_agent = Agent( name="Web Search Agent", role="Handle web search requests", model=OpenAIChat(id="gpt-4o"), tools=[DuckDuckGoTools()], instructions=["Always include sources"], ) team_with_knowledge = Team( name="Team with Knowledge", members=[web_agent], model=OpenAIChat(id="gpt-4o"), knowledge=agno_docs_knowledge, show_members_responses=True, markdown=True, ) if __name__ == "__main__": # Set to False after the knowledge base is loaded load_knowledge = True if load_knowledge: agno_docs_knowledge.load() team_with_knowledge.print_response("Tell me about the Agno framework", stream=True) \`\`\` The team can also manage user memories: \`\`\`python from agno.team import Team from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory [V0_FILE]markdown:file="agno/knowledge/initialize_memory_v2.md" --- title: Initialize memory.v2 category: knowledge source_lines: 62921-62927 line_count: 6 --- # Initialize memory.v2 memory = Memory( # Use any model for creating memories model=OpenAIChat(id="gpt-4.1"), db=SqliteMemoryDb(table_name="user_memories", db_file=db_file), ) [V0_FILE]markdown:file="agno/knowledge/initialize_memory_with_a_model_for_agentic_search.md" --- title: Initialize memory with a model for agentic search category: knowledge source_lines: 63207-63223 line_count: 16 --- # Initialize memory with a model for agentic search memory = Memory(model=Gemini(id="gemini-2.0-flash-exp")) john_doe_id = "john_doe@example.com" memory.add_user_memory( memory=UserMemory(memory="The user enjoys hiking in the mountains on weekends"), user_id=john_doe_id, ) memory.add_user_memory( memory=UserMemory( memory="The user enjoys reading science fiction novels before bed" ), user_id=john_doe_id, ) [V0_FILE]markdown:file="agno/knowledge/initialize_memory_with_the_storage_backend.md" --- title: Initialize Memory with the storage backend category: knowledge source_lines: 63335-63338 line_count: 3 --- # Initialize Memory with the storage backend memory = Memory(db=memory_db) [V0_FILE]markdown:file="agno/knowledge/initialize_the_agent_with_the_combined_knowledge_base.md" --- title: Initialize the Agent with the combined knowledge base category: knowledge source_lines: 13065-13073 line_count: 8 --- # Initialize the Agent with the combined knowledge base agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) knowledge_base.load(recreate=False) [V0_FILE]markdown:file="agno/knowledge/initialize_the_agent_with_the_knowledge_base.md" --- title: Initialize the Agent with the knowledge_base category: knowledge source_lines: 61620-61653 line_count: 33 --- # Initialize the Agent with the knowledge_base agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) if __name__ == "__main__": # Comment out after first run asyncio.run(knowledge_base.aload(recreate=False)) # Create and use the agent asyncio.run( agent.aprint_response( "Ask anything from the json knowledge base", markdown=True ) ) \`\`\` ## Params | Parameter | Type | Default | Description | | --------- | ------------------ | -------------- | ---------------------------------------------------------------------------------------- | | `path` | `Union[str, Path]` | - | Path to `JSON` files.
Can point to a single JSON file or a directory of JSON files. | | `reader` | `JSONReader` | `JSONReader()` | A `JSONReader` that converts the `JSON` files into `Documents` for the vector database. | `JSONKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Sync loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/json_kb.py) * View [Async loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/json_kb_async.py) [V0_FILE]markdown:file="agno/knowledge/initialize_the_agent_with_the_knowledge_base_and_filters.md" --- title: Initialize the Agent with the knowledge base and filters category: knowledge source_lines: 16320-16327 line_count: 7 --- # Initialize the Agent with the knowledge base and filters agent = Agent( knowledge=knowledge_base, search_knowledge=True, knowledge_filters={"user_id": "jordan_mitchell"}, ) [V0_FILE]markdown:file="agno/knowledge/initialize_the_agent_with_various_configurations_including_the_knowledge_base_and_storage.md" --- title: Initialize the Agent with various configurations including the knowledge base and storage category: knowledge source_lines: 21383-21394 line_count: 11 --- # Initialize the Agent with various configurations including the knowledge base and storage agent = Agent( session_id="session_id", # use any unique identifier to identify the run user_id="user", # user identifier to identify the user model=model, knowledge=knowledge_base, storage=storage, show_tool_calls=True, debug_mode=True, # Enable debug mode for additional information ) [V0_FILE]markdown:file="agno/knowledge/initialize_the_assistant_with_the_knowledge_base.md" --- title: Initialize the Assistant with the knowledge_base category: knowledge source_lines: 62562-62595 line_count: 33 --- # Initialize the Assistant with the knowledge_base agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) if __name__ == "__main__": # Comment out after first run asyncio.run(knowledge_base.aload(recreate=False)) asyncio.run( agent.aprint_response( "What knowledge is available in my knowledge base?", markdown=True ) ) \`\`\` ## Params | Parameter | Type | Default | Description | | --------- | ------------------ | -------------- | ------------------------------------------------------------------------------------- | | `path` | `Union[str, Path]` | - | Path to text files. Can point to a single text file or a directory of text files. | | `formats` | `List[str]` | `[".txt"]` | Formats accepted by this knowledge base. | | `reader` | `TextReader` | `TextReader()` | A `TextReader` that converts the text files into `Documents` for the vector database. | `TextKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Sync loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/text_kb.py) * View [Async loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/text_kb_async.py) [V0_FILE]markdown:file="agno/knowledge/initialize_the_docxknowledgebase.md" --- title: Initialize the DocxKnowledgeBase category: knowledge source_lines: 13738-13750 line_count: 12 --- # Initialize the DocxKnowledgeBase knowledge_base = DocxKnowledgeBase( vector_db=vector_db, num_documents=5, ) knowledge_base.load_document( path=downloaded_cv_paths[0], metadata={"user_id": "jordan_mitchell", "document_type": "cv", "year": 2025}, recreate=True, # Set to True only for the first run, then set to False ) [V0_FILE]markdown:file="agno/knowledge/initialize_the_jsonknowledgebase.md" --- title: Initialize the JSONKnowledgeBase category: knowledge source_lines: 15357-15369 line_count: 12 --- # Initialize the JSONKnowledgeBase knowledge_base = JSONKnowledgeBase( vector_db=vector_db, num_documents=5, ) knowledge_base.load_document( path=downloaded_cv_paths[0], metadata={"user_id": "jordan_mitchell", "document_type": "cv", "year": 2025}, recreate=True, # Set to True only for the first run, then set to False ) [V0_FILE]markdown:file="agno/knowledge/initialize_the_markdownknowledgebase.md" --- title: Initialize the MarkdownKnowledgeBase category: knowledge source_lines: 61929-61936 line_count: 7 --- # Initialize the MarkdownKnowledgeBase knowledge_base = MarkdownKnowledgeBase( path=Path("tmp/mds"), vector_db=vector_db, num_documents=5, ) [V0_FILE]markdown:file="agno/knowledge/initialize_the_pdfknowledgebase.md" --- title: Initialize the PDFKnowledgeBase category: knowledge source_lines: 15716-15728 line_count: 12 --- # Initialize the PDFKnowledgeBase knowledge_base = PDFKnowledgeBase( vector_db=vector_db, num_documents=5, ) knowledge_base.load_document( path=downloaded_cv_paths[0], metadata={"user_id": "jordan_mitchell", "document_type": "cv", "year": 2025}, recreate=True, # Set to True only for the first run, then set to False ) [V0_FILE]markdown:file="agno/knowledge/initialize_the_textknowledgebase.md" --- title: Initialize the TextKnowledgeBase category: knowledge source_lines: 62555-62562 line_count: 7 --- # Initialize the TextKnowledgeBase knowledge_base = TextKnowledgeBase( path=Path("tmp/docs"), vector_db=vector_db, num_documents=5, ) [V0_FILE]markdown:file="agno/knowledge/json_knowledge_base.md" --- title: JSON Knowledge Base category: knowledge source_lines: 67654-67664 line_count: 10 --- # JSON Knowledge Base Source: https://docs.agno.com/reference/knowledge/json JSONKnowledge is a knowledge base class that allows you to load and query data from JSON files. [V0_FILE]markdown:file="agno/knowledge/knowledge.md" --- title: Knowledge category: knowledge source_lines: 470-492 line_count: 22 --- # Knowledge Source: https://docs.agno.com/agents/knowledge **Knowledge** is domain-specific information that the Agent can **search** at runtime to make better decisions (dynamic few-shot learning) and provide accurate responses (agentic RAG). Knowledge is stored in a vector db and this **searching on demand** pattern is called Agentic RAG. Example: If we're building a Text2Sql Agent, we'll need to give the table schemas, column names, data types, example queries, common "gotchas" to help it generate the best-possible SQL query. We're obviously not going to put this all in the system prompt, instead we store this information in a vector database and let the Agent query it at runtime. Using this information, the Agent can then generate the best-possible SQL query. This is called dynamic few-shot learning. **Agno Agents use Agentic RAG** by default, meaning when we provide `knowledge` to an Agent, it will search this knowledge base, at runtime, for the specific information it needs to achieve its task. The pseudo steps for adding knowledge to an Agent are: \`\`\`python from agno.agent import Agent, AgentKnowledge [V0_FILE]markdown:file="agno/knowledge/knowledge_agent_prompt.md" --- title: knowledge_agent prompt category: knowledge source_lines: 9819-9820 line_count: 1 --- # knowledge_agent prompt [V0_FILE]markdown:file="agno/knowledge/knowledge_base.md" --- title: Knowledge Base category: knowledge source_lines: 80604-80691 line_count: 87 --- # Knowledge Base knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) def lancedb_agent(user: str = "user"): run_id: Optional[str] = None agent = Agent( run_id=run_id, user_id=user, knowledge=knowledge_base, show_tool_calls=True, debug_mode=True, ) if run_id is None: run_id = agent.run_id print(f"Started Run: {run_id}\n") else: print(f"Continuing Run: {run_id}\n") while True: message = Prompt.ask(f"[bold] :sunglasses: {user} [/bold]") if message in ("exit", "bye"): break agent.print_response(message) if __name__ == "__main__": # Comment out after first run knowledge_base.load(recreate=True) typer.run(lancedb_agent) \`\`\`

LanceDB also supports asynchronous operations, enabling concurrency and leading to better performance.

\`\`\`python async_lance_db.py # install lancedb - `pip install lancedb` import asyncio from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.lancedb import LanceDb # Initialize LanceDB vector_db = LanceDb( table_name="recipes", uri="tmp/lancedb", # You can change this path to store data elsewhere ) # Create knowledge base knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) agent = Agent(knowledge=knowledge_base, show_tool_calls=True, debug_mode=True) if __name__ == "__main__": # Load knowledge base asynchronously asyncio.run(knowledge_base.aload(recreate=False)) # Comment out after first run # Create and use the agent asynchronously asyncio.run(agent.aprint_response("How to make Tom Kha Gai", markdown=True)) \`\`\` Use aload() and aprint\_response() methods with asyncio.run() for non-blocking operations in high-throughput applications.

## LanceDb Params ## Developer Resources * View [Cookbook (Sync)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/lance_db/lance_db.py) * View [Cookbook (Async)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/lance_db/async_lance_db.py) [V0_FILE]markdown:file="agno/knowledge/knowledge_base_is_now_loaded.md" --- title: Knowledge base is now loaded category: knowledge source_lines: 61028-61029 line_count: 1 --- # Knowledge base is now loaded [V0_FILE]markdown:file="agno/knowledge/knowledge_base_load_recreate_false_comment_out_after_first_run.md" --- title: knowledge_base.load(recreate=False) # Comment out after first run category: knowledge source_lines: 80013-80099 line_count: 86 --- # knowledge_base.load(recreate=False) # Comment out after first run agent = Agent( model=MistralChat(provider="mistral-large-latest", api_key=os.getenv("MISTRAL_API_KEY")), knowledge=knowledge_base, show_tool_calls=True, ) agent.print_response( "What are the health benefits of Khao Niew Dam Piek Maphrao Awn?", markdown=True, show_full_reasoning=True ) \`\`\`

Cassandra also supports asynchronous operations, enabling concurrency and leading to better performance.

\`\`\`python async_cassandra.py import asyncio from agno.agent import Agent from agno.embedder.mistral import MistralEmbedder from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.models.mistral import MistralChat from agno.vectordb.cassandra import Cassandra try: from cassandra.cluster import Cluster # type: ignore except (ImportError, ModuleNotFoundError): raise ImportError( "Could not import cassandra-driver python package.Please install it with pip install cassandra-driver." ) cluster = Cluster() session = cluster.connect() session.execute( """ CREATE KEYSPACE IF NOT EXISTS testkeyspace WITH REPLICATION = { 'class' : 'SimpleStrategy', 'replication_factor' : 1 } """ ) knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=Cassandra( table_name="recipes", keyspace="testkeyspace", session=session, embedder=MistralEmbedder(), ), ) agent = Agent( model=MistralChat(), knowledge=knowledge_base, show_tool_calls=True, ) if __name__ == "__main__": # Comment out after first run asyncio.run(knowledge_base.aload(recreate=False)) # Create and use the agent asyncio.run( agent.aprint_response( "What are the health benefits of Khao Niew Dam Piek Maphrao Awn?", markdown=True, ) ) \`\`\` Use aload() and aprint\_response() methods with asyncio.run() for non-blocking operations in high-throughput applications.

## Developer Resources * View [Cookbook (Sync)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/cassandra_db/cassandra_db.py) * View [Cookbook (Async)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/cassandra_db/async_cassandra_db.py) [V0_FILE]markdown:file="agno/knowledge/knowledge_base_load_recreate_true_comment_out_after_first_run.md" --- title: knowledge_base.load(recreate=True) # Comment out after first run category: knowledge source_lines: 80831-80892 line_count: 61 --- # knowledge_base.load(recreate=True) # Comment out after first run agent = Agent(knowledge=knowledge_base, show_tool_calls=True) agent.print_response("How to make Thai curry?", markdown=True) \`\`\`

MongoDB also supports asynchronous operations, enabling concurrency and leading to better performance.

\`\`\`python async_mongodb.py import asyncio from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.mongodb import MongoDb # MongoDB Atlas connection string """ Example connection strings: "mongodb+srv://:@cluster0.mongodb.net/?retryWrites=true&w=majority" "mongodb://localhost:27017/agno?authSource=admin" """ mdb_connection_string = "mongodb+srv://:@cluster0.mongodb.net/?retryWrites=true&w=majority" knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=MongoDb( collection_name="recipes", db_url=mdb_connection_string, ), ) # Create and use the agent agent = Agent(knowledge=knowledge_base, show_tool_calls=True) if __name__ == "__main__": # Comment out after the first run asyncio.run(knowledge_base.aload(recreate=False)) asyncio.run(agent.aprint_response("How to make Thai curry?", markdown=True)) \`\`\` Use aload() and aprint\_response() methods with asyncio.run() for non-blocking operations in high-throughput applications.

## MongoDB Params ## Developer Resources * View [Cookbook (Sync)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/mongo_db/mongo_db.py) * View [Cookbook (Async)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/mongo_db/async_mongo_db.py) [V0_FILE]markdown:file="agno/knowledge/knowledge_base_setup_same_as_synchronous_example.md" --- title: Knowledge base setup (same as synchronous example) category: knowledge source_lines: 61097-61104 line_count: 7 --- # Knowledge base setup (same as synchronous example) embedder = OpenAIEmbedder(id="text-embedding-3-small") vector_db = Qdrant(collection="thai-recipes", url="http://localhost:6333", embedder=embedder) knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) [V0_FILE]markdown:file="agno/knowledge/knowledge_filtering_using_traditional_rag.md" --- title: Knowledge filtering using Traditional RAG category: knowledge source_lines: 13824-13840 line_count: 16 --- # Knowledge filtering using Traditional RAG Source: https://docs.agno.com/examples/concepts/knowledge/filters/filtering-traditional-RAG Learn how to filter knowledge in Traditional RAG using metadata like user IDs, document types, and years. This example demonstrates how to set up a knowledge base with filters and query it effectively. ## Code \`\`\`python filtering-traditional-RAG.py from agno.agent import Agent from agno.knowledge.text import TextKnowledgeBase from agno.utils.media import ( SampleDataFileExtension, download_knowledge_filters_sample_data, ) from agno.vectordb.lancedb import LanceDb [V0_FILE]markdown:file="agno/knowledge/knowledge_filters.md" --- title: Knowledge Filters category: knowledge source_lines: 59197-59277 line_count: 80 --- # Knowledge Filters Knowledge filters allow you to restrict and refine searches within your knowledge base using metadata such as user IDs, document types, years, and more. This feature is especially useful when you have a large collection of documents and want to retrieve information relevant to specific users or contexts. ## Why Use Knowledge Filters? * **Personalization:** Retrieve information for a specific user or group. * **Security:** Restrict access to sensitive documents. * **Efficiency:** Reduce noise by narrowing down search results. ## How Do Knowledge Filters Work? When you load documents into your knowledge base, you can attach metadata (like user ID, document type, year, etc.). Later, when querying, you can specify filters to only search documents matching certain criteria. **Example Metadata:** \`\`\`python { "user_id": "jordan_mitchell", "document_type": "cv", "year": 2025, } \`\`\` ## Ways to Apply Filters You can apply knowledge filters in two main ways: 1. **Manual Filters:** Explicitly pass filters when querying. 2. **Agentic Filters:** Let the Agent automatically extract filters from your query. > **Tip:** You can combine multiple filters for more precise results! ## Filters in Traditional RAG vs. Agentic RAG When configuring your Agent it is important to choose the right approach for your use case. There are two broad approaches to RAG with Agno agents: traditional RAG and agentic RAG. With a traditional RAG approach you set `add_references=True` to ensure that references are included in the system message sent to the LLM. For Agentic RAG, you set `search_knowledge=True` to leverage the agent's ability search the knowledge base directly. Example: \`\`\`python agent = Agent( name="KnowledgeFilterAgent", search_knowledge=False, # Do not use agentic search add_references=True, # Add knowledge base references to the system prompt knowledge_filters={"user_id": "jordan_mitchell"}, # Pass filters like this ) \`\`\` Remember to use only one of these configurations at a time, setting the other to false. By default, `search_knowledge=True` is preferred as it offers a more dynamic and interactive experience. Checkout an example [here](/examples/concepts/knowledge/filters/filtering-traditional-RAG) of how to set up knowledge filters in a Traditional RAG system ## Best Practices * Make your prompts descriptive (e.g., include user names, document types, years). * Use agentic filtering for interactive applications or chatbots. ## Manual vs. Agentic Filtering | Manual Filtering | Agentic Filtering | | ------------------------ | -------------------------------- | | Explicit filters in code | Filters inferred from query text | | Full control | More natural, less code | | Good for automation | Good for user-facing apps | 🚦 **Currently, knowledge filtering is supported on the following vector databases:** * **Qdrant** * **LanceDB** * **PgVector** * **MongoDB** * **Pinecone** * **Weaviate** * **ChromaDB** * **Milvus** [V0_FILE]markdown:file="agno/knowledge/knowledge_filters_cuisine_cape.md" --- title: knowledge_filters={"cuisine": "Cape"}, category: knowledge source_lines: 16065-16066 line_count: 1 --- # knowledge_filters={"cuisine": "Cape"}, [V0_FILE]markdown:file="agno/knowledge/knowledge_filters_user_id_jordan_mitchell.md" --- title: knowledge_filters = {"user_id": "jordan_mitchell"}, category: knowledge source_lines: 16461-16462 line_count: 1 --- # knowledge_filters = {"user_id": "jordan_mitchell"}, [V0_FILE]markdown:file="agno/knowledge/knowledge_filters_user_id_taylor_brooks.md" --- title: knowledge_filters={"user_id": "taylor_brooks"}, category: knowledge source_lines: 16342-16343 line_count: 1 --- # knowledge_filters={"user_id": "taylor_brooks"}, [V0_FILE]markdown:file="agno/knowledge/knowledge_knowledge_base.md" --- title: knowledge=knowledge_base, category: knowledge source_lines: 16456-16457 line_count: 1 --- # knowledge=knowledge_base, [V0_FILE]markdown:file="agno/knowledge/knowledge_tools.md" --- title: Knowledge Tools category: knowledge source_lines: 72526-72559 line_count: 33 --- # Knowledge Tools Source: https://docs.agno.com/tools/reasoning_tools/knowledge-tools The `KnowledgeTools` toolkit enables Agents to search, retrieve, and analyze information from knowledge bases. This toolkit integrates with `AgentKnowledge` and provides a structured workflow for finding and evaluating relevant information before responding to users. The toolkit implements a "Think → Search → Analyze" cycle that allows an Agent to: 1. Think through the problem and plan search queries 2. Search the knowledge base for relevant information 3. Analyze the results to determine if they are sufficient or if additional searches are needed This approach significantly improves an Agent's ability to provide accurate information by giving it tools to find, evaluate, and synthesize knowledge. The toolkit includes the following tools: * `think`: A scratchpad for planning, brainstorming keywords, and refining approaches. These thoughts remain internal to the Agent and are not shown to users. * `search`: Executes queries against the knowledge base to retrieve relevant documents. * `analyze`: Evaluates whether the returned documents are correct and sufficient, determining if further searches are needed. ## Example Here's an example of how to use the `KnowledgeTools` toolkit: \`\`\`python from agno.agent import Agent from agno.embedder.openai import OpenAIEmbedder from agno.knowledge.url import UrlKnowledge from agno.models.openai import OpenAIChat from agno.tools.knowledge import KnowledgeTools from agno.vectordb.lancedb import LanceDb, SearchType [V0_FILE]markdown:file="agno/knowledge/lancedb_agent_knowledge.md" --- title: LanceDB Agent Knowledge category: knowledge source_lines: 80574-80597 line_count: 23 --- # LanceDB Agent Knowledge Source: https://docs.agno.com/vectordb/lancedb ## Setup \`\`\`shell pip install lancedb \`\`\` ## Example \`\`\`python agent_with_knowledge.py import typer from typing import Optional from rich.prompt import Prompt from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.lancedb import LanceDb from agno.vectordb.search import SearchType [V0_FILE]markdown:file="agno/knowledge/langchain_knowledge_base.md" --- title: Langchain Knowledge Base category: knowledge source_lines: 67664-67674 line_count: 10 --- # Langchain Knowledge Base Source: https://docs.agno.com/reference/knowledge/langchain LangChainKnowledge is a knowledge base class that allows you to load and query data from Langchain supported knowledge bases. [V0_FILE]markdown:file="agno/knowledge/lightrag_knowledge_base.md" --- title: LightRAG Knowledge Base category: knowledge source_lines: 61717-61733 line_count: 16 --- # LightRAG Knowledge Base Source: https://docs.agno.com/knowledge/lightrag Learn how to use LightRAG, a fast graph-based retrieval-augmented generation system for enhanced knowledge querying. The **LightRAGKnowledgeBase** integrates with a [LightRAG Server](https://github.com/HKUDS/LightRAG), a simple and fast retrieval-augmented generation system that uses graph structures to enhance document retrieval and knowledge querying capabilities. ## Usage \`\`\`python knowledge_base.py import asyncio from agno.agent import Agent from agno.knowledge.light_rag import LightRagKnowledgeBase, lightrag_retriever from agno.models.anthropic import Claude [V0_FILE]markdown:file="agno/knowledge/llamaindex_knowledge_base.md" --- title: LlamaIndex Knowledge Base category: knowledge source_lines: 67674-67684 line_count: 10 --- # LlamaIndex Knowledge Base Source: https://docs.agno.com/reference/knowledge/llamaindex LlamaIndexKnowledge is a knowledge base class that allows you to load and query data from LlamaIndex supported knowledge bases. [V0_FILE]markdown:file="agno/knowledge/load_agno_documentation_in_a_knowledge_base.md" --- title: Load Agno documentation in a knowledge base category: knowledge source_lines: 60038-60039 line_count: 1 --- # Load Agno documentation in a knowledge base [V0_FILE]markdown:file="agno/knowledge/load_pdf_knowledge_base_using_hybrid_search.md" --- title: Load PDF knowledge base using hybrid search category: knowledge source_lines: 61400-61406 line_count: 6 --- # Load PDF knowledge base using hybrid search knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=hybrid_db, ) [V0_FILE]markdown:file="agno/knowledge/load_the_knowledge_base.md" --- title: Load the knowledge base category: knowledge source_lines: 80409-80412 line_count: 3 --- # Load the knowledge base knowledge_base.load(recreate=True) [V0_FILE]markdown:file="agno/knowledge/load_the_knowledge_base_comment_after_first_run_as_the_knowledge_base_is_already_loaded.md" --- title: Load the knowledge base: Comment after first run as the knowledge base is already loaded category: knowledge source_lines: 21530-21592 line_count: 62 --- # Load the knowledge base: Comment after first run as the knowledge base is already loaded knowledge_base.load(upsert=True) agent = Agent( model=OpenAIChat(id="gpt-4o"), knowledge=knowledge_base, # Enable RAG by adding context from the `knowledge` to the user prompt. add_references=True, # Set as False because Agents default to `search_knowledge=True` search_knowledge=False, markdown=True, ) agent.print_response( "How do I make chicken and galangal in coconut milk soup", stream=True ) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai sqlalchemy 'psycopg[binary]' pgvector pypdf agno \`\`\` \`\`\`bash docker run -d \ -e POSTGRES_DB=ai \ -e POSTGRES_USER=ai \ -e POSTGRES_PASSWORD=ai \ -e PGDATA=/var/lib/postgresql/data/pgdata \ -v pgvolume:/var/lib/postgresql/data \ -p 5532:5432 \ --name pgvector \ agnohq/pgvector:16 \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/rag/traditional_rag_pgvector.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/rag/traditional_rag_pgvector.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/load_the_knowledge_base_comment_out_after_first_run.md" --- title: Load the knowledge base: Comment out after first run category: knowledge source_lines: 80924-80988 line_count: 64 --- # Load the knowledge base: Comment out after first run knowledge_base.load(recreate=True, upsert=True) agent = Agent( model=OpenAIChat(id="gpt-4o"), knowledge=knowledge_base, # Add a tool to read chat history. read_chat_history=True, show_tool_calls=True, markdown=True, # debug_mode=True, ) agent.print_response("How do I make chicken and galangal in coconut milk soup", stream=True) agent.print_response("What was my last question?", stream=True) \`\`\`

PgVector also supports asynchronous operations, enabling concurrency and leading to better performance.

