🤖 Agentic RAG API

A powerful Retrieval-Augmented Generation (RAG) system built with CrewAI agents, featuring a REST API and a beautiful Gradio web interface. This project demonstrates how to create intelligent AI agents that can research queries, synthesize information, and provide comprehensive responses.

🌟 Features

• Multi-Agent Architecture: Researcher and Writer agents working together
• Web Search Integration: Powered by SerperDev tools for real-time information
• REST API: Clean API endpoints using LitServe
• Beautiful Web UI: Interactive Gradio interface
• Command Line Client: Simple CLI for testing and automation
• Production Ready: Proper error handling and logging

🔒 Privacy & Cost Benefits

This project leverages Qwen 3 4B via Ollama as a local language model, providing:

• 🔒 Complete Data Privacy - All inference happens on your machine
• 💰 Zero API Costs - No charges for LLM calls
• ⚡ Fast Responses - No network latency
• 🌐 Offline Capability - Works without internet (except web search)

🏗️ Architecture

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   Web Client    │    │   REST API       │    │   CrewAI Agents │
│   (Gradio UI)   │◄──►│   (LitServe)     │◄──►│   (Researcher   │
│                 │    │                  │    │    + Writer)    │
└─────────────────┘    └──────────────────┘    └─────────────────┘
                                │
                                ▼
                       ┌──────────────────┐
                       │   SerperDev      │
                       │   Web Search     │
                       └──────────────────┘

🤖 How the Agents Work

This project uses a multi-agent architecture powered by CrewAI, where specialized AI agents collaborate to provide intelligent responses:

Agent Roles and Workflow

1. Researcher Agent 🔍

   • Role: Information gathering and analysis
   • Tools: SerperDev web search integration
   • Goal: Research the user's query and generate comprehensive insights
   • Process:
     • Receives the user's query
     • Uses web search to find relevant, up-to-date information
     • Analyzes and synthesizes findings
     • Produces research insights and context

2. Writer Agent ✍️

   • Role: Response synthesis and communication
   • Goal: Transform research insights into clear, informative responses
   • Process:
     • Takes the researcher's insights as input
     • Crafts a concise, well-structured response
     • Ensures the response is accurate and easy to understand
     • Formats the final answer for the user

The Collaboration Process

User Query → Researcher Agent → Web Search → Analysis → Writer Agent → Final Response
     ↓              ↓                    ↓         ↓           ↓            ↓
   "What is     Researches &         Searches   Synthesizes  Writes      "Machine
   machine      gathers info          current    research    clear       learning
   learning?"   from web              data       results     response"   is..."

Key Features

• Sequential Processing: Agents work in sequence, each building on the previous agent's work
• Tool Integration: Real-time web search for current information
• Context Preservation: Information flows from researcher to writer seamlessly
• Quality Assurance: Each agent has specialized expertise for their role
• Error Handling: Robust error handling ensures reliable responses

Technical Implementation

• Framework: CrewAI for agent orchestration
• Language Model: Ollama Qwen 3 4B for intelligent responses
• Tools: SerperDev for web search capabilities
• API: LitServe for production-ready serving
• UI: Gradio for beautiful web interface

🚀 Quick Start

Prerequisites

• Python 3.11 or higher
• uv package manager (recommended)
• Internet connection for web search functionality

Installation

1. Clone the repository

   bash Copy

   git clone 
   cd deploy-agentic-rag

2. Install dependencies

   bash Copy

   uv sync

3. Set up environment variables

   bash Copy

   # Copy the example environment file
   cp .env.example .env
   
   # Edit .env and add your actual API key
   # Get your SerperDev API key from: https://serper.dev
   SERPER_API_KEY="your-actual-serper-api-key-here"

Run the Application

1. Start the API server

   bash Copy

   uv run python server.py

   The API will be available at http://localhost:8000

2. Start the web interface (in a new terminal)

   bash Copy

   uv run python gradio_ui.py

   The web UI will be available at http://localhost:7860

3. Test with the CLI client

   bash Copy

   uv run python client.py --query "What is machine learning?"

📱 Screenshots

Agentic RAG Interface showcasing intelligent AI responses to complex queries about AI breakthroughs

Usage

Web Interface

1. Open http://localhost:7860 in your browser
2. Enter your query in the text box
3. Click "Submit Query" or press Enter
4. View the intelligent response generated by the AI agents

REST API

Endpoint: POST /predict

Request:

{
  "query": "Explain quantum computing in simple terms"
}

Response:

{
  "output": {
    "raw": "Quantum computing is a type of computing that uses quantum mechanics...",
    "tasks_output": [...],
    "token_usage": {...}
  }
}

Command Line

# Simple query
uv run python client.py --query "What is the capital of France?"

# The client displays responses with a typewriter effect

🔧 Configuration

API Configuration

• Server URL: http://localhost:8000
• Timeout: 60 seconds
• Model: Ollama Qwen 3 4B (configurable in server.py)

Agent Configuration

Researcher Agent:

• Role: Research and gather information
• Tools: SerperDev web search
• Goal: Generate comprehensive insights

Writer Agent:

• Role: Synthesize information into clear responses
• Goal: Provide concise, informative answers

📁 Project Structure

deploy-agentic-rag/
├── server.py           # Main API server using LitServe
├── client.py           # Command-line client
├── gradio_ui.py        # Web interface
├── pyproject.toml      # Project configuration
└── README.md          # This file

🛠️ Development

Adding New Features

1. Custom Agents: Modify server.py to add new agent types
2. Additional Tools: Extend the Researcher agent with more tools
3. UI Enhancements: Customize gradio_ui.py for new features

Testing

# Run tests
uv run pytest

# Code formatting
uv run black .
uv run flake8 .

🌐 API Reference

Health Check

• GET /health - Check server status

Predictions

• POST /predict - Submit queries for processing

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• CrewAI for the amazing agent framework
• LitServe for the lightweight API serving
• Gradio for the beautiful web interface
• SerperDev for web search capabilities

Made with ❤️ using CrewAI Agents

🚀 Quick Start

git clone 
   cd deploy-agentic-rag