This repository implements a multi-Agents AI system using FastAPI, LangGraph, and Groq LLM.
It supports dynamic routing between agents for tasks such as web search and summarization, making it modular, scalable, and production-ready.

• ✅ Multi-Agent Orchestration with LangGraph
• ✅ Web Search Agent powered by Tavily
• ✅ Summarizer Agent using Groq LLM
• ✅ Router Agent that decides workflow dynamically
• ✅ FastAPI REST API with /multi-agents endpoint
• ✅ Dockerized Deployment with docker-compose
• ✅ Modular project structure for scalability

Router Agent: Decides whether to search or summarize.

Search Agent: Performs real-time web search using Tavily.

Summarizer Agent: Summarizes text with Groq LLM.

Workflow is managed with a LangGraph state machine.

├── endpoints/
│ └── endpoint.py # FastAPI route for multi-agent queries
├── schemas/
│ └── schema.py # Pydantic request schema
├── services/
│ ├── agents.py # Agent definitions (search, summarizer, router)
│ ├── graph.py # Build LangGraph state machine
│ ├── model.py # Groq LLM integration
│ └── tools.py # Web search & summarization tools
├── main.py # FastAPI entry point
├── requirements.txt # Python dependencies
├── Dockerfile # Docker image build
├── docker-compose.yml # Multi-service deployment
└── README.md # Project documentation

git clone https://github.com/imran-sony/multi-agent.git
cd multi-agent

python -m venv env
pip install -r requirements.txt

Create a .env file in the project root:

GROQ_API_KEY=your_groq_api_key
TAVILY_API_KEY=your_tavily_api_key

python main.py