\`\`\`python async_pgvector.py import asyncio from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.pgvector import PgVector db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai" vector_db = PgVector(table_name="recipes", db_url=db_url) knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) agent = Agent(knowledge=knowledge_base, show_tool_calls=True) if __name__ == "__main__": # Comment out after first run asyncio.run(knowledge_base.aload(recreate=False)) # Create and use the agent asyncio.run(agent.aprint_response("How to make Tom Kha Gai", markdown=True)) \`\`\` Use aload() and aprint\_response() methods with asyncio.run() for non-blocking operations in high-throughput applications.

## PgVector Params ## Developer Resources * View [Cookbook (Sync)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/pgvector_db/pg_vector.py) * View [Cookbook (Async)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/pgvector_db/async_pg_vector.py) [V0_FILE]markdown:file="agno/knowledge/load_the_knowledge_base_with_the_documents_from_the_local_directory_and_the_url.md" --- title: Load the knowledge base with the documents from the local directory and the URL category: knowledge source_lines: 61740-61743 line_count: 3 --- # Load the knowledge base with the documents from the local directory and the URL asyncio.run(knowledge_base.load()) [V0_FILE]markdown:file="agno/knowledge/load_the_knowledge_base_with_the_text.md" --- title: Load the knowledge base with the text category: knowledge source_lines: 61743-61753 line_count: 10 --- # Load the knowledge base with the text asyncio.run( knowledge_base.load_text(text="Agno is the best framework for building Agents") ) \`\`\` Then use the `lightrag_knowledge_base` with an Agent: \`\`\`python agent.py [V0_FILE]markdown:file="agno/knowledge/load_the_knowledge_base_without_recreating_it_if_it_already_exists_in_vector_lancedb.md" --- title: Load the knowledge base without recreating it if it already exists in Vector LanceDB category: knowledge source_lines: 21376-21379 line_count: 3 --- # Load the knowledge base without recreating it if it already exists in Vector LanceDB knowledge_base.load(recreate=False) [V0_FILE]markdown:file="agno/knowledge/manual_knowledge_filters.md" --- title: Manual Knowledge Filters category: knowledge source_lines: 59282-59407 line_count: 125 --- # Manual Knowledge Filters Manual filtering gives you full control over which documents are searched by specifying filters directly in your code. ## Step 1: Attach Metadata There are two ways to attach metadata to your documents: 1. **Attach Metadata When Initializing the Knowledge Base** \`\`\`python knowledge_base = PDFKnowledgeBase( path=[ { "path": "path/to/cv1.pdf", "metadata": { "user_id": "jordan_mitchell", "document_type": "cv", "year": 2025, }, }, # ... more documents ... ], vector_db=vector_db, ) knowledge_base.load(recreate=True) \`\`\` 2. **Attach Metadata When Loading Documents One by One** \`\`\`python # Initialize the PDFKnowledgeBase knowledge_base = PDFKnowledgeBase( vector_db=vector_db, num_documents=5, ) # Load first document with user_1 metadata knowledge_base.load_document( path=path/to/cv1.pdf, metadata={"user_id": "jordan_mitchell", "document_type": "cv", "year": 2025}, recreate=True, # Set to True only for the first run, then set to False ) # Load second document with user_2 metadata knowledge_base.load_document( path=path/to/cv2.pdf, metadata={"user_id": "taylor_brooks", "document_type": "cv", "year": 2025}, ) \`\`\` *** > 💡 **Tips:**\ > • Use **Option 1** if you have all your documents and metadata ready at once.\ > • Use **Option 2** if you want to add documents incrementally or as they become available. ## Step 2: Query with Filters You can pass filters in two ways: ### 1. On the Agent (applies to all queries) \`\`\`python agent = Agent( knowledge=knowledge_base, search_knowledge=True, knowledge_filters={"user_id": "jordan_mitchell"}, ) agent.print_response( "Tell me about Jordan Mitchell's experience and skills", markdown=True, ) \`\`\` ### 2. On Each Query (overrides Agent filters for that run) \`\`\`python agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) agent.print_response( "Tell me about Jordan Mitchell's experience and skills", knowledge_filters={"user_id": "jordan_mitchell"}, markdown=True, ) \`\`\` If you pass filters both on the Agent and on the query, the query-level filters take precedence. ## Combining Multiple Filters You can filter by multiple fields: \`\`\`python agent = Agent( knowledge=knowledge_base, search_knowledge=True, knowledge_filters={ "user_id": "jordan_mitchell", "document_type": "cv", "year": 2025, } ) agent.print_response( "Tell me about Jordan Mitchell's experience and skills", markdown=True, ) \`\`\` ## Try It Yourself! * Load documents with different metadata. * Query with different filter combinations. * Observe how the results change! *** ## Developer Resources * [Manual filtering](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/filters/pdf/filtering.py) * [Manual filtering on load](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/filters/pdf/filtering_on_load.py) [V0_FILE]markdown:file="agno/knowledge/markdown_knowledge_base.md" --- title: Markdown Knowledge Base category: knowledge source_lines: 61868-61929 line_count: 61 --- # Markdown Knowledge Base Source: https://docs.agno.com/knowledge/markdown Learn how to use Markdown files in your knowledge base. The **MarkdownKnowledgeBase** reads **local markdown** files, converts them into vector embeddings and loads them to a vector database. ## Usage We are using a local PgVector database for this example. [Make sure it's running](https://docs.agno.com/vectordb/pgvector) \`\`\`python knowledge_base.py from agno.knowledge.markdown import MarkdownKnowledgeBase from agno.vectordb.pgvector import PgVector knowledge_base = MarkdownKnowledgeBase( path="data/markdown_files", vector_db=PgVector( table_name="markdown_documents", db_url="postgresql+psycopg://ai:ai@localhost:5532/ai", ), ) \`\`\` Then use the `knowledge_base` with an Agent: \`\`\`python agent.py from agno.agent import Agent from knowledge_base import knowledge_base agent = Agent( knowledge_base=knowledge_base, search_knowledge=True, ) agent.knowledge.load(recreate=False) agent.print_response("Ask me about something from the knowledge base") \`\`\` #### MarkdownKnowledgeBase also supports async loading. \`\`\`shell pip install qdrant-client \`\`\` We are using a local Qdrant database for this example. [Make sure it's running](https://docs.agno.com/vectordb/qdrant) \`\`\`python async_knowledge_base.py import asyncio from pathlib import Path from agno.agent import Agent from agno.knowledge.markdown import MarkdownKnowledgeBase from agno.vectordb.qdrant import Qdrant COLLECTION_NAME = "essay-txt" vector_db = Qdrant(collection=COLLECTION_NAME, url="http://localhost:6333") [V0_FILE]markdown:file="agno/knowledge/mem0_memory.md" --- title: Mem0 Memory category: knowledge source_lines: 18283-18304 line_count: 21 --- # Mem0 Memory Source: https://docs.agno.com/examples/concepts/memory/mem0-memory ## Code \`\`\`python cookbook/agent_concepts/memory/mem0_memory.py from agno.agent import Agent, RunResponse from agno.models.openai import OpenAIChat from agno.utils.pprint import pprint_run_response from mem0 import MemoryClient client = MemoryClient() user_id = "agno" messages = [ {"role": "user", "content": "My name is John Billings."}, {"role": "user", "content": "I live in NYC."}, {"role": "user", "content": "I'm going to a concert tomorrow."}, ] [V0_FILE]markdown:file="agno/knowledge/mem0_memory_tools.md" --- title: Mem0 Memory Tools category: knowledge source_lines: 26141-26213 line_count: 72 --- # Mem0 Memory Tools Source: https://docs.agno.com/examples/concepts/tools/database/mem0 ## Code \`\`\`python cookbook/tools/mem0_tools.py from textwrap import dedent from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.mem0 import Mem0Tools USER_ID = "jane_doe" SESSION_ID = "agno_session" agent = Agent( model=OpenAIChat(id="gpt-4o"), tools=[Mem0Tools()], user_id=USER_ID, session_id=SESSION_ID, add_state_in_messages=True, markdown=True, instructions=dedent( """ You have an evolving memory of this user. Proactively capture new personal details, preferences, plans, and relevant context the user shares, and naturally bring them up in later conversation. Before answering questions about past details, recall from your memory to provide precise and personalized responses. Keep your memory concise: store only meaningful information that enhances long-term dialogue. If the user asks to start fresh, clear all remembered information and proceed anew. """ ), show_tool_calls=True, ) agent.print_response("I live in NYC") agent.print_response("I lived in San Francisco for 5 years previously") agent.print_response("I'm going to a Taylor Swift concert tomorrow") agent.print_response("Summarize all the details of the conversation") \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx export MEM0_API_KEY=xxx export MEM0_ORG_ID=xxx # Optional export MEM0_PROJECT_ID=xxx # Optional \`\`\` \`\`\`bash pip install -U mem0ai openai agno \`\`\` \`\`\`bash python cookbook/tools/mem0_tools.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/memory.md" --- title: Memory category: knowledge source_lines: 67744-67754 line_count: 10 --- # Memory Source: https://docs.agno.com/reference/memory/memory Memory is a class that manages conversation history, session summaries, and long-term user memories for AI agents. It provides comprehensive memory management capabilities including adding new memories, searching memories, and deleting memories. [V0_FILE]markdown:file="agno/knowledge/memory_add_user_memory.md" --- title: memory.add_user_memory( category: knowledge source_lines: 16670-16671 line_count: 1 --- # memory.add_user_memory( [V0_FILE]markdown:file="agno/knowledge/memory_memory.md" --- title: memory=memory, category: knowledge source_lines: 17791-17792 line_count: 1 --- # memory=memory, [V0_FILE]markdown:file="agno/knowledge/memory_operations_work_the_same_regardless_of_the_backend.md" --- title: Memory operations work the same regardless of the backend category: knowledge source_lines: 63348-63363 line_count: 15 --- # Memory operations work the same regardless of the backend print(f"User has {len(memory.get_user_memories(user_id=user_id))} memories stored") \`\`\` ## Integrating with Agent Storage When building agents with memory, you'll often want to store both agent sessions and memories. Agno makes this easy by allowing you to configure both storage systems: \`\`\`python from agno.agent import Agent from agno.memory.v2.memory import Memory from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.models.openai import OpenAIChat from agno.storage.sqlite import SqliteStorage [V0_FILE]markdown:file="agno/knowledge/memory_references.md" --- title: Memory References category: knowledge source_lines: 17757-17788 line_count: 31 --- # Memory References Source: https://docs.agno.com/examples/concepts/memory/12-memory-references This example shows how to use the `add_memory_references` parameter in the Agent config to add references to the user memories to the Agent. ## Code \`\`\`12_memory_references.py from agno.agent.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory, UserMemory from agno.models.google.gemini import Gemini memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") memory = Memory(db=memory_db) memory.add_user_memory( memory=UserMemory(memory="I like to play soccer", topics=["soccer"]), user_id="john_doe@example.com", ) agent = Agent( model=Gemini(id="gemini-2.0-flash-exp"), memory=memory, add_memory_references=True, # Add pre existing memories to the Agent but don't create new ones ) [V0_FILE]markdown:file="agno/knowledge/memory_search.md" --- title: Memory Search category: knowledge source_lines: 16827-16847 line_count: 20 --- # Memory Search Source: https://docs.agno.com/examples/concepts/memory/04-memory-search This example demonstrates how to search for user memories using different retrieval methods * last\_n: Retrieves the last n memories * first\_n: Retrieves the first n memories * semantic: Retrieves memories using semantic search ## Code \`\`\`python cookbook/agent_concepts/memory/05_memory_search.py from agno.memory.v2 import Memory, UserMemory from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.models.google.gemini import Gemini from rich.pretty import pprint memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") [V0_FILE]markdown:file="agno/knowledge/memory_storage.md" --- title: Memory Storage category: knowledge source_lines: 63247-63272 line_count: 25 --- # Memory Storage To persist memories across sessions and execution cycles, store memories in a persistent storage like a database. If you're using Memory in production, persistent storage is critical as you'd want to retain user memories across application restarts. Agno's memory system supports multiple persistent storage options. ## Storage Options The Memory class supports different backend storage options through a pluggable database interface. Currently, Agno provides: 1. [SQLite Storage](/reference/memory/storage/sqlite) 2. [PostgreSQL Storage](/reference/memory/storage/postgres) 3. [MongoDB Storage](/reference/memory/storage/mongo) 4. [Redis Storage](/reference/memory/storage/redis) ## Setting Up Storage To configure memory storage, you'll need to create a database instance and pass it to the Memory constructor: \`\`\`python from agno.memory.v2.memory import Memory from agno.memory.v2.db.sqlite import SqliteMemoryDb [V0_FILE]markdown:file="agno/knowledge/memory_usermemory_memory_the_user_works_at_a_softward_company_called_agno_topics_name.md" --- title: memory=UserMemory(memory="The user works at a softward company called Agno", topics=["name"]), category: knowledge source_lines: 16671-16672 line_count: 1 --- # memory=UserMemory(memory="The user works at a softward company called Agno", topics=["name"]), [V0_FILE]markdown:file="agno/knowledge/memory_v2.md" --- title: Memory V2 category: knowledge source_lines: 58755-58838 line_count: 83 --- # Memory V2 Source: https://docs.agno.com/faq/memoryv2 Starting with Agno version 1.4.0, **Memory V2** is now the default memory for the Agno Agent. This replaces the previous `AgentMemory` and `TeamMemory` classes which is now deprecated but still available to use. Memory V2 is a more powerful and flexible memory system that allows you to manage message history, session summaries, and long-term user memories. ## How to Continue Using AgentMemory (Memory V1) If you want to continue using `AgentMemory` and avoid breaking changes, you can do so by updating your imports. By default, the Agent now uses the `Memory` class: \`\`\`python from agno.memory.v2 import Memory \`\`\` To use the legacy AgentMemory class instead, import it like this: \`\`\`python from agno.memory import AgentMemory agent = Agent( memory=AgentMemory() ) \`\`\` ## Key Memory V2 Changes * **Accessing Messages:** * **Before:** \`\`\`python agent.memory.messages \`\`\` * **Now:** \`\`\`python [run.messages for run in agent.memory.runs] # or agent.get_messages_for_session() \`\`\` * **User Memories:** * **Before:** \`\`\`python from agno.memory import AgentMemory memory = AgentMemory(create_user_memories=True) agent = Agent(memory=memory) \`\`\` * **Now:** \`\`\`python from agno.memory.v2 import Memory memory = Memory() agent = Agent(create_user_memories=True, memory=memory) or team = Team(create_user_memories=True, memory=memory) \`\`\` * **Session Summaries:** * **Before:** \`\`\`python from agno.memory import AgentMemory memory = AgentMemory(create_session_summary=True) agent = Agent(memory=memory) \`\`\` * **Now:** \`\`\`python from agno.memory.v2 import Memory memory = Memory() agent = Agent(enable_session_summaries=True, memory=memory) or team = Team(enable_session_summaries=True, memory=memory) \`\`\` [V0_FILE]markdown:file="agno/knowledge/milvus_agent_knowledge.md" --- title: Milvus Agent Knowledge category: knowledge source_lines: 80691-80722 line_count: 31 --- # Milvus Agent Knowledge Source: https://docs.agno.com/vectordb/milvus ## Setup \`\`\`shell pip install pymilvus \`\`\` ## Initialize Milvus Set the uri and token for your Milvus server. * If you only need a local vector database for small scale data or prototyping, setting the uri as a local file, e.g.`./milvus.db`, is the most convenient method, as it automatically utilizes [Milvus Lite](https://milvus.io/docs/milvus_lite.md) to store all data in this file. * If you have large scale data, say more than a million vectors, you can set up a more performant Milvus server on [Docker or Kubernetes](https://milvus.io/docs/quickstart.md). In this setup, please use the server address and port as your uri, e.g.`http://localhost:19530`. If you enable the authentication feature on Milvus, use `your_username:your_password` as the token, otherwise don't set the token. * If you use [Zilliz Cloud](https://zilliz.com/cloud), the fully managed cloud service for Milvus, adjust the `uri` and `token`, which correspond to the [Public Endpoint and API key](https://docs.zilliz.com/docs/on-zilliz-cloud-console#cluster-details) in Zilliz Cloud. ## Example \`\`\`python agent_with_knowledge.py from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.milvus import Milvus vector_db = Milvus( collection="recipes", uri="./milvus.db", ) [V0_FILE]markdown:file="agno/knowledge/mongodb_agent_knowledge.md" --- title: MongoDB Agent Knowledge category: knowledge source_lines: 80791-80813 line_count: 22 --- # MongoDB Agent Knowledge Source: https://docs.agno.com/vectordb/mongodb ## Setup Follow the instructions in the [MongoDB Setup Guide](https://www.mongodb.com/docs/atlas/getting-started/) to get connection string Install MongoDB packages \`\`\`shell pip install "pymongo[srv]" \`\`\` ## Example \`\`\`python agent_with_knowledge.py from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.mongodb import MongoDb [V0_FILE]markdown:file="agno/knowledge/mongodb_memory_storage.md" --- title: MongoDB Memory Storage category: knowledge source_lines: 17904-17924 line_count: 20 --- # MongoDB Memory Storage Source: https://docs.agno.com/examples/concepts/memory/db/mem-mongodb-memory ## Code \`\`\`python cookbook/agent_concepts/memory/mongodb_memory.py """ This example shows how to use the Memory class with MongoDB storage. """ import asyncio import os from agno.agent.agent import Agent from agno.memory.v2.db.mongodb import MongoMemoryDb from agno.memory.v2.memory import Memory from agno.models.openai.chat import OpenAIChat [V0_FILE]markdown:file="agno/knowledge/mongomemorydb.md" --- title: MongoMemoryDb category: knowledge source_lines: 67754-67764 line_count: 10 --- # MongoMemoryDb Source: https://docs.agno.com/reference/memory/storage/mongo MongoMemoryDb is a class that implements the MemoryDb interface using MongoDB as the backend storage system. It provides persistent storage for agent memories with support for indexing and efficient querying. [V0_FILE]markdown:file="agno/knowledge/multiple_agents_sharing_memory.md" --- title: Multiple Agents Sharing Memory category: knowledge source_lines: 17247-17267 line_count: 20 --- # Multiple Agents Sharing Memory Source: https://docs.agno.com/examples/concepts/memory/08-agents-share-memory In this example, we have two agents that share the same memory. ## Code \`\`\`python cookbook/agent_concepts/memory/09_agents_share_memory.py from agno.agent.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory from agno.models.google.gemini import Gemini from agno.tools.duckduckgo import DuckDuckGoTools from rich.pretty import pprint memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") [V0_FILE]markdown:file="agno/knowledge/now_use_the_downloaded_paths_in_knowledge_base_initialization.md" --- title: Now use the downloaded paths in knowledge base initialization category: knowledge source_lines: 13600-13648 line_count: 48 --- # Now use the downloaded paths in knowledge base initialization knowledge_base = DocxKnowledgeBase( path=[ { "path": downloaded_cv_paths[0], "metadata": { "user_id": "jordan_mitchell", "document_type": "cv", "year": 2025, }, }, { "path": downloaded_cv_paths[1], "metadata": { "user_id": "taylor_brooks", "document_type": "cv", "year": 2025, }, }, { "path": downloaded_cv_paths[2], "metadata": { "user_id": "morgan_lee", "document_type": "cv", "year": 2025, }, }, { "path": downloaded_cv_paths[3], "metadata": { "user_id": "casey_jordan", "document_type": "cv", "year": 2025, }, }, { "path": downloaded_cv_paths[4], "metadata": { "user_id": "alex_rivera", "document_type": "cv", "year": 2025, }, }, ], vector_db=vector_db, ) [V0_FILE]markdown:file="agno/knowledge/pdf_bytes_knowledge_base.md" --- title: PDF Bytes Knowledge Base category: knowledge source_lines: 62070-62130 line_count: 60 --- # PDF Bytes Knowledge Base Source: https://docs.agno.com/knowledge/pdf-bytes Learn how to use in-memory PDF bytes in your knowledge base. The **PDFBytesKnowledgeBase** reads **PDF content from bytes or IO streams**, converts them into vector embeddings and loads them to a vector database. This is useful when working with dynamically generated PDFs, API responses, or file uploads without needing to save files to disk. ## Usage We are using a local LanceDB database for this example. [Make sure it's running](https://docs.agno.com/vectordb/lancedb) \`\`\`shell pip install pypdf \`\`\` \`\`\`python knowledge_base.py from agno.agent import Agent from agno.knowledge.pdf import PDFBytesKnowledgeBase from agno.vectordb.lancedb import LanceDb vector_db = LanceDb( table_name="recipes_async", uri="tmp/lancedb", ) with open("data/pdfs/ThaiRecipes.pdf", "rb") as f: pdf_bytes = f.read() knowledge_base = PDFBytesKnowledgeBase( pdfs=[pdf_bytes], vector_db=vector_db, ) knowledge_base.load(recreate=False) # Comment out after first run agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) agent.print_response("How to make Tom Kha Gai?", markdown=True) \`\`\` ## Params | Parameter | Type | Default | Description | | -------------- | --------------------------------- | ----------- | -------------------------------------------------------------------------------------------- | | pdfs | Union\[List\[bytes], List\[IO]] | - | List of PDF content as bytes or IO streams. | | exclude\_files | List\[str] | \[] | List of file patterns to exclude (inherited from base class). | | reader | Union\[PDFReader, PDFImageReader] | PDFReader() | A PDFReader or PDFImageReader that converts the PDFs into Documents for the vector database. | `PDFBytesKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Sync loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/pdf_bytes_kb.py) * View [Async loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/pdf_bytes_kb_async.py) [V0_FILE]markdown:file="agno/knowledge/pdf_knowledge_base.md" --- title: PDF Knowledge Base category: knowledge source_lines: 67684-67694 line_count: 10 --- # PDF Knowledge Base Source: https://docs.agno.com/reference/knowledge/pdf PDFKnowledge is a knowledge base class that allows you to load and query data from PDF files. [V0_FILE]markdown:file="agno/knowledge/pdf_url_knowledge_base.md" --- title: PDF URL Knowledge Base category: knowledge source_lines: 67694-67704 line_count: 10 --- # PDF URL Knowledge Base Source: https://docs.agno.com/reference/knowledge/pdf_url PDFUrlKnowledge is a knowledge base class that allows you to load and query data from PDF files stored in URLs. [V0_FILE]markdown:file="agno/knowledge/persistent_memory_with_sqlite.md" --- title: Persistent Memory with SQLite category: knowledge source_lines: 16639-16663 line_count: 24 --- # Persistent Memory with SQLite Source: https://docs.agno.com/examples/concepts/memory/02-persistent-memory This example shows how to use the Memory class to create a persistent memory. Every time you run this, the `Memory` object will be re-initialized from the DB. ## Code \`\`\`python cookbook/agent_concepts/memory/02_persistent_memory.py from typing import List from agno.memory.v2.db.schema import MemoryRow from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory from agno.memory.v2.schema import UserMemory memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") memory = Memory(db=memory_db) john_doe_id = "john_doe@example.com" [V0_FILE]markdown:file="agno/knowledge/pgvector_agent_knowledge.md" --- title: PgVector Agent Knowledge category: knowledge source_lines: 80892-80924 line_count: 32 --- # PgVector Agent Knowledge Source: https://docs.agno.com/vectordb/pgvector ## Setup \`\`\`shell docker run -d \ -e POSTGRES_DB=ai \ -e POSTGRES_USER=ai \ -e POSTGRES_PASSWORD=ai \ -e PGDATA=/var/lib/postgresql/data/pgdata \ -v pgvolume:/var/lib/postgresql/data \ -p 5532:5432 \ --name pgvector \ agnohq/pgvector:16 \`\`\` ## Example \`\`\`python agent_with_knowledge.py from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.pgvector import PgVector, SearchType db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai" knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=PgVector(table_name="recipes", db_url=db_url, search_type=SearchType.hybrid), ) [V0_FILE]markdown:file="agno/knowledge/pinecone_agent_knowledge.md" --- title: Pinecone Agent Knowledge category: knowledge source_lines: 80988-81131 line_count: 143 --- # Pinecone Agent Knowledge Source: https://docs.agno.com/vectordb/pinecone ## Setup Follow the instructions in the [Pinecone Setup Guide](https://docs.pinecone.io/guides/get-started/quickstart) to get started quickly with Pinecone. \`\`\`shell pip install pinecone \`\`\` We do not yet support Pinecone v6.x.x. We are actively working to achieve compatibility. In the meantime, we recommend using **Pinecone v5.4.2** for the best experience. ## Example \`\`\`python agent_with_knowledge.py import os import typer from typing import Optional from rich.prompt import Prompt from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.pineconedb import PineconeDb api_key = os.getenv("PINECONE_API_KEY") index_name = "thai-recipe-hybrid-search" vector_db = PineconeDb( name=index_name, dimension=1536, metric="cosine", spec={"serverless": {"cloud": "aws", "region": "us-east-1"}}, api_key=api_key, use_hybrid_search=True, hybrid_alpha=0.5, ) knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) def pinecone_agent(user: str = "user"): run_id: Optional[str] = None agent = Agent( run_id=run_id, user_id=user, knowledge=knowledge_base, show_tool_calls=True, debug_mode=True, ) if run_id is None: run_id = agent.run_id print(f"Started Run: {run_id}\n") else: print(f"Continuing Run: {run_id}\n") while True: message = Prompt.ask(f"[bold] :sunglasses: {user} [/bold]") if message in ("exit", "bye"): break agent.print_response(message) if __name__ == "__main__": # Comment out after first run knowledge_base.load(recreate=True, upsert=True) typer.run(pinecone_agent) \`\`\`

