Multi-Agent RAG & Web Search System 🤖🔍

A powerful multi-agent framework designed to handle complex queries by combining Retrieval-Augmented Generation (RAG) with real-time Web Search capabilities. This system uses specialized agents to decide when to look at local documents and when to fetch live information from the internet.

RAG capability

Web search capability

🚀 Features

• Multi-Agent Orchestration: Uses a supervisor or autonomous agents to delegate tasks.
• RAG Integration: Efficiently retrieves context from your local knowledge base (PDFs, Text, etc.).
• Real-time Web Search: Integrates with tools (Tavily) to provide up-to-date answers.
• Smart Routing: Automatically determines whether a query requires local data or a web search.
• Persistent Memory: Remembers previous interactions in a conversation.

🛠️ Tech Stack

• Framework: [LangGraph, LangChain]
• LLM: [openai/gpt-oss-20b from Groq API]
• Vector Database: [ChromaDB]
• Search Tool: [Tavily API]
• UI: [Streamlit]

📋 Prerequisites

Before running the project, ensure you have:

• Python 3.11
• API Keys for:
  • Groq
  • Tavily

This project uses uv as the python package manager

Install uv:

curl -LsSf https://astral.sh/uv/install.sh | sh

Verify the installation:

echo 'export PATH="$HOME/snap/code/221/.local/bin:$PATH"' >> ~/.bashrc && source ~/.bashrc

uv --version

Clone the repository or Download as zip file

https://github.com/yoursrealkiran/multi_agent.git

cd multi_agent

Environment Setup

Create a Virual Environment

In the terminal, run the below command to create a virtual environment

uv venv

Activate the environment

source .venv/bin/activate

Install Dependencies

uv sync

Running the Application

cd multi_agent/src

uv run streamlit run main.py