Pinecone also supports asynchronous operations, enabling concurrency and leading to better performance.

\`\`\`python async_pinecone.py import asyncio from os import getenv from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.pineconedb import PineconeDb api_key = getenv("PINECONE_API_KEY") index_name = "thai-recipe-index" vector_db = PineconeDb( name=index_name, dimension=1536, metric="cosine", spec={"serverless": {"cloud": "aws", "region": "us-east-1"}}, api_key=api_key, ) knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) agent = Agent( knowledge=knowledge_base, # Show tool calls in the response show_tool_calls=True, # Enable the agent to search the knowledge base search_knowledge=True, # Enable the agent to read the chat history read_chat_history=True, ) if __name__ == "__main__": # Comment out after first run asyncio.run(knowledge_base.aload(recreate=False, upsert=True)) # Create and use the agent asyncio.run(agent.aprint_response("How to make Tom Kha Gai", markdown=True)) \`\`\` Use aload() and aprint\_response() methods with asyncio.run() for non-blocking operations in high-throughput applications.

## PineconeDb Params ## Developer Resources * View [Cookbook (Sync)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/pinecone_db/pinecone_db.py) * View [Cookbook (Async)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/pinecone_db/async_pinecone_db.py) [V0_FILE]markdown:file="agno/knowledge/postgresmemorydb.md" --- title: PostgresMemoryDb category: knowledge source_lines: 67764-67774 line_count: 10 --- # PostgresMemoryDb Source: https://docs.agno.com/reference/memory/storage/postgres PostgresMemoryDb is a class that implements the MemoryDb interface using PostgreSQL as the backend storage system. It provides persistent storage for agent memories with support for JSONB data types, timestamps, and efficient querying. [V0_FILE]markdown:file="agno/knowledge/postgresql_memory_storage.md" --- title: PostgreSQL Memory Storage category: knowledge source_lines: 18002-18079 line_count: 77 --- # PostgreSQL Memory Storage Source: https://docs.agno.com/examples/concepts/memory/db/mem-postgres-memory ## Code \`\`\`python cookbook/agent_concepts/memory/postgres_memory.py """ This example shows how to use the Memory class with PostgreSQL storage. """ from agno.agent.agent import Agent from agno.memory.v2.db.postgres import PostgresMemoryDb from agno.memory.v2.memory import Memory from agno.models.openai.chat import OpenAIChat from agno.storage.postgres import PostgresStorage db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai" memory = Memory(db=PostgresMemoryDb(table_name="agent_memories", db_url=db_url)) session_id = "postgres_memories" user_id = "postgres_user" agent = Agent( model=OpenAIChat(id="gpt-4o-mini"), memory=memory, storage=PostgresStorage(table_name="agent_sessions", db_url=db_url), enable_user_memories=True, enable_session_summaries=True, ) agent.print_response( "My name is John Doe and I like to hike in the mountains on weekends.", stream=True, user_id=user_id, session_id=session_id, ) agent.print_response( "What are my hobbies?", stream=True, user_id=user_id, session_id=session_id ) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U agno openai sqlalchemy 'psycopg[binary]' \`\`\` \`\`\`bash Mac/Linux python cookbook/agent_concepts/memory/postgres_memory.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/memory/postgres_memory.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/qdrant_agent_knowledge.md" --- title: Qdrant Agent Knowledge category: knowledge source_lines: 81131-81252 line_count: 121 --- # Qdrant Agent Knowledge Source: https://docs.agno.com/vectordb/qdrant ## Setup Follow the instructions in the [Qdrant Setup Guide](https://qdrant.tech/documentation/guides/installation/) to install Qdrant locally. Here is a guide to get API keys: [Qdrant API Keys](https://qdrant.tech/documentation/cloud/authentication/). ## Example \`\`\`python agent_with_knowledge.py import os import typer from typing import Optional from rich.prompt import Prompt from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.qdrant import Qdrant api_key = os.getenv("QDRANT_API_KEY") qdrant_url = os.getenv("QDRANT_URL") collection_name = "thai-recipe-index" vector_db = Qdrant( collection=collection_name, url=qdrant_url, api_key=api_key, ) knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) def qdrant_agent(user: str = "user"): run_id: Optional[str] = None agent = Agent( run_id=run_id, user_id=user, knowledge=knowledge_base, tool_calls=True, use_tools=True, show_tool_calls=True, debug_mode=True, ) if run_id is None: run_id = agent.run_id print(f"Started Run: {run_id}\n") else: print(f"Continuing Run: {run_id}\n") while True: message = Prompt.ask(f"[bold] :sunglasses: {user} [/bold]") if message in ("exit", "bye"): break agent.print_response(message) if __name__ == "__main__": # Comment out after first run knowledge_base.load(recreate=True, upsert=True) typer.run(qdrant_agent) \`\`\`

Qdrant also supports asynchronous operations, enabling concurrency and leading to better performance.

\`\`\`python async_qdrant_db.py import asyncio from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.qdrant import Qdrant COLLECTION_NAME = "thai-recipes" # Initialize Qdrant with local instance vector_db = Qdrant( collection=COLLECTION_NAME, url="http://localhost:6333" ) # Create knowledge base knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=vector_db, ) agent = Agent(knowledge=knowledge_base, show_tool_calls=True) if __name__ == "__main__": # Load knowledge base asynchronously asyncio.run(knowledge_base.aload(recreate=False)) # Comment out after first run # Create and use the agent asynchronously asyncio.run(agent.aprint_response("How to make Tom Kha Gai", markdown=True)) \`\`\` Using aload() and aprint\_response() with asyncio provides non-blocking operations, making your application more responsive under load.

## Qdrant Params ## Developer Resources * View [Cookbook (Sync)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/qdrant_db/qdrant_db.py) * View [Cookbook (Async)](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/qdrant_db/async_qdrant_db.py) [V0_FILE]markdown:file="agno/knowledge/reasoning_agent_with_knowledge_tools.md" --- title: Reasoning Agent with Knowledge Tools category: knowledge source_lines: 22970-22986 line_count: 16 --- # Reasoning Agent with Knowledge Tools Source: https://docs.agno.com/examples/concepts/reasoning/tools/knowledge-tools ## Code \`\`\`python cookbook/reasoning/tools/knowledge_tools.py from agno.agent import Agent from agno.embedder.openai import OpenAIEmbedder from agno.knowledge.url import UrlKnowledge from agno.models.openai import OpenAIChat from agno.tools.knowledge import KnowledgeTools from agno.vectordb.lancedb import LanceDb, SearchType [V0_FILE]markdown:file="agno/knowledge/recursive_chunking.md" --- title: Recursive Chunking category: knowledge source_lines: 67307-67318 line_count: 11 --- # Recursive Chunking Source: https://docs.agno.com/reference/chunking/recursive Recursive chunking is a method of splitting documents into smaller chunks by recursively applying a chunking strategy. This is useful when you want to process large documents in smaller, manageable pieces. [V0_FILE]markdown:file="agno/knowledge/redis_memory_storage.md" --- title: Redis Memory Storage category: knowledge source_lines: 18079-18097 line_count: 18 --- # Redis Memory Storage Source: https://docs.agno.com/examples/concepts/memory/db/mem-redis-memory ## Code \`\`\`python cookbook/agent_concepts/memory/redis_memory.py """ This example shows how to use the Memory class with Redis storage. """ from agno.agent.agent import Agent from agno.memory.v2.db.redis import RedisMemoryDb from agno.memory.v2.memory import Memory from agno.models.openai import OpenAIChat from agno.storage.redis import RedisStorage [V0_FILE]markdown:file="agno/knowledge/redismemorydb.md" --- title: RedisMemoryDb category: knowledge source_lines: 67774-67784 line_count: 10 --- # RedisMemoryDb Source: https://docs.agno.com/reference/memory/storage/redis RedisMemoryDb is a class that implements the MemoryDb interface using Redis as the backend storage system. It provides persistent storage for agent memories with support for JSONB data types, timestamps, and efficient querying. [V0_FILE]markdown:file="agno/knowledge/replace_a_memory.md" --- title: Replace a memory category: knowledge source_lines: 16601-16639 line_count: 38 --- # Replace a memory print("\nReplacing memory") memory.replace_user_memory( memory_id=memory_id_1, memory=UserMemory(memory="The user's name is Jane Mary Doe", topics=["name"]), user_id=jane_doe_id, ) print("Memory replaced") memories = memory.get_user_memories(user_id=jane_doe_id) print("Memories:") pprint(memories) \`\`\` ## Usage \`\`\`bash pip install -U agno rich \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/memory/01_standalone_memory.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/memory/01s̄_standalone_memory.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/replace_a_user_memory.md" --- title: Replace a user memory category: knowledge source_lines: 63070-63090 line_count: 20 --- # Replace a user memory memory_id = memory.replace_user_memory( # The id of the memory to replace memory_id=previous_memory_id, # The new memory to replace it with memory=UserMemory( memory="The user's name is Verna Doe", topics=["personal", "name"] ), user_id="jane_doe@example.com" ) \`\`\` ### Deleting a memory \`\`\`python from agno.memory.v2.memory import Memory memory = Memory() [V0_FILE]markdown:file="agno/knowledge/replace_update_a_memory.md" --- title: Replace/update a memory category: knowledge source_lines: 63393-63400 line_count: 7 --- # Replace/update a memory memory.replace_user_memory( memory_id="memory_123", memory=UserMemory(memory="Updated information about the user"), user_id="user@example.com" ) [V0_FILE]markdown:file="agno/knowledge/reset_the_memory_for_this_example.md" --- title: Reset the memory for this example category: knowledge source_lines: 62988-62991 line_count: 3 --- # Reset the memory for this example memory.clear() [V0_FILE]markdown:file="agno/knowledge/s3_pdf_knowledge_base.md" --- title: S3 PDF Knowledge Base category: knowledge source_lines: 62231-62287 line_count: 56 --- # S3 PDF Knowledge Base Source: https://docs.agno.com/knowledge/s3_pdf Learn how to use PDFs from an S3 bucket in your knowledge base. The **S3PDFKnowledgeBase** reads **PDF** files from an S3 bucket, converts them into vector embeddings and loads them to a vector database. ## Usage We are using a local PgVector database for this example. [Make sure it's running](https://docs.agno.com/vectordb/pgvector) \`\`\`python from agno.knowledge.s3.pdf import S3PDFKnowledgeBase from agno.vectordb.pgvector import PgVector db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai" knowledge_base = S3PDFKnowledgeBase( bucket_name="agno-public", key="recipes/ThaiRecipes.pdf", vector_db=PgVector(table_name="recipes", db_url=db_url), ) \`\`\` Then use the `knowledge_base` with an `Agent`: \`\`\`python from agno.agent import Agent from knowledge_base import knowledge_base agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) agent.knowledge.load(recreate=False) agent.print_response("How to make Thai curry?") \`\`\` ## Params | Parameter | Type | Default | Description | | ------------- | ------------- | --------------- | ---------------------------------------------------------------------------------- | | `bucket_name` | `str` | `None` | The name of the S3 Bucket where the PDFs are. | | `key` | `str` | `None` | The key of the PDF file in the bucket. | | `reader` | `S3PDFReader` | `S3PDFReader()` | A `S3PDFReader` that converts the `PDFs` into `Documents` for the vector database. | `S3PDFKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/s3_pdf_kb.py) [V0_FILE]markdown:file="agno/knowledge/s3_text_knowledge_base.md" --- title: S3 Text Knowledge Base category: knowledge source_lines: 62287-62348 line_count: 61 --- # S3 Text Knowledge Base Source: https://docs.agno.com/knowledge/s3_text Learn how to use text files from an S3 bucket in your knowledge base. The **S3TextKnowledgeBase** reads **text** files from an S3 bucket, converts them into vector embeddings and loads them to a vector database. ## Usage We are using a local PgVector database for this example. [Make sure it's running](https://docs.agno.com/vectordb/pgvector) \`\`\`shell pip install textract \`\`\` \`\`\`python from agno.knowledge.s3.text import S3TextKnowledgeBase from agno.vectordb.pgvector import PgVector db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai" knowledge_base = S3TextKnowledgeBase( bucket_name="agno-public", key="recipes/recipes.docx", vector_db=PgVector(table_name="recipes", db_url=db_url), ) \`\`\` Then use the `knowledge_base` with an `Agent`: \`\`\`python from agno.agent import Agent from knowledge_base import knowledge_base agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) agent.knowledge.load(recreate=False) agent.print_response("How to make Hummus?") \`\`\` ## Params | Parameter | Type | Default | Description | | ------------- | -------------- | ------------------- | ----------------------------------------------------------------------------------------- | | `bucket_name` | `str` | `None` | The name of the S3 Bucket where the files are. | | `key` | `str` | `None` | The key of the file in the bucket. | | `formats` | `List[str]` | `[".doc", ".docx"]` | Formats accepted by this knowledge base. | | `reader` | `S3TextReader` | `S3TextReader()` | A `S3TextReader` that converts the `Text` files into `Documents` for the vector database. | `S3TextKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/s3_text_kb.py) [V0_FILE]markdown:file="agno/knowledge/search_knowledge_false.md" --- title: search_knowledge=False, category: knowledge source_lines: 13922-13923 line_count: 1 --- # search_knowledge=False, [V0_FILE]markdown:file="agno/knowledge/search_knowledge_true.md" --- title: search_knowledge=True, category: knowledge source_lines: 16457-16458 line_count: 1 --- # search_knowledge=True, [V0_FILE]markdown:file="agno/knowledge/second_interaction_-_testing_if_memory_was_stored.md" --- title: Second interaction - testing if memory was stored category: knowledge source_lines: 18237-18245 line_count: 8 --- # Second interaction - testing if memory was stored agent.print_response( "What are my hobbies?", stream=True, user_id=user_id, session_id=session_id ) [V0_FILE]markdown:file="agno/knowledge/semantic_chunking.md" --- title: Semantic Chunking category: knowledge source_lines: 67318-67330 line_count: 12 --- # Semantic Chunking Source: https://docs.agno.com/reference/chunking/semantic Semantic chunking is a method of splitting documents into smaller chunks by analyzing semantic similarity between text segments using embeddings. It uses the chonkie library to identify natural breakpoints where the semantic meaning changes significantly, based on a configurable similarity threshold. This helps preserve context and meaning better than fixed-size chunking by ensuring semantically related content stays together in the same chunk, while splitting occurs at meaningful topic transitions. [V0_FILE]markdown:file="agno/knowledge/send_a_message_to_the_agent_that_checks_the_memory_is_working.md" --- title: Send a message to the agent that checks the memory is working category: knowledge source_lines: 63007-63010 line_count: 3 --- # Send a message to the agent that checks the memory is working agent.print_response("What are my hobbies?", stream=True, user_id=john_doe_id) [V0_FILE]markdown:file="agno/knowledge/send_a_message_to_the_agent_that_would_require_the_memory_to_be_used.md" --- title: Send a message to the agent that would require the memory to be used category: knowledge source_lines: 63000-63007 line_count: 7 --- # Send a message to the agent that would require the memory to be used agent.print_response( "My name is John Doe and I like to hike in the mountains on weekends.", stream=True, user_id=john_doe_id, ) [V0_FILE]markdown:file="agno/knowledge/singlestore_agent_knowledge.md" --- title: SingleStore Agent Knowledge category: knowledge source_lines: 81252-81356 line_count: 104 --- # SingleStore Agent Knowledge Source: https://docs.agno.com/vectordb/singlestore ## Setup \`\`\`shell docker run -d --name singlestoredb \ -p 3306:3306 \ -p 8080:8080 \ -e ROOT_PASSWORD=admin \ -e SINGLESTORE_DB=AGNO \ -e SINGLESTORE_USER=root \ -e SINGLESTORE_PASSWORD=password \ singlestore/cluster-in-a-box docker start singlestoredb \`\`\` After running the container, set the environment variables: \`\`\`shell export SINGLESTORE_HOST="localhost" export SINGLESTORE_PORT="3306" export SINGLESTORE_USERNAME="root" export SINGLESTORE_PASSWORD="admin" export SINGLESTORE_DATABASE="AGNO" \`\`\` SingleStore supports both cloud-based and local deployments. For step-by-step guidance on setting up your cloud deployment, please refer to the [SingleStore Setup Guide](https://docs.singlestore.com/cloud/connect-to-singlestore/connect-with-mysql/connect-with-mysql-client/connect-to-singlestore-helios-using-tls-ssl/). ## Example \`\`\`python agent_with_knowledge.py import typer from typing import Optional from os import getenv from sqlalchemy.engine import create_engine from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.singlestore import SingleStore USERNAME = getenv("SINGLESTORE_USERNAME") PASSWORD = getenv("SINGLESTORE_PASSWORD") HOST = getenv("SINGLESTORE_HOST") PORT = getenv("SINGLESTORE_PORT") DATABASE = getenv("SINGLESTORE_DATABASE") SSL_CERT = getenv("SINGLESTORE_SSL_CERT", None) db_url = f"mysql+pymysql://{USERNAME}:{PASSWORD}@{HOST}:{PORT}/{DATABASE}?charset=utf8mb4" if SSL_CERT: db_url += f"&ssl_ca={SSL_CERT}&ssl_verify_cert=true" db_engine = create_engine(db_url) knowledge_base = PDFUrlKnowledgeBase( urls=["https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"], vector_db=SingleStore( collection="recipes", db_engine=db_engine, schema=DATABASE, ), ) def pdf_assistant(user: str = "user"): run_id: Optional[str] = None agent = Agent( run_id=run_id, user_id=user, knowledge_base=knowledge_base, use_tools=True, show_tool_calls=True, # Uncomment the following line to use traditional RAG # add_references_to_prompt=True, ) if run_id is None: run_id = agent.run_id print(f"Started Run: {run_id}\n") else: print(f"Continuing Run: {run_id}\n") while True: agent.cli_app(markdown=True) if __name__ == "__main__": # Comment out after first run knowledge_base.load(recreate=False) typer.run(pdf_assistant) \`\`\` ## SingleStore Params ## Developer Resources * View [Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/vector_dbs/singlestore_db.py) [V0_FILE]markdown:file="agno/knowledge/sqlite_memory_storage.md" --- title: SQLite Memory Storage category: knowledge source_lines: 18183-18201 line_count: 18 --- # SQLite Memory Storage Source: https://docs.agno.com/examples/concepts/memory/db/mem-sqlite-memory ## Code \`\`\`python cookbook/agent_concepts/memory/sqlite_memory.py """ This example shows how to use the Memory class with SQLite storage. """ from agno.agent.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory from agno.models.openai import OpenAIChat from agno.storage.sqlite import SqliteStorage [V0_FILE]markdown:file="agno/knowledge/sqlitememorydb.md" --- title: SqliteMemoryDb category: knowledge source_lines: 67784-67794 line_count: 10 --- # SqliteMemoryDb Source: https://docs.agno.com/reference/memory/storage/sqlite SqliteMemoryDb is a class that implements the MemoryDb interface using SQLite as the backend storage system. It provides lightweight, file-based or in-memory storage for agent memories with automatic timestamp management. [V0_FILE]markdown:file="agno/knowledge/standalone_memory_operations.md" --- title: Standalone Memory Operations category: knowledge source_lines: 16556-16571 line_count: 15 --- # Standalone Memory Operations Source: https://docs.agno.com/examples/concepts/memory/01-standalone-memory This example shows how to manually add, retrieve, delete, and replace user memories. ## Code \`\`\`python cookbook/agent_concepts/memory/01_standalone_memory.py from agno.memory.v2 import Memory, UserMemory from rich.pretty import pprint memory = Memory() [V0_FILE]markdown:file="agno/knowledge/step_1_initialize_knowledge_base_with_documents_and_metadata.md" --- title: Step 1: Initialize knowledge base with documents and metadata category: knowledge source_lines: 16398-16399 line_count: 1 --- # Step 1: Initialize knowledge base with documents and metadata [V0_FILE]markdown:file="agno/knowledge/step_1_initialize_knowledge_base_with_urls_and_metadata.md" --- title: Step 1: Initialize knowledge base with URLs and metadata category: knowledge source_lines: 15918-15919 line_count: 1 --- # Step 1: Initialize knowledge base with URLs and metadata [V0_FILE]markdown:file="agno/knowledge/step_2_query_the_knowledge_base_with_agent_using_filters_from_query_automatically.md" --- title: Step 2: Query the knowledge base with Agent using filters from query automatically category: knowledge source_lines: 16177-16178 line_count: 1 --- # Step 2: Query the knowledge base with Agent using filters from query automatically [V0_FILE]markdown:file="agno/knowledge/step_2_query_the_knowledge_base_with_different_filter_combinations.md" --- title: Step 2: Query the knowledge base with different filter combinations category: knowledge source_lines: 16439-16440 line_count: 1 --- # Step 2: Query the knowledge base with different filter combinations [V0_FILE]markdown:file="agno/knowledge/surrealdb_agent_knowledge.md" --- title: SurrealDB Agent Knowledge category: knowledge source_lines: 81356-81388 line_count: 32 --- # SurrealDB Agent Knowledge Source: https://docs.agno.com/vectordb/surrealdb ## Setup \`\`\`shell docker run --rm \ --pull always \ -p 8000:8000 \ surrealdb/surrealdb:latest \ start \ --user root \ --pass root \`\`\` or \`\`\`shell ./cookbook/scripts/run_surrealdb.sh \`\`\` ## Example \`\`\`python agent_with_knowledge.py from agno.agent import Agent from agno.embedder.openai import OpenAIEmbedder from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.surrealdb import SurrealDb from surrealdb import Surreal [V0_FILE]markdown:file="agno/knowledge/team_with_knowledge_tools.md" --- title: Team with Knowledge Tools category: knowledge source_lines: 22569-22699 line_count: 130 --- # Team with Knowledge Tools Source: https://docs.agno.com/examples/concepts/reasoning/teams/knowledge-tool-team This is a team reasoning example with knowledge tools. Enabling the reasoning option on the team leader helps optimize delegation and enhances multi-agent collaboration by selectively invoking deeper reasoning when required. ## Code \`\`\`python cookbook/reasoning/teams/knowledge_tool_team.py from textwrap import dedent from agno.agent import Agent from agno.knowledge.url import UrlKnowledge from agno.models.openai import OpenAIChat from agno.team.team import Team from agno.tools.duckduckgo import DuckDuckGoTools from agno.tools.knowledge import KnowledgeTools from agno.tools.yfinance import YFinanceTools from agno.vectordb.lancedb import LanceDb, SearchType agno_docs = UrlKnowledge( urls=["https://www.paulgraham.com/read.html"], # Use LanceDB as the vector database and store embeddings in the `agno_docs` table vector_db=LanceDb( uri="tmp/lancedb", table_name="agno_docs", search_type=SearchType.hybrid, ), ) knowledge_tools = KnowledgeTools( knowledge=agno_docs, think=True, search=True, analyze=True, add_few_shot=True, ) web_agent = Agent( name="Web Search Agent", role="Handle web search requests", model=OpenAIChat(id="gpt-4o-mini"), tools=[DuckDuckGoTools()], instructions="Always include sources", add_datetime_to_instructions=True, ) finance_agent = Agent( name="Finance Agent", role="Handle financial data requests", model=OpenAIChat(id="gpt-4o-mini"), tools=[ YFinanceTools(stock_price=True, analyst_recommendations=True, company_info=True) ], add_datetime_to_instructions=True, ) team_leader = Team( name="Reasoning Finance Team", mode="coordinate", model=OpenAIChat(id="gpt-4o"), members=[ web_agent, finance_agent, ], tools=[knowledge_tools], instructions=[ "Only output the final answer, no other text.", "Use tables to display data", ], markdown=True, show_members_responses=True, enable_agentic_context=True, add_datetime_to_instructions=True, success_criteria="The team has successfully completed the task.", debug_mode=True, ) def run_team(task: str): # Comment out after first run agno_docs.load(recreate=True) team_leader.print_response( task, stream=True, stream_intermediate_steps=True, show_full_reasoning=True, ) if __name__ == "__main__": run_team("What does Paul Graham talk about the need to read in this essay?") \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno \`\`\` \`\`\`bash Mac python cookbook/reasoning/teams/knowledge_tool_team.py \`\`\` \`\`\`bash Windows python cookbook/reasoning/teams/knowledge_tool_team.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/text_knowledge_base.md" --- title: Text Knowledge Base category: knowledge source_lines: 67704-67714 line_count: 10 --- # Text Knowledge Base Source: https://docs.agno.com/reference/knowledge/text TextKnowledge is a knowledge base class that allows you to load and query data from text files. [V0_FILE]markdown:file="agno/knowledge/this_section_loads_the_knowledge_base_skip_if_your_knowledge_base_was_populated_elsewhere.md" --- title: This section loads the knowledge base. Skip if your knowledge base was populated elsewhere. category: knowledge source_lines: 61015-61016 line_count: 1 --- # This section loads the knowledge base. Skip if your knowledge base was populated elsewhere. [V0_FILE]markdown:file="agno/knowledge/use_an_embedder_in_a_knowledge_base.md" --- title: Use an embedder in a knowledge base category: knowledge source_lines: 58887-58901 line_count: 14 --- # Use an embedder in a knowledge base knowledge_base = AgentKnowledge( vector_db=PgVector( db_url="postgresql+psycopg://ai:ai@localhost:5532/ai", table_name="gemini_embeddings", embedder=GeminiEmbedder(), ), num_documents=2, ) \`\`\` For more details on configuring different model providers, check our [Embeddings documentation](../embedder/) [V0_FILE]markdown:file="agno/knowledge/use_default_memory_manager.md" --- title: Use default memory manager category: knowledge source_lines: 16764-16768 line_count: 4 --- # Use default memory manager memory = Memory(model=Claude(id="claude-3-5-sonnet-latest"), db=memory_db) jane_doe_id = "jane_doe@example.com" [V0_FILE]markdown:file="agno/knowledge/user_id_for_the_memory.md" --- title: User ID for the memory category: knowledge source_lines: 62991-63000 line_count: 9 --- # User ID for the memory john_doe_id = "john_doe@example.com" agent = Agent( model=OpenAIChat(id="gpt-4o-mini"), memory=memory, # This will trigger the MemoryManager after each user message enable_user_memories=True, ) [V0_FILE]markdown:file="agno/knowledge/we_can_get_the_session_summary_from_memory_as_well.md" --- title: We can get the session summary from memory as well category: knowledge source_lines: 17209-17247 line_count: 38 --- # We can get the session summary from memory as well session_summary = memory.get_session_summary( session_id=session_id_2, user_id=mark_gonzales_id ) pprint(session_summary) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U agno openai \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/memory/08_agent_with_summaries.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/memory/08_agent_with_summaries.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/weaviate_agent_knowledge.md" --- title: Weaviate Agent Knowledge category: knowledge source_lines: 81511-81557 line_count: 46 --- # Weaviate Agent Knowledge Source: https://docs.agno.com/vectordb/weaviate Follow steps mentioned in [Weaviate setup guide](https://weaviate.io/developers/weaviate/quickstart) to setup Weaviate. ## Setup Install weaviate packages \`\`\`shell pip install weaviate-client \`\`\` Run weaviate \`\`\`shell docker run -d \ -p 8080:8080 \ -p 50051:50051 \ --name weaviate \ cr.weaviate.io/semitechnologies/weaviate:1.28.4 \`\`\` or \`\`\`shell ./cookbook/scripts/run_weaviate.sh \`\`\` ## Example \`\`\`python agent_with_knowledge.py from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.search import SearchType from agno.vectordb.weaviate import Distance, VectorIndex, Weaviate vector_db = Weaviate( collection="recipes", search_type=SearchType.hybrid, vector_index=VectorIndex.HNSW, distance=Distance.COSINE, local=True, # Set to False if using Weaviate Cloud and True if using local instance ) [V0_FILE]markdown:file="agno/knowledge/website_knowledge_base.md" --- title: Website Knowledge Base category: knowledge source_lines: 67714-67724 line_count: 10 --- # Website Knowledge Base Source: https://docs.agno.com/reference/knowledge/website WebsiteKnowledge is a knowledge base class that allows you to load and query data from websites. [V0_FILE]markdown:file="agno/knowledge/what_is_knowledge.md" --- title: What is Knowledge? category: knowledge source_lines: 61419-61441 line_count: 22 --- # What is Knowledge? Source: https://docs.agno.com/knowledge/introduction Knowledge is domain-specific information that the Agent can search at runtime to make better decisions (dynamic few-shot learning) and provide accurate responses (agentic RAG). Knowledge is stored in a vector db and this searching on demand pattern is called Agentic RAG. Example: If we're building a Text2Sql Agent, we'll need to give the table schemas, column names, data types, example queries, common "gotchas" to help it generate the best-possible SQL query. We're obviously not going to put this all in the system prompt, instead we store this information in a vector database and let the Agent query it at runtime. Using this information, the Agent can then generate the best-possible SQL query. This is called dynamic few-shot learning. **Agno Agents use Agentic RAG** by default, meaning if you add `knowledge` to an Agent, it will search this knowledge base, at runtime, for the specific information it needs to achieve its task. The pseudo steps for adding knowledge to an Agent are: \`\`\`python from agno.agent import Agent, AgentKnowledge [V0_FILE]markdown:file="agno/knowledge/what_is_memory.md" --- title: What is Memory? category: knowledge source_lines: 62869-62891 line_count: 22 --- # What is Memory? Source: https://docs.agno.com/memory/introduction Memory gives an Agent the ability to recall relevant information. Memory is a part of the Agent's context that helps it provide the best, most personalized response. If the user tells the Agent they like to ski, then future responses can reference this information to provide a more personalized experience. 1. **Session Storage (chat history and session state):** Session storage saves an Agent's sessions in a database and enables Agents to have multi-turn conversations. Session storage also holds the session state, which is persisted across runs because it is saved to the database after each run. Session storage is a form of short-term memory **called "Storage" in Agno**. 2. **User Memories (user preferences):** The Agent can store insights and facts about the user that it learns through conversation. This helps the agents personalize its response to the user it is interacting with. Think of this as adding "ChatGPT like memory" to your agent. **This is called "Memory" in Agno**. 3. **Session Summaries (chat summary):** The Agent can store a condensed representations of the session, useful when chat histories gets too long. **This is called "Summary" in Agno**. If you haven't, we also recommend reading the Memory section of the [Agents](/agents/memory) to get familiar with the basics. [V0_FILE]markdown:file="agno/knowledge/when_initializing_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md" --- title: When initializing the knowledge base, we can attach metadata that will be used for filtering category: knowledge source_lines: 16264-16265 line_count: 1 --- # When initializing the knowledge base, we can attach metadata that will be used for filtering [V0_FILE]markdown:file="agno/knowledge/when_loading_the_knowledge_base_we_can_attach_metadata_that_will_be_used_for_filtering.md" --- title: When loading the knowledge base, we can attach metadata that will be used for filtering category: knowledge source_lines: 16400-16401 line_count: 1 --- # When loading the knowledge base, we can attach metadata that will be used for filtering [V0_FILE]markdown:file="agno/knowledge/wikipedia_knowledge_base.md" --- title: Wikipedia Knowledge Base category: knowledge source_lines: 67724-67734 line_count: 10 --- # Wikipedia Knowledge Base Source: https://docs.agno.com/reference/knowledge/wikipedia WikipediaKnowledge is a knowledge base class that allows you to load and query data from Wikipedia articles. [V0_FILE]markdown:file="agno/knowledge/wikipedia_knowledgebase.md" --- title: Wikipedia KnowledgeBase category: knowledge source_lines: 62702-62762 line_count: 60 --- # Wikipedia KnowledgeBase Source: https://docs.agno.com/knowledge/wikipedia Learn how to use Wikipedia topics in your knowledge base. The **WikipediaKnowledgeBase** reads wikipedia topics, converts them into vector embeddings and loads them to a vector database. ## Usage We are using a local PgVector database for this example. [Make sure it's running](http://localhost:3333/vectordb/pgvector) \`\`\`shell pip install wikipedia \`\`\` \`\`\`python knowledge_base.py from agno.knowledge.wikipedia import WikipediaKnowledgeBase from agno.vectordb.pgvector import PgVector knowledge_base = WikipediaKnowledgeBase( topics=["Manchester United", "Real Madrid"], # Table name: ai.wikipedia_documents vector_db=PgVector( table_name="wikipedia_documents", db_url="postgresql+psycopg://ai:ai@localhost:5532/ai", ), ) \`\`\` Then use the `knowledge_base` with an Agent: \`\`\`python agent.py from agno.agent import Agent from knowledge_base import knowledge_base agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) agent.knowledge.load(recreate=False) agent.print_response("Ask me about something from the knowledge base") \`\`\` ## Params | Parameter | Type | Default | Description | | --------- | ----------- | ------- | -------------- | | `topics` | `List[str]` | \[] | Topics to read | `WikipediaKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Sync loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/wikipedia_kb.py) * View [Async loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/wikipedia_kb_async.py) [V0_FILE]markdown:file="agno/knowledge/you_can_also_override_the_entire_system_message_for_the_memory_manager.md" --- title: You can also override the entire `system_message` for the memory manager category: knowledge source_lines: 17365-17373 line_count: 8 --- # You can also override the entire `system_message` for the memory manager memory_manager = MemoryManager( model=OpenRouter(id="meta-llama/llama-3.3-70b-instruct"), additional_instructions=""" IMPORTANT: Don't store any memories about the user's name. Just say "The User" instead of referencing the user's name. """, ) [V0_FILE]markdown:file="agno/knowledge/youtube_knowledge_base.md" --- title: YouTube Knowledge Base category: knowledge source_lines: 67734-67744 line_count: 10 --- # YouTube Knowledge Base Source: https://docs.agno.com/reference/knowledge/youtube YouTubeKnowledge is a knowledge base class that allows you to load and query data from YouTube videos. [V0_FILE]markdown:file="agno/knowledge/youtube_knowledgebase.md" --- title: Youtube KnowledgeBase category: knowledge source_lines: 62762-62869 line_count: 107 --- # Youtube KnowledgeBase Source: https://docs.agno.com/knowledge/youtube Learn how to use YouTube video transcripts in your knowledge base. The **YouTubeKnowledgeBase** iterates over a list of YouTube URLs, extracts the video transcripts, converts them into vector embeddings and loads them to a vector database. ## Usage We are using a local PgVector database for this example. [Make sure it's running](http://localhost:3333/vectordb/pgvector) \`\`\`shell pip install bs4 \`\`\` \`\`\`python knowledge_base.py from agno.knowledge.youtube import YouTubeKnowledgeBase from agno.vectordb.pgvector import PgVector knowledge_base = YouTubeKnowledgeBase( urls=["https://www.youtube.com/watch?v=CDC3GOuJyZ0"], # Table name: ai.website_documents vector_db=PgVector( table_name="youtube_documents", db_url="postgresql+psycopg://ai:ai@localhost:5532/ai", ), ) \`\`\` Then use the `knowledge_base` with an `Agent`: \`\`\`python agent.py from agno.agent import Agent from knowledge_base import knowledge_base agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) agent.knowledge.load(recreate=False) agent.print_response("Ask me about something from the knowledge base") \`\`\` #### YouTubeKnowledgeBase also supports async loading. \`\`\`shell pip install qdrant-client \`\`\` We are using a local Qdrant database for this example. [Make sure it's running](https://docs.agno.com/vectordb/qdrant) \`\`\`python async_knowledge_base.py import asyncio from agno.agent import Agent from agno.knowledge.youtube import YouTubeKnowledgeBase, YouTubeReader from agno.vectordb.qdrant import Qdrant COLLECTION_NAME = "youtube-reader" vector_db = Qdrant(collection=COLLECTION_NAME, url="http://localhost:6333") knowledge_base = YouTubeKnowledgeBase( urls=[ "https://www.youtube.com/watch?v=CDC3GOuJyZ0", "https://www.youtube.com/watch?v=JbF_8g1EXj4", ], vector_db=vector_db, reader=YouTubeReader(chunk=True), ) agent = Agent( knowledge=knowledge_base, search_knowledge=True, ) if __name__ == "__main__": # Comment out after first run asyncio.run(knowledge_base.aload(recreate=False)) # Create and use the agent asyncio.run( agent.aprint_response( "What is the major focus of the knowledge provided in both the videos, explain briefly.", markdown=True, ) ) \`\`\` ## Params | Parameter | Type | Default | Description | | --------- | ------------------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------ | | `urls` | `List[str]` | `[]` | URLs of the videos to read | | `reader` | `Optional[YouTubeReader]` | `None` | A `YouTubeReader` that reads transcripts of the videos at the urls and converts them into `Documents` for the vector database. | `YouTubeKnowledgeBase` is a subclass of the [AgentKnowledge](/reference/knowledge/base) class and has access to the same params. ## Developer Resources * View [Sync loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/youtube_kb.py) * View [Async loading Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/agent_concepts/knowledge/youtube_kb_async.py) [V0_FILE]markdown:file="agno/knowledge/zep_async_memory_tools.md" --- title: Zep Async Memory Tools category: knowledge source_lines: 26454-26536 line_count: 82 --- # Zep Async Memory Tools Source: https://docs.agno.com/examples/concepts/tools/database/zep_async ## Code \`\`\`python cookbook/tools/zep_async_tools.py import asyncio import time from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.zep import ZepAsyncTools async def main(): # Initialize the ZepAsyncTools zep_tools = ZepAsyncTools( user_id="agno", session_id="agno-async-session", add_instructions=True ) # Initialize the Agent agent = Agent( model=OpenAIChat(), tools=[zep_tools], context={ "memory": lambda: zep_tools.get_zep_memory(memory_type="context"), }, add_context=True, ) # Interact with the Agent await agent.aprint_response("My name is John Billings") await agent.aprint_response("I live in NYC") await agent.aprint_response("I'm going to a concert tomorrow") # Allow the memories to sync with Zep database time.sleep(10) # Refresh the context agent.context["memory"] = await zep_tools.get_zep_memory(memory_type="context") # Ask the Agent about the user await agent.aprint_response("What do you know about me?") if __name__ == "__main__": asyncio.run(main()) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx export ZEP_API_KEY=xxx \`\`\` \`\`\`bash pip install -U zep-cloud openai agno \`\`\` \`\`\`bash Mac python cookbook/tools/zep_async_tools.py \`\`\` \`\`\`bash Windows python cookbook/tools/zep_async_tools.py \`\`\` [V0_FILE]markdown:file="agno/knowledge/zep_memory_tools.md" --- title: Zep Memory Tools category: knowledge source_lines: 26383-26396 line_count: 13 --- # Zep Memory Tools Source: https://docs.agno.com/examples/concepts/tools/database/zep ## Code \`\`\`python cookbook/tools/zep_tools.py import time from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.zep import ZepTools [V0_FILE]markdown:file="agno/metrics/access_metrics_for_a_specific_member.md" --- title: Access metrics for a specific member category: metrics source_lines: 70102-70131 line_count: 29 --- # Access metrics for a specific member for member_response in team.run_response.member_responses: print(f"Member: {member_response.member_id}") print(f"Member metrics: {member_response.metrics}") # Access individual messages for message in member_response.messages: if message.role == "assistant": print(f"Message metrics: {message.metrics}") \`\`\` ## Metrics Comparison | Metric Level | Access Method | Description | | ----------------------- | -------------------------------- | ---------------------------------------------- | | **Team Leader Run** | `team.run_response.metrics` | Aggregated metrics for the current run | | **Team Leader Session** | `team.session_metrics` | Aggregated metrics across all team leader runs | | **Individual Member** | `member_response.metrics` | Metrics for a specific team member's run | | **Full Team Session** | `team.full_team_session_metrics` | Complete team metrics including all members | ## `MessageMetrics` Params ## `SessionMetrics` Params [V0_FILE]markdown:file="agno/metrics/agent_extra_metrics.md" --- title: Agent Extra Metrics category: metrics source_lines: 20102-20117 line_count: 15 --- # Agent Extra Metrics Source: https://docs.agno.com/examples/concepts/others/agent_extra_metrics This example shows how to get special token metrics like audio, cached and reasoning tokens. ## Code \`\`\`python cookbook/agent_concepts/other/agent_extra_metrics.py import requests from agno.agent import Agent from agno.media import Audio from agno.models.openai import OpenAIChat [V0_FILE]markdown:file="agno/metrics/agent_metrics.md" --- title: Agent Metrics category: metrics source_lines: 20196-20225 line_count: 29 --- # Agent Metrics Source: https://docs.agno.com/examples/concepts/others/agent_metrics This example shows how to get the metrics of an agent run. ## Code \`\`\`python cookbook/agent_concepts/other/agent_metrics.py from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.openai import OpenAIChat from agno.tools.yfinance import YFinanceTools from agno.utils.pprint import pprint_run_response from rich.pretty import pprint agent = Agent( model=OpenAIChat(id="gpt-4o"), tools=[YFinanceTools(stock_price=True)], markdown=True, show_tool_calls=True, ) run_stream: Iterator[RunResponse] = agent.run( "What is the stock price of NVDA", stream=True ) pprint_run_response(run_stream, markdown=True) [V0_FILE]markdown:file="agno/metrics/metrics.md" --- title: Metrics category: metrics source_lines: 69930-69972 line_count: 42 --- # Metrics Source: https://docs.agno.com/teams/metrics Understanding team run and session metrics in Agno ## Overview When you run a team in Agno, the response you get (**TeamRunResponse**) includes detailed metrics about the run. These metrics help you understand resource usage (like **token usage** and **time**), performance, and other aspects of the model and tool calls across both the team leader and team members. Metrics are available at multiple levels: * **Per-message**: Each message (assistant, tool, etc.) has its own metrics. * **Per-tool call**: Each tool execution has its own metrics. * **Per-member run**: Each team member run has its own metrics. * **Team-level**: The `TeamRunResponse` aggregates metrics across all team leader messages. * **Session-level**: Aggregated metrics across all runs in the session, for both the team leader and all team members. Where Metrics Live * `TeamRunResponse.metrics`: Aggregated metrics for the team leader's run, as a dictionary. * `TeamRunResponse.member_responses`: Individual member responses with their own metrics. * `ToolExecution.metrics`: Metrics for each tool call. * `Message.metrics`: Metrics for each message (assistant, tool, etc.). * `Team.session_metrics`: Session-level metrics for the team leader. * `Team.full_team_session_metrics`: Session-level metrics including all team member metrics. ## Example Usage Suppose you have a team that performs some tasks and you want to analyze the metrics after running it. Here's how you can access and print the metrics: \`\`\`python from typing import Iterator from agno.agent import Agent, RunResponse from agno.models.openai import OpenAIChat from agno.team.team import Team from agno.tools.yfinance import YFinanceTools from agno.utils.pprint import pprint_run_response from rich.pretty import pprint [V0_FILE]markdown:file="agno/metrics/print_aggregated_team_leader_metrics.md" --- title: Print aggregated team leader metrics category: metrics source_lines: 70007-70011 line_count: 4 --- # Print aggregated team leader metrics print("---" * 5, "Aggregated Metrics of Team Agent", "---" * 5) pprint(team.run_response.metrics) [V0_FILE]markdown:file="agno/metrics/print_full_team_session_metrics_including_all_members.md" --- title: Print full team session metrics (including all members) category: metrics source_lines: 70030-70102 line_count: 72 --- # Print full team session metrics (including all members) print("---" * 5, "Full Team Session Metrics", "---" * 5) pprint(team.full_team_session_metrics) \`\`\` ## Team Leader Metrics ### Team Leader Message Metrics This section provides metrics for each message response from the team leader. All "assistant" responses will have metrics like this, helping you understand the performance and resource usage at the message level. ![Team Leader Message Metrics](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/team-leader-message-metrics.png) ### Aggregated Team Leader Metrics The aggregated metrics provide a comprehensive view of the team leader's run. This includes a summary of all messages and tool calls, giving you an overall picture of the team leader's performance and resource usage. ![Aggregated Team Leader Metrics](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/team-leader-aggregated-metrics.png) ## Team Member Metrics ### Individual Member Metrics Each team member has their own metrics that can be accessed through `team.run_response.member_responses`. This allows you to analyze the performance of individual team members. ![Team Member Message Metrics](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/team-member-message-metrics.png) ### Member Response Structure Each member response contains: * `messages`: List of messages with individual metrics * `metrics`: Aggregated metrics for that member's run * `tools`: Tool executions with their own metrics ## Session-Level Metrics ### Team Leader Session Metrics The `team.session_metrics` provides aggregated metrics across all runs in the session for the team leader only. ![Team Leader Session Metrics](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/team-leader-session-metrics.png) ### Full Team Session Metrics The `team.full_team_session_metrics` provides comprehensive metrics that include both the team leader and all team members across all runs in the session. ![Full Team Session Metrics](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/full-team-session-metrics.png) ## How Metrics Are Aggregated ### Team Leader Level * **Per-message**: Each message (assistant, tool, etc.) has its own metrics object. * **Run-level**: `TeamRunResponse.metrics` is a dictionary where each key (e.g., input\_tokens) maps to a list of values from all assistant messages in the run. * **Session-level**: `team.session_metrics` aggregates metrics across all team leader runs in the session. ### Team Member Level * **Per-member**: Each team member has their own metrics tracked separately. * **Member aggregation**: Individual member metrics are aggregated within their respective `RunResponse` objects. * **Full team aggregation**: `team.full_team_session_metrics` combines metrics from the team leader and all team members. ### Cross-Member Aggregation * **Session-level**: `team.full_team_session_metrics` provides a complete view of all token usage and performance metrics across the entire team. ## Accessing Member Metrics Programmatically You can access individual member metrics in several ways: \`\`\`python [V0_FILE]markdown:file="agno/metrics/print_metrics_per_message.md" --- title: Print metrics per message category: metrics source_lines: 20225-20237 line_count: 12 --- # Print metrics per message if agent.run_response.messages: for message in agent.run_response.messages: if message.role == "assistant": if message.content: print(f"Message: {message.content}") elif message.tool_calls: print(f"Tool calls: {message.tool_calls}") print("---" * 5, "Metrics", "---" * 5) pprint(message.metrics) print("---" * 20) [V0_FILE]markdown:file="agno/metrics/print_team_leader_message_metrics.md" --- title: Print team leader message metrics category: metrics source_lines: 69994-70007 line_count: 13 --- # Print team leader message metrics print("---" * 5, "Team Leader Message Metrics", "---" * 5) if team.run_response.messages: for message in team.run_response.messages: if message.role == "assistant": if message.content: print(f"Message: {message.content}") elif message.tool_calls: print(f"Tool calls: {message.tool_calls}") print("---" * 5, "Metrics", "---" * 5) pprint(message.metrics) print("---" * 20) [V0_FILE]markdown:file="agno/metrics/print_team_leader_session_metrics.md" --- title: Print team leader session metrics category: metrics source_lines: 70011-70015 line_count: 4 --- # Print team leader session metrics print("---" * 5, "Session Metrics", "---" * 5) pprint(team.session_metrics) [V0_FILE]markdown:file="agno/metrics/print_team_member_message_metrics.md" --- title: Print team member message metrics category: metrics source_lines: 70015-70030 line_count: 15 --- # Print team member message metrics print("---" * 5, "Team Member Message Metrics", "---" * 5) if team.run_response.member_responses: for member_response in team.run_response.member_responses: if member_response.messages: for message in member_response.messages: if message.role == "assistant": if message.content: print(f"Message: {message.content}") elif message.tool_calls: print(f"Tool calls: {message.tool_calls}") print("---" * 5, "Metrics", "---" * 5) pprint(message.metrics) print("---" * 20) [V0_FILE]markdown:file="agno/metrics/print_the_aggregated_metrics_for_the_whole_run.md" --- title: Print the aggregated metrics for the whole run category: metrics source_lines: 1159-1162 line_count: 3 --- # Print the aggregated metrics for the whole run print("---" * 5, "Collected Metrics", "---" * 5) pprint(agent.run_response.metrics) [V0_FILE]markdown:file="agno/metrics/print_the_aggregated_metrics_for_the_whole_session.md" --- title: Print the aggregated metrics for the whole session category: metrics source_lines: 1162-1200 line_count: 38 --- # Print the aggregated metrics for the whole session print("---" * 5, "Session Metrics", "---" * 5) pprint(agent.session_metrics) \`\`\` You'd see the outputs with following information: ### Tool Execution Metrics This section provides metrics for each tool execution. It includes details about the resource usage and performance of individual tool calls. ![Tool Run Message Metrics](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/tools-run-message-metrics.png) ### Message Metrics Here, you can see the metrics for each message response from the agent. All "assistant" responses will have metrics like this, helping you understand the performance and resource usage at the message level. ![Agent Run Message Metrics](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/agent-run-message-metrics.png) ### Aggregated Run Metrics The aggregated metrics provide a comprehensive view of the entire run. This includes a summary of all messages and tool calls, giving you an overall picture of the agent's performance and resource usage. ![Aggregated Run Metrics](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/agent-run-aggregated-metrics.png) Similarly for the session metrics, you can see the aggregated metrics across all runs in the session, providing insights into the overall performance and resource usage of the agent across multiple runs. ## How Metrics Are Aggregated * **Per-message**: Each message (assistant, tool, etc.) has its own metrics object. * **Run-level**: RunResponse.metrics is a dictionary where each key (e.g., input\_tokens) maps to a list of values from all assistant messages in the run. * **Session-level**: `SessionMetrics` (see `agent.session_metrics`) aggregates metrics across all runs in the session. ## `MessageMetrics` Params [V0_FILE]markdown:file="agno/metrics/print_the_metrics.md" --- title: Print the metrics category: metrics source_lines: 20237-20240 line_count: 3 --- # Print the metrics print("---" * 5, "Aggregated Metrics", "---" * 5) pprint(agent.run_response.metrics) [V0_FILE]markdown:file="agno/metrics/print_the_session_metrics.md" --- title: Print the session metrics category: metrics source_lines: 20240-20276 line_count: 36 --- # Print the session metrics print("---" * 5, "Session Metrics", "---" * 5) pprint(agent.session_metrics) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno yfinance rich \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/other/agent_metrics.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/other/agent_metrics.py \`\`\` [V0_FILE]markdown:file="agno/metrics/showing_cached_tokens_metrics.md" --- title: Showing cached tokens metrics category: metrics source_lines: 20161-20196 line_count: 35 --- # Showing cached tokens metrics print(f"Cached tokens: {agent.run_response.metrics['cached_tokens']}") \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U requests openai agno \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/other/agent_extra_metrics.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/other/agent_extra_metrics.py \`\`\` [V0_FILE]markdown:file="agno/metrics/showing_input_audio_output_audio_and_total_audio_tokens_metrics.md" --- title: Showing input audio, output audio and total audio tokens metrics category: metrics source_lines: 20136-20152 line_count: 16 --- # Showing input audio, output audio and total audio tokens metrics print(f"Input audio tokens: {agent.run_response.metrics['input_audio_tokens']}") print(f"Output audio tokens: {agent.run_response.metrics['output_audio_tokens']}") print(f"Audio tokens: {agent.run_response.metrics['audio_tokens']}") agent = Agent( model=OpenAIChat(id="o3-mini"), markdown=True, telemetry=False, monitoring=False, debug_mode=True, ) agent.print_response( "Solve the trolley problem. Evaluate multiple ethical frameworks. Include an ASCII diagram of your solution.", stream=False, ) [V0_FILE]markdown:file="agno/metrics/showing_reasoning_tokens_metrics.md" --- title: Showing reasoning tokens metrics category: metrics source_lines: 20152-20161 line_count: 9 --- # Showing reasoning tokens metrics print(f"Reasoning tokens: {agent.run_response.metrics['reasoning_tokens']}") agent = Agent( model=OpenAIChat(id="gpt-4o-mini"), markdown=True, telemetry=False, monitoring=False ) agent.run("Share a 2 sentence horror story" * 150) agent.print_response("Share a 2 sentence horror story" * 150) [V0_FILE]markdown:file="agno/misc/------------------------------------------------------------------------------------------------.md" --- title: ------------------------------------------------------------------------------------------------ # category: misc source_lines: 56936-56938 line_count: 2 --- # ------------------------------------------------------------------------------------------------ # [V0_FILE]markdown:file="agno/misc/-----------------------------------------------------------------------------------.md" --- title: ----------------------------------------------------------------------------------- category: misc source_lines: 16178-16180 line_count: 2 --- # ----------------------------------------------------------------------------------- [V0_FILE]markdown:file="agno/misc/------------------------------------------------------------------------------.md" --- title: ------------------------------------------------------------------------------ category: misc source_lines: 16440-16441 line_count: 1 --- # ------------------------------------------------------------------------------ [V0_FILE]markdown:file="agno/misc/---------------------------------------------------------.md" --- title: --------------------------------------------------------- category: misc source_lines: 61104-61106 line_count: 2 --- # --------------------------------------------------------- [V0_FILE]markdown:file="agno/misc/---_agents_---.md" --- title: --- Agents --- category: misc source_lines: 83500-83629 line_count: 129 --- # --- Agents --- research_agent = Agent( name="Blog Research Agent", model=OpenAIChat(id="gpt-4o-mini"), tools=[GoogleSearchTools()], description=dedent("""\ You are BlogResearch-X, an elite research assistant specializing in discovering high-quality sources for compelling blog content. Your expertise includes: - Finding authoritative and trending sources - Evaluating content credibility and relevance - Identifying diverse perspectives and expert opinions - Discovering unique angles and insights - Ensuring comprehensive topic coverage """), instructions=dedent("""\ 1. Search Strategy 🔍 - Find 10-15 relevant sources and select the 5-7 best ones - Prioritize recent, authoritative content - Look for unique angles and expert insights 2. Source Evaluation 📊 - Verify source credibility and expertise - Check publication dates for timeliness - Assess content depth and uniqueness 3. Diversity of Perspectives 🌐 - Include different viewpoints - Gather both mainstream and expert opinions - Find supporting data and statistics """), response_model=SearchResults, ) content_scraper_agent = Agent( name="Content Scraper Agent", model=OpenAIChat(id="gpt-4o-mini"), tools=[Newspaper4kTools()], description=dedent("""\ You are ContentBot-X, a specialist in extracting and processing digital content for blog creation. Your expertise includes: - Efficient content extraction - Smart formatting and structuring - Key information identification - Quote and statistic preservation - Maintaining source attribution """), instructions=dedent("""\ 1. Content Extraction 📑 - Extract content from the article - Preserve important quotes and statistics - Maintain proper attribution - Handle paywalls gracefully 2. Content Processing 🔄 - Format text in clean markdown - Preserve key information - Structure content logically 3. Quality Control ✅ - Verify content relevance - Ensure accurate extraction - Maintain readability """), response_model=ScrapedArticle, ) blog_writer_agent = Agent( name="Blog Writer Agent", model=OpenAIChat(id="gpt-4o"), description=dedent("""\ You are BlogMaster-X, an elite content creator combining journalistic excellence with digital marketing expertise. Your strengths include: - Crafting viral-worthy headlines - Writing engaging introductions - Structuring content for digital consumption - Incorporating research seamlessly - Optimizing for SEO while maintaining quality - Creating shareable conclusions """), instructions=dedent("""\ 1. Content Strategy 📝 - Craft attention-grabbing headlines - Write compelling introductions - Structure content for engagement - Include relevant subheadings 2. Writing Excellence ✍️ - Balance expertise with accessibility - Use clear, engaging language - Include relevant examples - Incorporate statistics naturally 3. Source Integration 🔍 - Cite sources properly - Include expert quotes - Maintain factual accuracy 4. Digital Optimization 💻 - Structure for scanability - Include shareable takeaways - Optimize for SEO - Add engaging subheadings Format your blog post with this structure: # {Viral-Worthy Headline} ## Introduction {Engaging hook and context} ## {Compelling Section 1} {Key insights and analysis} {Expert quotes and statistics} ## {Engaging Section 2} {Deeper exploration} {Real-world examples} ## {Practical Section 3} {Actionable insights} {Expert recommendations} ## Key Takeaways - {Shareable insight 1} - {Practical takeaway 2} - {Notable finding 3} ## Sources {Properly attributed sources with links} """), markdown=True, ) [V0_FILE]markdown:file="agno/misc/---_execution_function_---.md" --- title: --- Execution function --- category: misc source_lines: 58202-58408 line_count: 206 --- # --- Execution function --- async def startup_validation_execution( workflow: Workflow, execution_input: WorkflowExecutionInput, startup_idea: str, **kwargs: Any, ) -> str: """Execute the complete startup idea validation workflow""" # Get inputs message: str = execution_input.message idea: str = startup_idea if not idea: return "❌ No startup idea provided" print(f"🚀 Starting startup idea validation for: {idea}") print(f"💡 Validation request: {message}") # Phase 1: Idea Clarification print(f"\n🎯 PHASE 1: IDEA CLARIFICATION & REFINEMENT") print("=" * 60) clarification_prompt = f""" {message} Please analyze and refine the following startup idea: STARTUP IDEA: {idea} Evaluate: 1. The originality of this idea compared to existing solutions 2. Define a clear mission statement for this startup 3. Outline specific, measurable objectives Provide insights on how to strengthen and focus the core concept. """ print(f"🔍 Analyzing and refining the startup concept...") try: clarification_result = await idea_clarifier_agent.arun(clarification_prompt) idea_clarification = clarification_result.content print(f"✅ Idea clarification completed") print(f"📝 Mission: {idea_clarification.mission[:100]}...") except Exception as e: return f"❌ Failed to clarify idea: {str(e)}" # Phase 2: Market Research print(f"\n📊 PHASE 2: MARKET RESEARCH & ANALYSIS") print("=" * 60) market_research_prompt = f""" Based on the refined startup idea and clarification below, conduct comprehensive market research: STARTUP IDEA: {idea} ORIGINALITY: {idea_clarification.originality} MISSION: {idea_clarification.mission} OBJECTIVES: {idea_clarification.objectives} Please research and provide: 1. Total Addressable Market (TAM) - overall market size 2. Serviceable Available Market (SAM) - portion you could serve 3. Serviceable Obtainable Market (SOM) - realistic market share 4. Target customer segments with detailed characteristics Use web search to find current market data and trends. """ print(f"📈 Researching market size and customer segments...") try: market_result = await market_research_agent.arun(market_research_prompt) market_research = market_result.content print(f"✅ Market research completed") print(f"🎯 TAM: {market_research.total_addressable_market[:100]}...") except Exception as e: return f"❌ Failed to complete market research: {str(e)}" # Phase 3: Competitor Analysis print(f"\n🏢 PHASE 3: COMPETITIVE LANDSCAPE ANALYSIS") print("=" * 60) competitor_prompt = f""" Based on the startup idea and market research below, analyze the competitive landscape: STARTUP IDEA: {idea} TAM: {market_research.total_addressable_market} SAM: {market_research.serviceable_available_market} SOM: {market_research.serviceable_obtainable_market} TARGET SEGMENTS: {market_research.target_customer_segments} Please research and provide: 1. Identify direct and indirect competitors 2. SWOT analysis for each major competitor 3. Assessment of startup's potential competitive positioning 4. Market gaps and opportunities Use web search to find current competitor information. """ print(f"🔎 Analyzing competitive landscape...") try: competitor_result = await competitor_analysis_agent.arun(competitor_prompt) competitor_analysis = competitor_result.content print(f"✅ Competitor analysis completed") print(f"🏆 Positioning: {competitor_analysis.positioning[:100]}...") except Exception as e: return f"❌ Failed to complete competitor analysis: {str(e)}" # Phase 4: Final Validation Report print(f"\n📋 PHASE 4: COMPREHENSIVE VALIDATION REPORT") print("=" * 60) report_prompt = f""" Synthesize all the research and analysis into a comprehensive startup validation report: STARTUP IDEA: {idea} IDEA CLARIFICATION: - Originality: {idea_clarification.originality} - Mission: {idea_clarification.mission} - Objectives: {idea_clarification.objectives} MARKET RESEARCH: - TAM: {market_research.total_addressable_market} - SAM: {market_research.serviceable_available_market} - SOM: {market_research.serviceable_obtainable_market} - Target Segments: {market_research.target_customer_segments} COMPETITOR ANALYSIS: - Competitors: {competitor_analysis.competitors} - SWOT: {competitor_analysis.swot_analysis} - Positioning: {competitor_analysis.positioning} Create a professional validation report with: 1. Executive summary 2. Idea assessment (strengths/weaknesses) 3. Market opportunity analysis 4. Competitive landscape overview 5. Strategic recommendations 6. Specific next steps for the entrepreneur """ print(f"📝 Generating comprehensive validation report...") try: final_result = await report_agent.arun(report_prompt) validation_report = final_result.content print(f"✅ Validation report completed") except Exception as e: return f"❌ Failed to generate final report: {str(e)}" # Final summary summary = f""" 🎉 STARTUP IDEA VALIDATION COMPLETED! 📊 Validation Summary: • Startup Idea: {idea} • Idea Clarification: ✅ Completed • Market Research: ✅ Completed • Competitor Analysis: ✅ Completed • Final Report: ✅ Generated 📈 Key Market Insights: • TAM: {market_research.total_addressable_market[:150]}... • Target Segments: {market_research.target_customer_segments[:150]}... 🏆 Competitive Positioning: {competitor_analysis.positioning[:200]}... 📋 COMPREHENSIVE VALIDATION REPORT: ## Executive Summary {validation_report.executive_summary} ## Idea Assessment {validation_report.idea_assessment} ## Market Opportunity {validation_report.market_opportunity} ## Competitive Landscape {validation_report.competitive_landscape} ## Strategic Recommendations {validation_report.recommendations} ## Next Steps {validation_report.next_steps} ⚠️ Disclaimer: This validation is for informational purposes only. Conduct additional due diligence before making investment decisions. """ return summary [V0_FILE]markdown:file="agno/misc/---_helper_functions_---.md" --- title: --- Helper Functions --- category: misc source_lines: 83629-83791 line_count: 162 --- # --- Helper Functions --- def get_cached_blog_post(workflow: Workflow, topic: str) -> Optional[str]: """Get cached blog post from workflow session state""" logger.info("Checking if cached blog post exists") return workflow.workflow_session_state.get("blog_posts", {}).get(topic) def cache_blog_post(workflow: Workflow, topic: str, blog_post: str): """Cache blog post in workflow session state""" logger.info(f"Saving blog post for topic: {topic}") if "blog_posts" not in workflow.workflow_session_state: workflow.workflow_session_state["blog_posts"] = {} workflow.workflow_session_state["blog_posts"][topic] = blog_post def get_cached_search_results( workflow: Workflow, topic: str ) -> Optional[SearchResults]: """Get cached search results from workflow session state""" logger.info("Checking if cached search results exist") search_results = workflow.workflow_session_state.get("search_results", {}).get( topic ) if search_results and isinstance(search_results, dict): try: return SearchResults.model_validate(search_results) except Exception as e: logger.warning(f"Could not validate cached search results: {e}") return search_results if isinstance(search_results, SearchResults) else None def cache_search_results(workflow: Workflow, topic: str, search_results: SearchResults): """Cache search results in workflow session state""" logger.info(f"Saving search results for topic: {topic}") if "search_results" not in workflow.workflow_session_state: workflow.workflow_session_state["search_results"] = {} workflow.workflow_session_state["search_results"][topic] = ( search_results.model_dump() ) def get_cached_scraped_articles( workflow: Workflow, topic: str ) -> Optional[Dict[str, ScrapedArticle]]: """Get cached scraped articles from workflow session state""" logger.info("Checking if cached scraped articles exist") scraped_articles = workflow.workflow_session_state.get("scraped_articles", {}).get( topic ) if scraped_articles and isinstance(scraped_articles, dict): try: return { url: ScrapedArticle.model_validate(article) for url, article in scraped_articles.items() } except Exception as e: logger.warning(f"Could not validate cached scraped articles: {e}") return scraped_articles if isinstance(scraped_articles, dict) else None def cache_scraped_articles( workflow: Workflow, topic: str, scraped_articles: Dict[str, ScrapedArticle] ): """Cache scraped articles in workflow session state""" logger.info(f"Saving scraped articles for topic: {topic}") if "scraped_articles" not in workflow.workflow_session_state: workflow.workflow_session_state["scraped_articles"] = {} workflow.workflow_session_state["scraped_articles"][topic] = { url: article.model_dump() for url, article in scraped_articles.items() } async def get_search_results( workflow: Workflow, topic: str, use_cache: bool = True, num_attempts: int = 3 ) -> Optional[SearchResults]: """Get search results with caching support""" # Check cache first if use_cache: cached_results = get_cached_search_results(workflow, topic) if cached_results: logger.info(f"Found {len(cached_results.articles)} articles in cache.") return cached_results # Search for new results for attempt in range(num_attempts): try: print( f"🔍 Searching for articles about: {topic} (attempt {attempt + 1}/{num_attempts})" ) response = await research_agent.arun(topic) if ( response and response.content and isinstance(response.content, SearchResults) ): article_count = len(response.content.articles) logger.info(f"Found {article_count} articles on attempt {attempt + 1}") print(f"✅ Found {article_count} relevant articles") # Cache the results cache_search_results(workflow, topic, response.content) return response.content else: logger.warning( f"Attempt {attempt + 1}/{num_attempts} failed: Invalid response type" ) except Exception as e: logger.warning(f"Attempt {attempt + 1}/{num_attempts} failed: {str(e)}") logger.error(f"Failed to get search results after {num_attempts} attempts") return None async def scrape_articles( workflow: Workflow, topic: str, search_results: SearchResults, use_cache: bool = True, ) -> Dict[str, ScrapedArticle]: """Scrape articles with caching support""" # Check cache first if use_cache: cached_articles = get_cached_scraped_articles(workflow, topic) if cached_articles: logger.info(f"Found {len(cached_articles)} scraped articles in cache.") return cached_articles scraped_articles: Dict[str, ScrapedArticle] = {} print(f"📄 Scraping {len(search_results.articles)} articles...") for i, article in enumerate(search_results.articles, 1): try: print( f"📖 Scraping article {i}/{len(search_results.articles)}: {article.title[:50]}..." ) response = await content_scraper_agent.arun(article.url) if ( response and response.content and isinstance(response.content, ScrapedArticle) ): scraped_articles[response.content.url] = response.content logger.info(f"Scraped article: {response.content.url}") print(f"✅ Successfully scraped: {response.content.title[:50]}...") else: print(f"❌ Failed to scrape: {article.title[:50]}...") except Exception as e: logger.warning(f"Failed to scrape {article.url}: {str(e)}") print(f"❌ Error scraping: {article.title[:50]}...") # Cache the scraped articles cache_scraped_articles(workflow, topic, scraped_articles) return scraped_articles [V0_FILE]markdown:file="agno/misc/---_main_execution_function_---.md" --- title: --- Main Execution Function --- category: misc source_lines: 83791-83881 line_count: 90 --- # --- Main Execution Function --- async def blog_generation_execution( workflow: Workflow, topic: str = None, use_search_cache: bool = True, use_scrape_cache: bool = True, use_blog_cache: bool = True, ) -> str: """ Blog post generation workflow execution function. Args: workflow: The workflow instance execution_input: Standard workflow execution input topic: Blog post topic (if not provided, uses execution_input.message) use_search_cache: Whether to use cached search results use_scrape_cache: Whether to use cached scraped articles use_blog_cache: Whether to use cached blog posts **kwargs: Additional parameters """ blog_topic = topic if not blog_topic: return "❌ No blog topic provided. Please specify a topic." print(f"🎨 Generating blog post about: {blog_topic}") print("=" * 60) # Check for cached blog post first if use_blog_cache: cached_blog = get_cached_blog_post(workflow, blog_topic) if cached_blog: print("📋 Found cached blog post!") return cached_blog # Phase 1: Research and gather sources print(f"\n🔍 PHASE 1: RESEARCH & SOURCE GATHERING") print("=" * 50) search_results = await get_search_results(workflow, blog_topic, use_search_cache) if not search_results or len(search_results.articles) == 0: return f"❌ Sorry, could not find any articles on the topic: {blog_topic}" print(f"📊 Found {len(search_results.articles)} relevant sources:") for i, article in enumerate(search_results.articles, 1): print(f" {i}. {article.title[:60]}...") # Phase 2: Content extraction print(f"\n📄 PHASE 2: CONTENT EXTRACTION") print("=" * 50) scraped_articles = await scrape_articles( workflow, blog_topic, search_results, use_scrape_cache ) if not scraped_articles: return f"❌ Could not extract content from any articles for topic: {blog_topic}" print(f"📖 Successfully extracted content from {len(scraped_articles)} articles") # Phase 3: Blog post writing print(f"\n✍️ PHASE 3: BLOG POST CREATION") print("=" * 50) # Prepare input for the writer writer_input = { "topic": blog_topic, "articles": [article.model_dump() for article in scraped_articles.values()], } print("🤖 AI is crafting your blog post...") writer_response = await blog_writer_agent.arun(json.dumps(writer_input, indent=2)) if not writer_response or not writer_response.content: return f"❌ Failed to generate blog post for topic: {blog_topic}" blog_post = writer_response.content # Cache the blog post cache_blog_post(workflow, blog_topic, blog_post) print("✅ Blog post generated successfully!") print(f"📝 Length: {len(blog_post)} characters") print(f"📚 Sources: {len(scraped_articles)} articles") return blog_post [V0_FILE]markdown:file="agno/misc/---_pdf_utility_---.md" --- title: --- PDF utility --- category: misc source_lines: 57756-57768 line_count: 12 --- # --- PDF utility --- def extract_text_from_pdf(url: str) -> str: try: resp = requests.get(url) resp.raise_for_status() reader = PdfReader(io.BytesIO(resp.content)) return "\n".join(page.extract_text() or "" for page in reader.pages) except Exception as e: print(f"Error extracting PDF from {url}: {e}") return "" [V0_FILE]markdown:file="agno/misc/---_response_models_---.md" --- title: --- Response Models --- category: misc source_lines: 83475-83500 line_count: 25 --- # --- Response Models --- class NewsArticle(BaseModel): title: str = Field(..., description="Title of the article.") url: str = Field(..., description="Link to the article.") summary: Optional[str] = Field( ..., description="Summary of the article if available." ) class SearchResults(BaseModel): articles: list[NewsArticle] class ScrapedArticle(BaseModel): title: str = Field(..., description="Title of the article.") url: str = Field(..., description="Link to the article.") summary: Optional[str] = Field( ..., description="Summary of the article if available." ) content: Optional[str] = Field( ..., description="Full article content in markdown format. None if content is unavailable.", ) [V0_FILE]markdown:file="agno/misc/---_workflow_definition_---.md" --- title: --- Workflow Definition --- category: misc source_lines: 83881-83937 line_count: 56 --- # --- Workflow Definition --- blog_generator_workflow = Workflow( name="Blog Post Generator v2.0", description="Advanced blog post generator with research and content creation capabilities", storage=SqliteStorage( table_name="blog_generator_v2", db_file="tmp/blog_generator_v2.db", mode="workflow_v2", ), steps=blog_generation_execution, workflow_session_state={}, # Initialize empty session state for caching ) if __name__ == "__main__": import random async def main(): # Fun example topics to showcase the generator's versatility example_topics = [ "The Rise of Artificial General Intelligence: Latest Breakthroughs", "How Quantum Computing is Revolutionizing Cybersecurity", "Sustainable Living in 2024: Practical Tips for Reducing Carbon Footprint", "The Future of Work: AI and Human Collaboration", "Space Tourism: From Science Fiction to Reality", "Mindfulness and Mental Health in the Digital Age", "The Evolution of Electric Vehicles: Current State and Future Trends", "Why Cats Secretly Run the Internet", "The Science Behind Why Pizza Tastes Better at 2 AM", "How Rubber Ducks Revolutionized Software Development", ] # Test with a random topic topic = random.choice(example_topics) print("🧪 Testing Blog Post Generator v2.0") print("=" * 60) print(f"📝 Topic: {topic}") print() # Generate the blog post resp = await blog_generator_workflow.arun( topic=topic, use_search_cache=True, use_scrape_cache=True, use_blog_cache=True, ) pprint_run_response(resp, markdown=True, show_time=True) asyncio.run(main()) \`\`\` For more examples and advanced patterns, see [here](/examples/workflows_2). Each file demonstrates a specific pattern with detailed comments and real-world use cases. [V0_FILE]markdown:file="agno/misc/-_-_ask_a_follow_up_question_that_continues_the_conversation.md" --- title: -*- Ask a follow up question that continues the conversation category: misc source_lines: 16520-16522 line_count: 2 --- # -*- Ask a follow up question that continues the conversation agent.print_response("What was my first message?", stream=True) [V0_FILE]markdown:file="agno/misc/-_-_build_container_environment.md" --- title: -*- Build container environment category: misc source_lines: 85675-85750 line_count: 75 --- # -*- Build container environment container_env = { ... # Migrate database on startup using alembic "MIGRATE_DB": ws_settings.prd_db_enabled, } ... \`\`\` ### Update the ECS Task Definition Because we updated the Environment Variables, we need to update the Task Definition: \`\`\`bash terminal ag ws patch --env prd --infra aws --name td \`\`\` \`\`\`bash shorthand ag ws patch -e prd -i aws -n td \`\`\` ### Update the ECS Service After updating the task definition, redeploy the production application: \`\`\`bash terminal ag ws patch --env prd --infra aws --name service \`\`\` \`\`\`bash shorthand ag ws patch -e prd -i aws -n service \`\`\` ## Manually migrate prodution database Another approach is to SSH into the production container to run the migration manually. Your ECS tasks are already enabled with SSH access. Run the alembic command to migrate the production database: \`\`\`bash ECS_CLUSTER=ai-app-prd-cluster TASK_ARN=$(aws ecs list-tasks --cluster ai-app-prd-cluster --query "taskArns[0]" --output text) CONTAINER_NAME=ai-api-prd aws ecs execute-command --cluster $ECS_CLUSTER \ --task $TASK_ARN \ --container $CONTAINER_NAME \ --interactive \ --command "alembic -c db/alembic.ini upgrade head" \`\`\` *** ## How the migrations directory was created These commands have been run and are described for completeness The migrations directory was created using: \`\`\`bash docker exec -it ai-api cd db && alembic init migrations \`\`\` * After running the above command, the `db/migrations` directory should be created. * Update `alembic.ini` * set `script_location = db/migrations` * uncomment `black` hook in `[post_write_hooks]` * Update `db/migrations/env.py` file following [this link](https://alembic.sqlalchemy.org/en/latest/autogenerate.html) * Add the following function to `configure` to only include tables in the target\_metadata \`\`\`python db/migrations/env.py [V0_FILE]markdown:file="agno/misc/-_-_create_a_retriever_from_the_vector_store.md" --- title: -*- Create a retriever from the vector store category: misc source_lines: 61691-61694 line_count: 3 --- # -*- Create a retriever from the vector store retriever = db.as_retriever() [V0_FILE]markdown:file="agno/misc/-_-_create_a_run.md" --- title: -*- Create a run category: misc source_lines: 16510-16512 line_count: 2 --- # -*- Create a run agent.print_response("Share a 2 sentence horror story", stream=True) [V0_FILE]markdown:file="agno/misc/-_-_fastapi_running_on_ecs.md" --- title: -*- FastAPI running on ECS category: misc source_lines: 85920-85963 line_count: 43 --- # -*- FastAPI running on ECS prd_fastapi = FastApi( ... # To enable HTTPS, create an ACM certificate and add the ARN below: load_balancer_enable_https=True, load_balancer_certificate_arn="arn:aws:acm:us-east-1:497891874516:certificate/6598c24a-d4fc-4f17-8ee0-0d3906eb705f", ... ) \`\`\` 4. Create new Loadbalancer Listeners Create new listeners for the loadbalancer to pickup the HTTPs configuration. \`\`\`bash terminal ag ws up --env prd --infra aws --name listener \`\`\` \`\`\`bash shorthand ag ws up -e prd -i aws -n listener \`\`\` The certificate should be `Issued` before applying it. After this, `https` should be working on your custom domain. 5. Update existing listeners to redirect HTTP to HTTPS \`\`\`bash terminal ag ws patch --env prd --infra aws --name listener \`\`\` \`\`\`bash shorthand ag ws patch -e prd -i aws -n listener \`\`\` After this, all HTTP requests should redirect to HTTPS automatically. [V0_FILE]markdown:file="agno/misc/-_-_get_the_vectordb.md" --- title: -*- Get the vectordb category: misc source_lines: 61689-61691 line_count: 2 --- # -*- Get the vectordb db = Chroma(embedding_function=OpenAIEmbeddings(), persist_directory=str(chroma_db_dir)) [V0_FILE]markdown:file="agno/misc/-_-_only_include_tables_that_are_in_the_target_metadata.md" --- title: -*- Only include tables that are in the target_metadata category: misc source_lines: 85750-85760 line_count: 10 --- # -*- Only include tables that are in the target_metadata def include_name(name, type_, parent_names): if type_ == "table": return name in target_metadata.tables else: return True ... \`\`\` [V0_FILE]markdown:file="agno/misc/-_-_print_a_response_to_the_cli.md" --- title: -*- Print a response to the cli category: misc source_lines: 11135-11164 line_count: 29 --- # -*- Print a response to the cli asyncio.run(agent.aprint_response("Share a breakfast recipe.", stream=True)) \`\`\` ## Usage \`\`\`bash pip install -U openai agno \`\`\` \`\`\`bash Mac python cookbook/agent_concepts/async/basic.py \`\`\` \`\`\`bash Windows python cookbook/agent_concepts/async/basic.py \`\`\` [V0_FILE]markdown:file="agno/misc/-_-_print_the_chat_history.md" --- title: -*- Print the chat history category: misc source_lines: 16980-16984 line_count: 4 --- # -*- Print the chat history session_run = memory.runs[session_id][-1] print_chat_history(session_run) [V0_FILE]markdown:file="agno/misc/-_-_secrets_for_production_application.md" --- title: -*- Secrets for production application category: misc source_lines: 86938-86947 line_count: 9 --- # -*- Secrets for production application prd_secret = SecretsManager( ... # Create secret from workspace/secrets/prd_app_secrets.yml secret_files=[ ws_settings.ws_root.joinpath("workspace/secrets/prd_app_secrets.yml") ], ) [V0_FILE]markdown:file="agno/misc/-_-_secrets_for_production_database.md" --- title: -*- Secrets for production database category: misc source_lines: 86947-86978 line_count: 31 --- # -*- Secrets for production database prd_db_secret = SecretsManager( ... # Create secret from workspace/secrets/prd_db_secrets.yml secret_files=[ws_settings.ws_root.joinpath("workspace/secrets/prd_db_secrets.yml")], ) \`\`\` Read the secret in production apps using: \`\`\`python FastApi prd_fastapi = FastApi( ... aws_secrets=[prd_secret], ... ) \`\`\` \`\`\`python RDS prd_db = DbInstance( ... aws_secret=prd_db_secret, ... ) \`\`\` Production resources can also read secrets using yaml files but we highly recommend using [AWS Secrets](https://docs.aws.amazon.com/secretsmanager/latest/userguide/intro.html). [V0_FILE]markdown:file="agno/misc/-_-_streamlit_running_on_ecs.md" --- title: -*- Streamlit running on ECS category: misc source_lines: 85911-85920 line_count: 9 --- # -*- Streamlit running on ECS prd_streamlit = Streamlit( ... # To enable HTTPS, create an ACM certificate and add the ARN below: load_balancer_enable_https=True, load_balancer_certificate_arn="arn:aws:acm:us-east-1:497891874516:certificate/6598c24a-d4fc-4f17-8ee0-0d3906eb705f", ... ) [V0_FILE]markdown:file="agno/misc/-_get_the_top_5_stories_you_can_try_accepting_and_declining_the_confirmation.md" --- title: - "Get the top 5 stories (you can try accepting and declining the confirmation)" category: misc source_lines: 35084-35109 line_count: 25 --- # - "Get the top 5 stories (you can try accepting and declining the confirmation)" agent.print_response( "What are the top 2 hackernews stories?", stream=True, console=console ) \`\`\` ## Usage \`\`\`bash pip install openai agno \`\`\` \`\`\`bash python human_in_the_loop.py \`\`\` [V0_FILE]markdown:file="agno/misc/-_set_to_false_to_send_to_yourself.md" --- title: - set to False to send to yourself category: misc source_lines: 53184-53188 line_count: 4 --- # - set to False to send to yourself DEMO_MODE = True today = datetime.now().strftime("%Y-%m-%d") [V0_FILE]markdown:file="agno/misc/-_set_to_true_to_print_email_to_console.md" --- title: - set to True to print email to console category: misc source_lines: 53183-53184 line_count: 1 --- # - set to True to print email to console [V0_FILE]markdown:file="agno/misc/-_show_me_the_most_recent_story_from_hacker_news.md" --- title: - "Show me the most recent story from Hacker News" category: misc source_lines: 35083-35084 line_count: 1 --- # - "Show me the most recent story from Hacker News" [V0_FILE]markdown:file="agno/misc/-_summarize_the_top_5_stories_on_hacker_news.md" --- title: - "Summarize the top 5 stories on Hacker News" category: misc source_lines: 34946-34947 line_count: 1 --- # - "Summarize the top 5 stories on Hacker News" [V0_FILE]markdown:file="agno/misc/-_what_are_the_top_3_hn_stories_right_now.md" --- title: - "What are the top 3 HN stories right now?" category: misc source_lines: 35082-35083 line_count: 1 --- # - "What are the top 3 HN stories right now?" [V0_FILE]markdown:file="agno/misc/-_what_are_the_trending_tech_discussions_on_hn_right_now.md" --- title: - "What are the trending tech discussions on HN right now?" category: misc source_lines: 34945-34946 line_count: 1 --- # - "What are the trending tech discussions on HN right now?" [V0_FILE]markdown:file="agno/misc/-_what_s_the_most_upvoted_story_today.md" --- title: - "What's the most upvoted story today?" category: misc source_lines: 34947-34970 line_count: 23 --- # - "What's the most upvoted story today?" agent.print_response("Summarize the top 5 stories on hackernews?", stream=True) \`\`\` ## Usage \`\`\`bash pip install openai httpx agno \`\`\` \`\`\`bash python custom_tools.py \`\`\` [V0_FILE]markdown:file="agno/misc/.md" --- title: ) category: misc source_lines: 51180-51211 line_count: 31 --- # ) \`\`\` ## Usage \`\`\`bash pip install openai anthropic mistralai \`\`\` \`\`\`bash export OPENAI_API_KEY=**** export ANTHROPIC_API_KEY=**** export DEEPSEEK_API_KEY=**** export MISTRAL_API_KEY=**** \`\`\` \`\`\`bash python multi_language_team.py \`\`\` [V0_FILE]markdown:file="agno/misc/1_create_a_new_bot_with_botfather_on_telegram_https_core_telegram_org_bots_features_creating-a-new-bot.md" --- title: 1. Create a new bot with BotFather on Telegram. https://core.telegram.org/bots/features#creating-a-new-bot category: misc source_lines: 78298-78299 line_count: 1 --- # 1. Create a new bot with BotFather on Telegram. https://core.telegram.org/bots/features#creating-a-new-bot [V0_FILE]markdown:file="agno/misc/2_get_the_token_from_botfather.md" --- title: 2. Get the token from BotFather. category: misc source_lines: 78299-78300 line_count: 1 --- # 2. Get the token from BotFather. [V0_FILE]markdown:file="agno/misc/3_send_a_message_to_the_bot.md" --- title: 3. Send a message to the bot. category: misc source_lines: 78300-78301 line_count: 1 --- # 3. Send a message to the bot. [V0_FILE]markdown:file="agno/misc/4_get_the_chat_id_by_going_to_the_url.md" --- title: 4. Get the chat_id by going to the URL: category: misc source_lines: 78301-78302 line_count: 1 --- # 4. Get the chat_id by going to the URL: [V0_FILE]markdown:file="agno/misc/access_multiple_previous_steps_output.md" --- title: Access Multiple Previous Steps Output category: misc source_lines: 56278-56303 line_count: 25 --- # Access Multiple Previous Steps Output Source: https://docs.agno.com/examples/workflows_2/06-workflows-advanced-concepts/access_multiple_previous_steps_output This example demonstrates **Workflows 2.0** advanced data flow capabilities This example demonstrates **Workflows 2.0** shows how to: 1. Access outputs from **specific named steps** (`get_step_content()`) 2. Aggregate **all previous outputs** (`get_all_previous_content()`) 3. Create comprehensive reports by combining multiple research sources ## Key Features: * **Step Output Access**: Retrieve data from any previous step by name or collectively. * **Custom Reporting**: Combine and analyze outputs from parallel or sequential steps. * **Streaming Support**: Real-time updates during execution. \`\`\`python access_multiple_previous_steps_output.py from agno.agent.agent import Agent from agno.tools.duckduckgo import DuckDuckGoTools from agno.tools.hackernews import HackerNewsTools from agno.workflow.v2.step import Step from agno.workflow.v2.types import StepInput, StepOutput from agno.workflow.v2.workflow import Workflow [V0_FILE]markdown:file="agno/misc/accuracy_with_given_answer.md" --- title: Accuracy with Given Answer category: misc source_lines: 33387-33415 line_count: 28 --- # Accuracy with Given Answer Source: https://docs.agno.com/examples/evals/accuracy/accuracy_with_given_answer Learn how to evaluate the accuracy of an Agno Agent's response with a given answer. For this example an agent won't be executed, but the given result will be evaluated against the expected output for correctness. ## Code \`\`\`python from typing import Optional from agno.eval.accuracy import AccuracyEval, AccuracyResult from agno.models.openai import OpenAIChat evaluation = AccuracyEval( model=OpenAIChat(id="o4-mini"), input="What is 10*5 then to the power of 2? do it step by step", expected_output="2500", num_iterations=1, ) result_with_given_answer: Optional[AccuracyResult] = evaluation.run_with_output( output="2500", print_results=True ) assert result_with_given_answer is not None and result_with_given_answer.avg_score >= 8 \`\`\` [V0_FILE]markdown:file="agno/misc/accuracy_with_teams.md" --- title: Accuracy with Teams category: misc source_lines: 33415-33430 line_count: 15 --- # Accuracy with Teams Source: https://docs.agno.com/examples/evals/accuracy/accuracy_with_teams Learn how to evaluate the accuracy of an Agno Team. ## Code \`\`\`python from typing import Optional from agno.agent import Agent from agno.eval.accuracy import AccuracyEval, AccuracyResult from agno.models.openai import OpenAIChat from agno.team.team import Team [V0_FILE]markdown:file="agno/misc/add_embedding_to_database.md" --- title: Add embedding to database category: misc source_lines: 5681-5683 line_count: 2 --- # Add embedding to database embeddings = CohereEmbedder().get_embedding("The quick brown fox jumps over the lazy dog.") [V0_FILE]markdown:file="agno/misc/add_in_the_query_and_the_agent_redirects_it_to_the_appropriate_agent.md" --- title: Add in the query and the agent redirects it to the appropriate agent category: misc source_lines: 51015-51020 line_count: 5 --- # Add in the query and the agent redirects it to the appropriate agent customer_support_team.print_response( "Hi Team, I want to build an educational platform where the models are have access to tons of study materials, How can Agno platform help me build this?", stream=True, ) [V0_FILE]markdown:file="agno/misc/add_memories_for_jane_doe.md" --- title: Add memories for Jane Doe category: misc source_lines: 16578-16593 line_count: 15 --- # Add memories for Jane Doe jane_doe_id = "jane_doe@example.com" print(f"\nUser: {jane_doe_id}") memory_id_1 = memory.add_user_memory( memory=UserMemory(memory="The user's name is Jane Doe", topics=["name"]), user_id=jane_doe_id, ) memory_id_2 = memory.add_user_memory( memory=UserMemory(memory="She likes to play tennis", topics=["hobbies"]), user_id=jane_doe_id, ) memories = memory.get_user_memories(user_id=jane_doe_id) print("Memories:") pprint(memories) [V0_FILE]markdown:file="agno/misc/add_python_libraries.md" --- title: Add Python Libraries category: misc source_lines: 86791-86906 line_count: 115 --- # Add Python Libraries Source: https://docs.agno.com/workspaces/workspace-management/python-packages Agno templates are setup to manage dependencies using a [pyproject.toml](https://packaging.python.org/en/latest/specifications/declaring-project-metadata/#declaring-project-metadata) file, **which is used to generate the `requirements.txt` file using [uv](https://github.com/astral-sh/uv) or [pip-tools](https://pip-tools.readthedocs.io/en/latest/).** Adding or Updating a python library is a 2 step process: 1. Add library to the `pyproject.toml` file 2. Auto-Generate the `requirements.txt` file We highly recommend auto-generating the `requirements.txt` file using this process. ## Update pyproject.toml * Open the `pyproject.toml` file * Add new libraries to the dependencies section. ## Generate requirements After updating the `dependencies` in the `pyproject.toml` file, auto-generate the `requirements.txt` file using a helper script or running `pip-compile` directly. \`\`\`bash terminal ./scripts/generate_requirements.sh \`\`\` \`\`\`bash pip compile pip-compile \ --no-annotate \ --pip-args "--no-cache-dir" \ -o requirements.txt pyproject.toml \`\`\` If you'd like to upgrade all python libraries to their latest version, run: \`\`\`bash terminal ./scripts/generate_requirements.sh upgrade \`\`\` \`\`\`bash pip compile pip-compile \ --upgrade \ --no-annotate \ --pip-args "--no-cache-dir" \ -o requirements.txt pyproject.toml \`\`\` ## Rebuild Images After updating the `requirements.txt` file, rebuild your images. ### Rebuild dev images \`\`\`bash terminal ag ws up --env dev --infra docker --type image \`\`\` \`\`\`bash short options ag ws up -e dev -i docker -t image \`\`\` ### Rebuild production images Remember to [authenticate with ECR](workspaces/workspace-management/production-app#ecr-images) if needed. \`\`\`bash terminal ag ws up --env prd --infra aws --type image \`\`\` \`\`\`bash short options ag ws up -e prd -i aws -t image \`\`\` ## Recreate Resources After rebuilding images, recreate the resources. ### Recreate dev containers \`\`\`bash terminal ag ws restart --env dev --infra docker --type container \`\`\` \`\`\`bash short options ag ws restart -e dev -c docker -t container \`\`\` ### Update ECS services \`\`\`bash terminal ag ws patch --env prd --infra aws --name service \`\`\` \`\`\`bash short options ag ws patch -e prd -i aws -n service \`\`\` [V0_FILE]markdown:file="agno/misc/add_references_true.md" --- title: add_references=True, category: misc source_lines: 13921-13922 line_count: 1 --- # add_references=True, [V0_FILE]markdown:file="agno/misc/add_secrets.md" --- title: Add Secrets category: misc source_lines: 86906-86938 line_count: 32 --- # Add Secrets Source: https://docs.agno.com/workspaces/workspace-management/secrets Secret management is a critical part of your application security and should be taken seriously. Local secrets are defined in the `worspace/secrets` directory which is excluded from version control (see `.gitignore`). Its contents should be handled with the same security as passwords. Production secrets are managed by [AWS Secrets Manager](https://docs.aws.amazon.com/secretsmanager/latest/userguide/intro.html). Incase you're missing the secrets dir, copy `workspace/example_secrets` ## Development Secrets Apps running locally can read secrets using a `yaml` file, for example: \`\`\`python dev_resources.py dev_fastapi = FastApi( ... # Read secrets from secrets/dev_app_secrets.yml secrets_file=ws_settings.ws_root.joinpath("workspace/secrets/dev_app_secrets.yml"), ) \`\`\` ## Production Secrets `AWS Secrets` are used to manage production secrets, which are read by the production apps. \`\`\`python prd_resources.py [V0_FILE]markdown:file="agno/misc/add_storage_to_the_agent.md" --- title: Add storage to the Agent category: misc source_lines: 69225-69238 line_count: 13 --- # Add storage to the Agent agent = Agent(storage=storage) \`\`\` ## Params ## Developer Resources * View [Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/storage/sqllite_storage/sqlite_storage_for_agent.py) [V0_FILE]markdown:file="agno/misc/add_storage_to_the_team.md" --- title: Add storage to the Team category: misc source_lines: 24523-24536 line_count: 13 --- # Add storage to the Team team = Team(storage=storage) \`\`\` ## Params ## Developer Resources * View [Cookbook](https://github.com/agno-agi/agno/blob/main/cookbook/storage/mysql_storage/mysql_storage_for_team.py) [V0_FILE]markdown:file="agno/misc/add_user_memories.md" --- title: Add user memories category: misc source_lines: 63338-63348 line_count: 10 --- # Add user memories user_id = "user@example.com" memory.add_user_memory( memory=UserMemory( memory="The user has a premium subscription", topics=["subscription", "account"] ), user_id=user_id ) [V0_FILE]markdown:file="agno/misc/additional_agents_for_multi-step_condition.md" --- title: Additional agents for multi-step condition category: misc source_lines: 55350-55361 line_count: 11 --- # Additional agents for multi-step condition trend_analyzer_agent = Agent( name="Trend Analyzer", instructions="Analyze trends and patterns from research data", ) fact_checker_agent = Agent( name="Fact Checker", instructions="Verify facts and cross-reference information", ) [V0_FILE]markdown:file="agno/misc/advanced.md" --- title: Advanced category: misc source_lines: 81631-81654 line_count: 23 --- # Advanced Source: https://docs.agno.com/workflows/advanced **Workflows are all about control and flexibility.** Your workflow logic is just a python function, so you have full control over the workflow logic. You can: * Validate input before processing * Depending on the input, spawn agents and run them in parallel * Cache results as needed * Correct any intermediate errors * Stream the output * Return a single or multiple outputs **This level of control is critical for reliability.** ## Streaming It is important to understand that when you build a workflow, you are writing a python function, meaning you decide if the function streams the output or not. To stream the output, yield an `Iterator[RunResponse]` from the `run()` method of your workflow. \`\`\`python news_report_generator.py [V0_FILE]markdown:file="agno/misc/advanced_concepts_in_workflows.md" --- title: Advanced Concepts in Workflows category: misc source_lines: 82322-82435 line_count: 113 --- # Advanced Concepts in Workflows Source: https://docs.agno.com/workflows_2/advanced Explore advanced features and concepts in the Agno workflow system, including custom functions, error handling, and streaming capabilities. ## How Custom Functions Work Custom functions provide flexibility by allowing developers to define specific logic for step execution. They can be used to `preprocess inputs`, `call agents`, and `postprocess outputs`. * **executor**: Step can be defined with a custom execution function that handles the step logic. * **Integration with Agents and Teams**: Custom functions can interact with agents and teams, leveraging their capabilities. While defining a `Step`, you can specify a custom function as an `executor`. This function should accept a `StepInput` object and return a `StepOutput` object. \`\`\`python content_planning_step = Step( name="Content Planning Step", executor=custom_content_planning_function, ) def custom_content_planning_function(step_input: StepInput) -> StepOutput: """ Custom function that does intelligent content planning with context awareness """ message = step_input.message previous_step_content = step_input.previous_step_content # Create intelligent planning prompt planning_prompt = f""" STRATEGIC CONTENT PLANNING REQUEST: Core Topic: {message} Research Results: {previous_step_content[:500] if previous_step_content else "No research results"} Planning Requirements: 1. Create a comprehensive content strategy based on the research 2. Leverage the research findings effectively 3. Identify content formats and channels 4. Provide timeline and priority recommendations 5. Include engagement and distribution strategies Please create a detailed, actionable content plan. """ try: response = content_planner.run(planning_prompt) enhanced_content = f""" ## Strategic Content Plan **Planning Topic:** {message} **Research Integration:** {"✓ Research-based" if previous_step_content else "✗ No research foundation"} **Content Strategy:** {response.content} **Custom Planning Enhancements:** - Research Integration: {"High" if previous_step_content else "Baseline"} - Strategic Alignment: Optimized for multi-channel distribution - Execution Ready: Detailed action items included """.strip() return StepOutput(content=enhanced_content, response=response) except Exception as e: return StepOutput( content=f"Custom content planning failed: {str(e)}", success=False, ) \`\`\` Just make sure to follow this structure and return the output as a `StepOutput` object. \`\`\`python def custom_content_planning_function(step_input: StepInput) -> StepOutput: # Custom preprocessing # Call the agent # Custom postprocessing return StepOutput(content=enhanced_content) \`\`\` **More Examples**: * [Step with a Custom Function](/examples/workflows_2/01-basic-workflows/step_with_function) ## Run a Workflow non-blocking (in the background) You can run a workflow as a non-blocking task by passing `background=True` to `Workflow.arun()`. This will return a `WorkflowRunResponse` object with a `run_id` that you can use to poll for the result of the workflow until it is completed. This feature is only available for async workflows using `.arun()`. For long-running workflows, you can poll for the result using `result = workflow.get_run(run_id)` which returns the updated `WorkflowRunResponse`. Use `.has_completed()` to check if the workflow has finished executing. This is particularly useful for workflows that involve time-consuming operations like large-scale data processing, multi-step research tasks, or batch operations that you don't want to block your main application thread. \`\`\`python import asyncio from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.run.base import RunStatus from agno.run.v2.workflow import WorkflowRunResponse from agno.storage.sqlite import SqliteStorage from agno.team import Team from agno.tools.googlesearch import GoogleSearchTools from agno.tools.hackernews import HackerNewsTools from agno.utils.pprint import pprint_run_response from agno.workflow.v2.step import Step from agno.workflow.v2.workflow import Workflow [V0_FILE]markdown:file="agno/misc/advanced_mcp_usage.md" --- title: Advanced MCP Usage category: misc source_lines: 71720-71767 line_count: 47 --- # Advanced MCP Usage Source: https://docs.agno.com/tools/mcp/advanced_usage Agno's MCP integration also supports handling connections to multiple servers, specifying server parameters and using your own MCP servers: ## Connecting to Multiple MCP Servers You can use multiple MCP servers in a single agent by using the `MultiMCPTools` class. \`\`\`python multiple_mcp_servers.py import asyncio from os import getenv from agno.agent import Agent from agno.tools.mcp import MultiMCPTools async def run_agent(message: str) -> None: # Initialize the MCP tools mcp_tools = MultiMCPTools( [ "npx -y @openbnb/mcp-server-airbnb --ignore-robots-txt", "npx -y @modelcontextprotocol/server-brave-search", ], env={ "BRAVE_API_KEY": getenv("BRAVE_API_KEY"), }, timeout_seconds=30, ) # Connect to the MCP servers await mcp_tools.connect() # Use the MCP tools with an Agent agent = Agent( tools=[mcp_tools], markdown=True, show_tool_calls=True, ) await agent.aprint_response(message) # Close the MCP connection await mcp_tools.close() [V0_FILE]markdown:file="agno/misc/advanced_research_topics_to_explore.md" --- title: Advanced research topics to explore: category: misc source_lines: 7476-7530 line_count: 54 --- # Advanced research topics to explore: """ Quantum Science & Computing: 1. "Investigate recent breakthroughs in quantum error correction" 2. "Analyze the development of topological quantum computing" 3. "Research quantum machine learning algorithms and applications" 4. "Explore advances in quantum sensing technologies" Biotechnology & Medicine: 1. "Examine recent developments in mRNA vaccine technology" 2. "Analyze breakthroughs in organoid research" 3. "Investigate advances in precision medicine" 4. "Research developments in neurotechnology" Materials Science: 1. "Explore recent advances in metamaterials" 2. "Analyze developments in 2D materials beyond graphene" 3. "Research progress in self-healing materials" 4. "Investigate new battery technologies" Artificial Intelligence: 1. "Examine recent advances in foundation models" 2. "Analyze developments in AI safety research" 3. "Research progress in neuromorphic computing" 4. "Investigate advances in explainable AI" """ \`\`\` ## Usage \`\`\`bash pip install openai exa_py agno \`\`\` \`\`\`bash export OPENAI_API_KEY=**** export EXA_API_KEY=**** \`\`\` \`\`\`bash python research_agent_exa.py \`\`\` [V0_FILE]markdown:file="agno/misc/ag-ui_app.md" --- title: AG-UI App category: misc source_lines: 3052-3150 line_count: 98 --- # AG-UI App Source: https://docs.agno.com/applications/ag-ui/introduction Expose your Agno Agent as a AG-UI compatible app AG-UI, or [Agent-User Interaction Protocol](https://github.com/ag-ui-protocol/ag-ui), is a protocol standarizing how AI agents connect to front-end applications. ## Example usage \`\`\`bash pip install agno ag-ui-protocol \`\`\` Now let's run a `AGUIApp` exposing an Agno Agent. You can use the previous code! You can use [Dojo](https://github.com/ag-ui-protocol/ag-ui/tree/main/typescript-sdk/apps/dojo), an advanced and customizable option to use as frontend for AG-UI agents. 1. Clone the project: `git clone https://github.com/ag-ui-protocol/ag-ui.git` 2. Follow the instructions [here](https://github.com/ag-ui-protocol/ag-ui/tree/main/typescript-sdk/apps/dojo) to learn how to install the needed dependencies and run the project. 3. Remember to install the dependencies in `/ag-ui/typescript-sdk` with `pnpm install`, and to build the Agno package in `/integrations/agno` with `pnpm run build`. 4. You can now run your Dojo! It will show our Agno agent as one of the available options. Done! If you are running Dojo as your front-end, you can now go to [http://localhost:3000](http://localhost:3000) in your browser and chat with your Agno Agent. ![AG-UI Dojo screenshot](https://mintlify.s3.us-west-1.amazonaws.com/agno/images/agui-dojo.png) You can see more examples in our [AG-UI integration examples](/examples/applications/ag-ui) section. ## Core Components * `AGUIApp`: Wraps Agno agents/teams for in a FastAPI app. * `serve`: Serves the FastAPI AG-UI app using Uvicorn. `AGUIApp` uses helper functions for routing. ## `AGUIApp` Class Main entry point for Agno AG-UI apps. ### Initialization Parameters | Parameter | Type | Default | Description | | ------------- | -------------------------- | ------- | ------------------------------------------------ | | `agent` | `Optional[Agent]` | `None` | Agno `Agent` instance. | | `team` | `Optional[Team]` | `None` | Agno `Team` instance. | | `settings` | `Optional[APIAppSettings]` | `None` | API configuration. Defaults if `None`. | | `api_app` | `Optional[FastAPI]` | `None` | Existing FastAPI app. New one created if `None`. | | `router` | `Optional[APIRouter]` | `None` | Existing APIRouter. New one created if `None`. | | `app_id` | `Optional[str]` | `None` | App identifier (autogenerated if not set). | | `name` | `Optional[str]` | `None` | Name for the App. | | `description` | `Optional[str]` | `None` | Description for the App. | *Provide `agent` or `team`, not both.* ### Key Method | Method | Parameters | Return Type | Description | | --------- | ------------------------ | ----------- | ------------------------------------------------------------------------------------------- | | `get_app` | `use_async: bool = True` | `FastAPI` | Returns configured FastAPI app (async by default). Sets prefix, error handlers, CORS, docs. | ## Endpoints Endpoints are available at the specified `prefix` (default `/v1`). ### 1. `POST /agui` This is the main entrypoint to interact with your Agno Agent or Team. It expects a `RunAgentInput` object (from the `ag-ui-protocol` package) as defined by the protocol. You can read more about it in [their docs](https://docs.ag-ui.com/quickstart/server). ## Serving the Application (`serve`) Serves the FastAPI app using Uvicorn. ### Parameters | Parameter | Type | Default | Description | | --------- | --------------------- | ------------- | ------------------------------------------------- | | `app` | `Union[str, FastAPI]` | `N/A` | FastAPI app instance or import string (Required). | | `host` | `str` | `"localhost"` | Host to bind. | | `port` | `int` | `7777` | Port to bind. | | `reload` | `bool` | `False` | Enable auto-reload for development. | You can check some usage examples in our [AG-UI integration examples](/examples/applications/ag-ui) section. You can also check the [CopilotKit docs](https://docs.copilotkit.ai/agno) on working with Agno, to learn more on how to build the UI side. [V0_FILE]markdown:file="agno/misc/agent.md" --- title: Agent category: misc source_lines: 67096-67104 line_count: 8 --- # Agent Source: https://docs.agno.com/reference/agents/agent [V0_FILE]markdown:file="agno/misc/agent_agent.md" --- title: agent = Agent( category: misc source_lines: 17789-17790 line_count: 1 --- # agent = Agent( [V0_FILE]markdown:file="agno/misc/agent_app_fastapi_streamlit_and_postgres.md" --- title: Agent App: FastAPI, Streamlit and Postgres category: misc source_lines: 85312-85346 line_count: 34 --- # Agent App: FastAPI, Streamlit and Postgres Source: https://docs.agno.com/workspaces/agent-app/local The Agent App is our go-to workspace for building agentic systems. It contains: * A FastAPI server for serving Agents, Teams and Workflows. * A streamlit application for debugging and testing. This streamlit app is very versatile and can be used as an admin interface for the agentic system and shows all sorts of data. * A postgres database for session and vector storage. It's designed to run locally using docker and in production on AWS. ## Next Congratulations on running your AI App locally. Next Steps: * [Run your Agent App on AWS](/workspaces/agent-app/aws) * Read how to [update workspace settings](/workspaces/workspace-management/workspace-settings) * Read how to [create a git repository for your workspace](/workspaces/workspace-management/git-repo) * Read how to [manage the development application](/workspaces/workspace-management/development-app) * Read how to [format and validate your code](/workspaces/workspace-management/format-and-validate) * Read how to [add python libraries](/workspaces/workspace-management/install) * Chat with us on [discord](https://agno.link/discord) [V0_FILE]markdown:file="agno/misc/agent_print_response.md" --- title: agent.print_response( category: misc source_lines: 16459-16460 line_count: 1 --- # agent.print_response( [V0_FILE]markdown:file="agno/misc/agent_print_response_share_the_nvda_stock_price_and_analyst_recommendations_stream_true.md" --- title: agent.print_response("Share the NVDA stock price and analyst recommendations", stream=True) category: misc source_lines: 43523-43559 line_count: 36 --- # agent.print_response("Share the NVDA stock price and analyst recommendations", stream=True) agent.print_response("Summarize fundamentals for TSLA", stream=True) \`\`\` ## Usage \`\`\`bash export LANGDB_API_KEY=xxx export LANGDB_PROJECT_ID=xxx \`\`\` \`\`\`bash pip install -U yfinance agno \`\`\` \`\`\`bash Mac python cookbook/models/langdb/finance_agent.py \`\`\` \`\`\`bash Windows python cookbook/models/langdb/finance_agent.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_session.md" --- title: Agent Session category: misc source_lines: 33933-34061 line_count: 128 --- # Agent Session Source: https://docs.agno.com/examples/getting-started/agent-session This example shows how to create an agent with persistent memory stored in a SQLite database. We set the session\_id on the agent when resuming the conversation, this way the previous chat history is preserved. Key features: * Stores conversation history in a SQLite database * Continues conversations across multiple sessions * References previous context in responses ## Code \`\`\`python agent_session.py import json from typing import Optional import typer from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.storage.sqlite import SqliteStorage from rich.console import Console from rich.json import JSON from rich.panel import Panel from rich.prompt import Prompt from rich import print console = Console() def create_agent(user: str = "user"): session_id: Optional[str] = None # Ask if user wants to start new session or continue existing one new = typer.confirm("Do you want to start a new session?") # Get existing session if user doesn't want a new one agent_storage = SqliteStorage( table_name="agent_sessions", db_file="tmp/agents.db" ) if not new: existing_sessions = agent_storage.get_all_session_ids(user) if len(existing_sessions) > 0: session_id = existing_sessions[0] agent = Agent( user_id=user, # Set the session_id on the agent to resume the conversation session_id=session_id, model=OpenAIChat(id="gpt-4o"), storage=agent_storage, # Add chat history to messages add_history_to_messages=True, num_history_responses=3, markdown=True, ) if session_id is None: session_id = agent.session_id if session_id is not None: print(f"Started Session: {session_id}\n") else: print("Started Session\n") else: print(f"Continuing Session: {session_id}\n") return agent def print_messages(agent): """Print the current chat history in a formatted panel""" console.print( Panel( JSON( json.dumps( [ m.model_dump(include={"role", "content"}) for m in agent.memory.messages ] ), indent=4, ), title=f"Chat History for session_id: {agent.session_id}", expand=True, ) ) def main(user: str = "user"): agent = create_agent(user) print("Chat with an OpenAI agent!") exit_on = ["exit", "quit", "bye"] while True: message = Prompt.ask(f"[bold] :sunglasses: {user} [/bold]") if message in exit_on: break agent.print_response(message=message, stream=True, markdown=True) print_messages(agent) if __name__ == "__main__": typer.run(main) \`\`\` ## Usage \`\`\`bash pip install openai sqlalchemy agno \`\`\` \`\`\`bash python agent_session.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_state.md" --- title: Agent State category: misc source_lines: 34061-34085 line_count: 24 --- # Agent State Source: https://docs.agno.com/examples/getting-started/agent-state This example shows how to create an agent that maintains state across interactions. It demonstrates a simple counter mechanism, but this pattern can be extended to more complex state management like maintaining conversation context, user preferences, or tracking multi-step processes. Example prompts to try: * "Increment the counter 3 times and tell me the final count" * "What's our current count? Add 2 more to it" * "Let's increment the counter 5 times, but tell me each step" * "Add 4 to our count and remind me where we started" * "Increase the counter twice and summarize our journey" ## Code \`\`\`python agent_state.py from textwrap import dedent from agno.agent import Agent from agno.models.openai import OpenAIChat [V0_FILE]markdown:file="agno/misc/agent_team.md" --- title: Agent Team category: misc source_lines: 34157-34273 line_count: 116 --- # Agent Team Source: https://docs.agno.com/examples/getting-started/agent-team This example shows how to create a powerful team of AI agents working together to provide comprehensive financial analysis and news reporting. The team consists of: 1. Web Agent: Searches and analyzes latest news 2. Finance Agent: Analyzes financial data and market trends 3. Lead Editor: Coordinates and combines insights from both agents Example prompts to try: * "What's the latest news and financial performance of Apple (AAPL)?" * "Analyze the impact of AI developments on NVIDIA's stock (NVDA)" * "How are EV manufacturers performing? Focus on Tesla (TSLA) and Rivian (RIVN)" * "What's the market outlook for semiconductor companies like AMD and Intel?" * "Summarize recent developments and stock performance of Microsoft (MSFT)" ## Code \`\`\`python agent_team.py from textwrap import dedent from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.duckduckgo import DuckDuckGoTools from agno.tools.yfinance import YFinanceTools web_agent = Agent( name="Web Agent", role="Search the web for information", model=OpenAIChat(id="gpt-4o"), tools=[DuckDuckGoTools()], instructions=dedent("""\ You are an experienced web researcher and news analyst! 🔍 Follow these steps when searching for information: 1. Start with the most recent and relevant sources 2. Cross-reference information from multiple sources 3. Prioritize reputable news outlets and official sources 4. Always cite your sources with links 5. Focus on market-moving news and significant developments Your style guide: - Present information in a clear, journalistic style - Use bullet points for key takeaways - Include relevant quotes when available - Specify the date and time for each piece of news - Highlight market sentiment and industry trends - End with a brief analysis of the overall narrative - Pay special attention to regulatory news, earnings reports, and strategic announcements\ """), show_tool_calls=True, markdown=True, ) finance_agent = Agent( name="Finance Agent", role="Get financial data", model=OpenAIChat(id="gpt-4o"), tools=[ YFinanceTools(stock_price=True, analyst_recommendations=True, company_info=True) ], instructions=dedent("""\ You are a skilled financial analyst with expertise in market data! 📊 Follow these steps when analyzing financial data: 1. Start with the latest stock price, trading volume, and daily range 2. Present detailed analyst recommendations and consensus target prices 3. Include key metrics: P/E ratio, market cap, 52-week range 4. Analyze trading patterns and volume trends 5. Compare performance against relevant sector indices Your style guide: - Use tables for structured data presentation - Include clear headers for each data section - Add brief explanations for technical terms - Highlight notable changes with emojis (📈 📉) - Use bullet points for quick insights - Compare current values with historical averages - End with a data-driven financial outlook\ """), show_tool_calls=True, markdown=True, ) agent_team = Agent( team=[web_agent, finance_agent], model=OpenAIChat(id="gpt-4o"), instructions=dedent("""\ You are the lead editor of a prestigious financial news desk! 📰 Your role: 1. Coordinate between the web researcher and financial analyst 2. Combine their findings into a compelling narrative 3. Ensure all information is properly sourced and verified 4. Present a balanced view of both news and data 5. Highlight key risks and opportunities Your style guide: - Start with an attention-grabbing headline - Begin with a powerful executive summary - Present financial data first, followed by news context - Use clear section breaks between different types of information - Include relevant charts or tables when available - Add 'Market Sentiment' section with current mood - Include a 'Key Takeaways' section at the end - End with 'Risk Factors' when appropriate - Sign off with 'Market Watch Team' and the current date\ """), add_datetime_to_instructions=True, show_tool_calls=True, markdown=True, ) [V0_FILE]markdown:file="agno/misc/agent_teams_deprecated.md" --- title: Agent Teams [Deprecated] category: misc source_lines: 2481-2555 line_count: 74 --- # Agent Teams [Deprecated] Source: https://docs.agno.com/agents/teams Agent Teams were an initial implementation of our multi-agent architecture (2023-2025) that use a transfer/handoff mechanism. After 2 years of experimentation, we've learned that this mechanism is not scalable and do NOT recommend it for complex multi-agent systems. Please use the new [Teams](/teams) architecture instead. We can combine multiple Agents to form a team and tackle tasks as a cohesive unit. Here's a simple example that converts an agent into a team to write an article about the top stories on hackernews. \`\`\`python hackernews_team.py from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.hackernews import HackerNewsTools from agno.tools.duckduckgo import DuckDuckGoTools from agno.tools.newspaper4k import Newspaper4kTools hn_researcher = Agent( name="HackerNews Researcher", model=OpenAIChat("gpt-4o"), role="Gets top stories from hackernews.", tools=[HackerNewsTools()], ) web_searcher = Agent( name="Web Searcher", model=OpenAIChat("gpt-4o"), role="Searches the web for information on a topic", tools=[DuckDuckGoTools()], add_datetime_to_instructions=True, ) article_reader = Agent( name="Article Reader", model=OpenAIChat("gpt-4o"), role="Reads articles from URLs.", tools=[Newspaper4kTools()], ) hn_team = Agent( name="Hackernews Team", model=OpenAIChat("gpt-4o"), team=[hn_researcher, web_searcher, article_reader], instructions=[ "First, search hackernews for what the user is asking about.", "Then, ask the article reader to read the links for the stories to get more information.", "Important: you must provide the article reader with the links to read.", "Then, ask the web searcher to search for each story to get more information.", "Finally, provide a thoughtful and engaging summary.", ], show_tool_calls=True, markdown=True, ) hn_team.print_response("Write an article about the top 2 stories on hackernews", stream=True) \`\`\` Run the script to see the output. \`\`\`bash pip install -U openai duckduckgo-search newspaper4k lxml_html_clean agno python hackernews_team.py \`\`\` ## How to build Agent Teams 1. Add a `name` and `role` parameter to the member Agents. 2. Create a Team Leader that can delegate tasks to team-members. 3. Use your Agent team just like you would use a regular Agent. [V0_FILE]markdown:file="agno/misc/agent_that_returns_a_structured_output.md" --- title: Agent that returns a structured output category: misc source_lines: 49526-49561 line_count: 35 --- # Agent that returns a structured output structured_output_agent = Agent( model=xAI(id="grok-2-latest"), description="You write movie scripts.", response_model=MovieScript, ) structured_output_agent.print_response("Llamas ruling the world") \`\`\` ## Usage \`\`\`bash export XAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno rich \`\`\` \`\`\`bash python cookbook/models/xai/structured_output.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_that_uses_a_json_schema_output.md" --- title: Agent that uses a JSON schema output category: misc source_lines: 39700-39741 line_count: 41 --- # Agent that uses a JSON schema output json_schema_output_agent = Agent( model=Cerebras(id="llama-4-scout-17b-16e-instruct"), description="You are a helpful assistant. Summarize the movie script based on the location in a JSON object.", response_model=MovieScript, ) json_schema_output_agent.print_response("New York") \`\`\` ## Usage \`\`\`bash export CEREBRAS_API_KEY=xxx \`\`\` \`\`\`bash pip install -U cerebras-cloud-sdk agno \`\`\` \`\`\`bash Mac python cookbook/models/cerebras/basic_json_schema.py \`\`\` \`\`\`bash Windows python cookbook/models/cerebras/basic_json_schema.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_that_uses_a_structured_output.md" --- title: Agent that uses a structured output category: misc source_lines: 45633-45675 line_count: 42 --- # Agent that uses a structured output structured_output_agent = Agent( model=Nebius(id="Qwen/Qwen3-30B-A3B"), description="You are a helpful assistant. Summarize the movie script based on the location in a JSON object.", response_model=MovieScript, debug_mode=True, ) structured_output_agent.print_response("New York") \`\`\` ## Usage \`\`\`bash export NEBIUS_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno \`\`\` \`\`\`bash Mac python cookbook/models/nebius/structured_output.py \`\`\` \`\`\`bash Windows python cookbook/models/nebius/structured_output.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_that_uses_json_mode.md" --- title: Agent that uses JSON mode category: misc source_lines: 48003-48010 line_count: 7 --- # Agent that uses JSON mode json_mode_agent = Agent( model=Together(id="meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo"), description="You write movie scripts.", response_model=MovieScript, ) [V0_FILE]markdown:file="agno/misc/agent_that_uses_structured_outputs.md" --- title: Agent that uses structured outputs category: misc source_lines: 47652-47660 line_count: 8 --- # Agent that uses structured outputs structured_output_agent = Agent( model=OpenAIResponses(id="gpt-4o"), description="You write movie scripts.", response_model=MovieScript, ) [V0_FILE]markdown:file="agno/misc/agent_with_async_tool_usage.md" --- title: Agent with Async Tool Usage category: misc source_lines: 42858-42914 line_count: 56 --- # Agent with Async Tool Usage Source: https://docs.agno.com/examples/models/ibm/async_tool_use ## Code \`\`\`python cookbook/models/ibm/watsonx/async_tool_use.py import asyncio from agno.agent import Agent from agno.models.ibm import WatsonX from agno.tools.duckduckgo import DuckDuckGoTools agent = Agent( model=WatsonX(id="meta-llama/llama-3-3-70b-instruct"), tools=[DuckDuckGoTools()], show_tool_calls=True, markdown=True, ) asyncio.run(agent.aprint_response("Whats happening in France?", stream=True)) \`\`\` ## Usage \`\`\`bash export IBM_WATSONX_API_KEY=xxx export IBM_WATSONX_PROJECT_ID=xxx \`\`\` \`\`\`bash pip install -U ibm-watsonx-ai duckduckgo-search agno \`\`\` \`\`\`bash Mac python cookbook/models/ibm/watsonx/async_tool_use.py \`\`\` \`\`\`bash Windows python cookbook\models\ibm\watsonx\async_tool_use.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_with_image_input.md" --- title: Agent with Image Input category: misc source_lines: 37881-37905 line_count: 24 --- # Agent with Image Input Source: https://docs.agno.com/examples/models/aws/bedrock/image_agent AWS Bedrock supports image input with models like `amazon.nova-pro-v1:0`. You can use this to analyze images and get information about them. ## Code \`\`\`python cookbook/models/aws/bedrock/image_agent.py from pathlib import Path from agno.agent import Agent from agno.media import Image from agno.models.aws import AwsBedrock from agno.tools.duckduckgo import DuckDuckGoTools agent = Agent( model=AwsBedrock(id="amazon.nova-pro-v1:0"), tools=[DuckDuckGoTools()], markdown=True, ) image_path = Path(__file__).parent.joinpath("sample.jpg") [V0_FILE]markdown:file="agno/misc/agent_with_media.md" --- title: Agent With Media category: misc source_lines: 10508-10576 line_count: 68 --- # Agent With Media Source: https://docs.agno.com/examples/applications/whatsapp/agent_with_media This example shows how to use the files with whatsapp app. ## Code \`\`\`python cookbook/apps/whatsapp/agent_with_media.py from agno.agent import Agent from agno.app.whatsapp.app import WhatsappAPI from agno.models.google import Gemini media_agent = Agent( name="Media Agent", model=Gemini(id="gemini-2.0-flash"), add_history_to_messages=True, num_history_responses=3, add_datetime_to_instructions=True, markdown=True, ) whatsapp_app = WhatsappAPI( agent=media_agent, name="Media Agent", app_id="media_agent", description="A agent that can send media to the user.", ) app = whatsapp_app.get_app() if __name__ == "__main__": whatsapp_app.serve(app="agent_with_media:app", port=8000, reload=True) \`\`\` ## Usage \`\`\`bash export GOOGLE_API_KEY=xxx \`\`\` \`\`\`bash pip install -U agno google-generativeai "uvicorn[standard]" \`\`\` \`\`\`bash Mac python cookbook/apps/whatsapp/agent_with_media.py \`\`\` \`\`\`bash Windows python cookbook/apps/whatsapp/agent_with_media.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_with_pdf_input_local_file.md" --- title: Agent with PDF Input (Local File) category: misc source_lines: 47414-47431 line_count: 17 --- # Agent with PDF Input (Local File) Source: https://docs.agno.com/examples/models/openai/responses/pdf_input_local ## Code \`\`\`python cookbook/models/openai/responses/pdf_input_local.py from pathlib import Path from agno.agent import Agent from agno.media import File from agno.models.openai.responses import OpenAIResponses from agno.utils.media import download_file pdf_path = Path(__file__).parent.joinpath("ThaiRecipes.pdf") [V0_FILE]markdown:file="agno/misc/agent_with_pdf_input_url.md" --- title: Agent with PDF Input (URL) category: misc source_lines: 47481-47539 line_count: 58 --- # Agent with PDF Input (URL) Source: https://docs.agno.com/examples/models/openai/responses/pdf_input_url ## Code \`\`\`python cookbook/models/openai/responses/pdf_input_url.py from agno.agent import Agent from agno.media import File from agno.models.openai.responses import OpenAIResponses agent = Agent( model=OpenAIResponses(id="gpt-4o-mini"), tools=[{"type": "file_search"}, {"type": "web_search_preview"}], markdown=True, ) agent.print_response( "Summarize the contents of the attached file and search the web for more information.", files=[File(url="https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf")], ) print("Citations:") print(agent.run_response.citations) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai agno \`\`\` \`\`\`bash Mac python cookbook/models/openai/responses/pdf_input_url.py \`\`\` \`\`\`bash Windows python cookbook/models/openai/responses/pdf_input_url.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_with_reasoning_effort.md" --- title: Agent with Reasoning Effort category: misc source_lines: 46832-46885 line_count: 53 --- # Agent with Reasoning Effort Source: https://docs.agno.com/examples/models/openai/chat/reasoning_effort ## Code \`\`\`python cookbook/reasoning/models/openai/reasoning_effort.py from agno.agent import Agent from agno.models.openai import OpenAIChat from agno.tools.yfinance import YFinanceTools agent = Agent( model=OpenAIChat(id="o3-mini", reasoning_effort="high"), tools=[YFinanceTools(enable_all=True)], show_tool_calls=True, markdown=True, ) agent.print_response("Write a report on the NVDA, is it a good buy?", stream=True) \`\`\` ## Usage \`\`\`bash export OPENAI_API_KEY=xxx \`\`\` \`\`\`bash pip install -U openai yfinance agno \`\`\` \`\`\`bash Mac python cookbook/reasoning/models/openai/reasoning_effort.py \`\`\` \`\`\`bash Windows python cookbook/reasoning/models/openai/reasoning_effort.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_with_session_summaries.md" --- title: Agent with Session Summaries category: misc source_lines: 17130-17157 line_count: 27 --- # Agent with Session Summaries Source: https://docs.agno.com/examples/concepts/memory/07-agent-with-summaries This example demonstrates how to create session summaries. To enable this, set `enable_session_summaries=True` in the Agent config. ## Code \`\`\`python cookbook/agent_concepts/memory/08_agent_with_summaries.py from agno.agent.agent import Agent from agno.memory.v2.db.sqlite import SqliteMemoryDb from agno.memory.v2.memory import Memory from agno.memory.v2.summarizer import SessionSummarizer from agno.models.anthropic.claude import Claude from rich.pretty import pprint memory_db = SqliteMemoryDb(table_name="memory", db_file="tmp/memory.db") memory = Memory( db=memory_db, summarizer=SessionSummarizer(model=Claude(id="claude-3-5-sonnet-20241022")), ) [V0_FILE]markdown:file="agno/misc/agent_with_storage.md" --- title: Agent with Storage category: misc source_lines: 48597-48663 line_count: 66 --- # Agent with Storage Source: https://docs.agno.com/examples/models/vllm/storage ## Code \`\`\`python cookbook/models/vllm/storage.py from agno.agent import Agent from agno.models.vllm import vLLM from agno.storage.postgres import PostgresStorage from agno.tools.duckduckgo import DuckDuckGoTools DB_URL = "postgresql+psycopg://ai:ai@localhost:5532/ai" agent = Agent( model=vLLM(id="Qwen/Qwen2.5-7B-Instruct"), storage=PostgresStorage(table_name="agent_sessions", db_url=DB_URL), tools=[DuckDuckGoTools()], add_history_to_messages=True, ) agent.print_response("How many people live in Canada?") agent.print_response("What is their national anthem called?") \`\`\` Ensure Postgres database is running. ## Usage \`\`\`bash pip install -U agno openai vllm sqlalchemy psycopg[binary] duckduckgo-search \`\`\` \`\`\`bash ./cookbook/scripts/run_pgvector.sh \`\`\` \`\`\`bash vllm serve Qwen/Qwen2.5-7B-Instruct \ --enable-auto-tool-choice \ --tool-call-parser hermes \ --dtype float16 \ --max-model-len 8192 \ --gpu-memory-utilization 0.9 \`\`\` \`\`\`bash python cookbook/models/vllm/storage.py \`\`\` [V0_FILE]markdown:file="agno/misc/agent_with_structured_output.md" --- title: Agent with Structured Output category: misc source_lines: 43240-43281 line_count: 41 --- # Agent with Structured Output Source: https://docs.agno.com/examples/models/ibm/structured_output ## Code \`\`\`python cookbook/models/ibm/watsonx/structured_output.py from typing import List from agno.agent import Agent, RunResponse from agno.models.ibm import WatsonX from pydantic import BaseModel, Field from rich.pretty import pprint class MovieScript(BaseModel): setting: str = Field( ..., description="Provide a nice setting for a blockbuster movie." ) ending: str = Field( ..., description="Ending of the movie. If not available, provide a happy ending.", ) genre: str = Field( ..., description="Genre of the movie. If not available, select action, thriller or romantic comedy.", ) name: str = Field(..., description="Give a name to this movie") characters: List[str] = Field(..., description="Name of characters for this movie.") storyline: str = Field( ..., description="3 sentence storyline for the movie. Make it exciting!" ) movie_agent = Agent( model=WatsonX(id="ibm/granite-20b-code-instruct"), description="You help people write movie scripts.", response_model=MovieScript, ) [V0_FILE]markdown:file="agno/misc/agent_with_structured_outputs.md" --- title: Agent with Structured Outputs category: misc source_lines: 47968-48003 line_count: 35 --- # Agent with Structured Outputs Source: https://docs.agno.com/examples/models/together/structured_output ## Code \`\`\`python cookbook/models/together/structured_output.py from typing import List from agno.agent import Agent, RunResponse # noqa from agno.models.together import Together from pydantic import BaseModel, Field from rich.pretty import pprint # noqa class MovieScript(BaseModel): setting: str = Field( ..., description="Provide a nice setting for a blockbuster movie." ) ending: str = Field( ..., description="Ending of the movie. If not available, provide a happy ending.", ) genre: str = Field( ..., description="Genre of the movie. If not available, select action, thriller or romantic comedy.", ) name: str = Field(..., description="Give a name to this movie") characters: List[str] = Field(..., description="Name of characters for this movie.") storyline: str = Field( ..., description="3 sentence storyline for the movie. Make it exciting!" ) [V0_FILE]markdown:file="agno/misc/agentic_deep_researcher.md" --- title: Agentic Deep Researcher category: misc source_lines: 51562-51748 line_count: 186 --- # Agentic Deep Researcher Source: https://docs.agno.com/examples/workflows/agentic-deep-researcher A multi-stage AI workflow for deep web research, analysis, and report generation using Agno, Scrapegraph, and Nebius AI Studio. This advanced example demonstrates how to build a multi-stage, AI-powered research workflow agent that automates comprehensive web research, analysis, and report generation using Agno, Scrapegraph, and Nebius AI Studio. Key capabilities: * Multi-stage research workflow: Automated pipeline for searching, analyzing, and reporting * Web scraping: Advanced data extraction with Scrapegraph * AI-powered analysis: Uses Nebius AI for intelligent synthesis ## Workflow stages 1. **Searcher**: Finds and extracts high-quality, up-to-date information from the web using Scrapegraph and Nebius AI. 2. **Analyst**: Synthesizes, interprets, and organizes the research findings, highlighting key insights and trends. 3. **Writer**: Crafts a clear, structured, and actionable report, including references and recommendations. ## How it works 1. Input a research topic or question. 2. The agent orchestrates web search, analysis, and report writing in sequence. 3. Results are presented in a user-friendly format. \`\`\`python deep_researcher_agent.py import os from agno.agent import Agent from agno.models.nebius import Nebius from dotenv import load_dotenv from typing import Iterator from agno.utils.log import logger from agno.utils.pprint import pprint_run_response from agno.tools.scrapegraph import ScrapeGraphTools from agno.workflow import RunEvent, RunResponse, Workflow from pydantic import BaseModel, Field load_dotenv() class DeepResearcherAgent(Workflow): """ A multi-stage research workflow that: 1. Gathers information from the web using advanced scraping tools. 2. Analyzes and synthesizes the findings. 3. Produces a clear, well-structured report. """ # Searcher: Finds and extracts relevant information from the web searcher: Agent = Agent( tools=[ScrapeGraphTools(api_key=os.getenv("SGAI_API_KEY"))], model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), show_tool_calls=True, markdown=True, description=( "You are ResearchBot-X, an expert at finding and extracting high-quality, " "up-to-date information from the web. Your job is to gather comprehensive, " "reliable, and diverse sources on the given topic." ), instructions=( "1. Search for the most recent and authoritative and up-to-date sources (news, blogs, official docs, research papers, forums, etc.) on the topic.\n" "2. Extract key facts, statistics, and expert opinions.\n" "3. Cover multiple perspectives and highlight any disagreements or controversies.\n" "4. Include relevant statistics, data, and expert opinions where possible.\n" "5. Organize your findings in a clear, structured format (e.g., markdown table or sections by source type).\n" "6. If the topic is ambiguous, clarify with the user before proceeding.\n" "7. Be as comprehensive and verbose as possible—err on the side of including more detail.\n" "8. Mention the References & Sources of the Content. (It's Must)" ), ) # Analyst: Synthesizes and interprets the research findings analyst: Agent = Agent( model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), markdown=True, description=( "You are AnalystBot-X, a critical thinker who synthesizes research findings " "into actionable insights. Your job is to analyze, compare, and interpret the " "information provided by the researcher." ), instructions=( "1. Identify key themes, trends, and contradictions in the research.\n" "2. Highlight the most important findings and their implications.\n" "3. Suggest areas for further investigation if gaps are found.\n" "4. Present your analysis in a structured, easy-to-read format.\n" "5. Extract and list ONLY the reference links or sources that were ACTUALLY found and provided by the researcher in their findings. Do NOT create, invent, or hallucinate any links.\n" "6. If no links were provided by the researcher, do not include a References section.\n" "7. Don't add hallucinations or make up information. Use ONLY the links that were explicitly passed to you by the researcher.\n" "8. Verify that each link you include was actually present in the researcher's findings before listing it.\n" "9. If there's no Link found from the previous agent then just say, No reference Found." ), ) # Writer: Produces a final, polished report writer: Agent = Agent( model=Nebius( id="deepseek-ai/DeepSeek-V3-0324", api_key=os.getenv("NEBIUS_API_KEY") ), markdown=True, description=( "You are WriterBot-X, a professional technical writer. Your job is to craft " "a clear, engaging, and well-structured report based on the analyst's summary." ), instructions=( "1. Write an engaging introduction that sets the context.\n" "2. Organize the main findings into logical sections with headings.\n" "3. Use bullet points, tables, or lists for clarity where appropriate.\n" "4. Conclude with a summary and actionable recommendations.\n" "5. Include a References & Sources section ONLY if the analyst provided actual links from their analysis.\n" "6. Use ONLY the reference links that were explicitly provided by the analyst in their analysis. Do NOT create, invent, or hallucinate any links.\n" "7. If the analyst provided links, format them as clickable markdown links in the References section.\n" "8. If no links were provided by the analyst, do not include a References section at all.\n" "9. Never add fake or made-up links - only use links that were actually found and passed through the research chain." ), ) def run(self, topic: str) -> Iterator[RunResponse]: """ Orchestrates the research, analysis, and report writing process for a given topic. """ logger.info(f"Running deep researcher agent for topic: {topic}") # Step 1: Research research_content = self.searcher.run(topic) # logger.info(f"Searcher content: {research_content.content}") logger.info("Analysis started") # Step 2: Analysis analysis = self.analyst.run(research_content.content) # logger.info(f"Analyst analysis: {analysis.content}") logger.info("Report Writing Started") # Step 3: Report Writing report = self.writer.run(analysis.content, stream=True) yield from report def run_research(query: str) -> str: agent = DeepResearcherAgent() final_report_iterator = agent.run( topic=query, ) logger.info("Report Generated") # Collect all streaming content into a single string full_report = "" for chunk in final_report_iterator: if chunk.content: full_report += chunk.content return full_report if __name__ == "__main__": topic = "Extract information about Nebius AI Studio, including its features, capabilities, and applications from available sources." response = run_research(topic) print(response) \`\`\` ## Prerequisites * Python 3.10+ * API keys for [Nebius AI Studio](https://studio.nebius.com/?modals=create-api-key) and [Scrapegraph AI](https://scrapegraphai.com/) ## Usage \`\`\`bash pip install agno openai scrapegraph-py \`\`\` \`\`\`bash python deep_researcher_agent.py \`\`\` [V0_FILE]markdown:file="agno/misc/agentic_filtering_with_docx.md" --- title: Agentic filtering with Docx category: misc source_lines: 13432-13450 line_count: 18 --- # Agentic filtering with Docx Source: https://docs.agno.com/examples/concepts/knowledge/filters/docx/agentic_filtering Learn how to do agentic knowledge filtering using Docx documents with user-specific metadata. ## Code \`\`\`python from pathlib import Path from agno.agent import Agent from agno.knowledge.docx import DocxKnowledgeBase from agno.utils.media import ( SampleDataFileExtension, download_knowledge_filters_sample_data, ) from agno.vectordb.lancedb import LanceDb [V0_FILE]markdown:file="agno/misc/agentic_filtering_with_json.md" --- title: Agentic filtering with Json category: misc source_lines: 15047-15065 line_count: 18 --- # Agentic filtering with Json Source: https://docs.agno.com/examples/concepts/knowledge/filters/json/agentic_filtering Learn how to do agentic knowledge filtering using Json documents with user-specific metadata. ## Code \`\`\`python from pathlib import Path from agno.agent import Agent from agno.knowledge.json import JSONKnowledgeBase from agno.utils.media import ( SampleDataFileExtension, download_knowledge_filters_sample_data, ) from agno.vectordb.lancedb import LanceDb [V0_FILE]markdown:file="agno/misc/agentic_filtering_with_pdf-url.md" --- title: Agentic filtering with Pdf-Url category: misc source_lines: 15792-15804 line_count: 12 --- # Agentic filtering with Pdf-Url Source: https://docs.agno.com/examples/concepts/knowledge/filters/pdf_url/agentic_filtering Learn how to do agentic knowledge filtering using Pdf-Url documents with user-specific metadata. ## Code \`\`\`python from agno.agent import Agent from agno.knowledge.pdf_url import PDFUrlKnowledgeBase from agno.vectordb.lancedb import LanceDb [V0_FILE]markdown:file="agno/misc/agentic_filtering_with_pdf.md" --- title: Agentic filtering with Pdf category: misc source_lines: 15443-15461 line_count: 18 --- # Agentic filtering with Pdf Source: https://docs.agno.com/examples/concepts/knowledge/filters/pdf/agentic_filtering Learn how to do agentic knowledge filtering using Pdf documents with user-specific metadata. ## Code \`\`\`python from pathlib import Path from agno.agent import Agent from agno.knowledge.pdf import PDFKnowledgeBase from agno.utils.media import ( SampleDataFileExtension, download_knowledge_filters_sample_data, ) from agno.vectordb.lancedb import LanceDb [V0_FILE]markdown:file="agno/misc/agentic_filtering_with_text.md" --- title: Agentic filtering with Text category: misc source_lines: 16093-16111 line_count: 18 --- # Agentic filtering with Text Source: https://docs.agno.com/examples/concepts/knowledge/filters/text/agentic_filtering Learn how to do agentic knowledge filtering using Text documents with user-specific metadata. ## Code \`\`\`python from pathlib import Path from agno.agent import Agent from agno.knowledge.text import TextKnowledgeBase from agno.utils.media import ( SampleDataFileExtension, download_knowledge_filters_sample_data, ) from agno.vectordb.lancedb import LanceDb [V0_FILE]markdown:file="agno/misc/agentic_rag.md" --- title: Agentic RAG category: misc source_lines: 49607-49718 line_count: 111 --- # Agentic RAG Source: https://docs.agno.com/examples/streamlit/agentic-rag This example application shows how to build a sophisticated RAG (Retrieval Augmented Generation) system that leverages search of a knowledge base with LLMs to provide deep insights into the data. ## The agent can: * Process and understand documents from multiple sources (PDFs, websites, text files) * Build a searchable knowledge base using vector embeddings * Maintain conversation context and memory across sessions * Provide relevant citations and sources for its responses * Generate summaries and extract key insights * Answer follow-up questions and clarifications ## The agent uses: * Vector similarity search for relevant document retrieval * Conversation memory for contextual responses * Citation tracking for source attribution * Dynamic knowledge base updates