🤖 LangGraph Multi-Agent Chatbot

A production-grade Multi-Agent AI Chatbot built with LangGraph — featuring intelligent routing, RAG pipeline, real-time web search, Python code execution via MCP, and persistent memory backed by PostgreSQL.

⭐ Star this repo if it helps you!

---

✨ Features

Feature	Description
🔀 Smart Router	Classifies every query and sends it to the right agent automatically
🐍 Python Executor	Runs Python code in real-time via a remote FastMCP server
🔍 Web Search	Google search with full page content extraction (Apify)
📄 RAG Pipeline	Document Q&A using Pinecone Vector DB + BM25 hybrid retrieval
🧠 Persistent Memory	Multi-thread conversation memory backed by PostgreSQL (Neon)
📊 Observability	Full tracing via LangSmith
🌐 Bilingual	Responds in Roman Urdu or English based on user language

---

🏗️ Architecture

User Query
    │
    ▼
┌─────────────┐
│ router_node │  Classifies → python_tool / web_search / rag / direct
└──────┬──────┘
       │
  ┌────┴──────────┬──────────────┐
  ▼               ▼              ▼
python_tool    web_search      rag
(FastMCP)      (Apify)      (Pinecone)
       │
       ▼
  answer_node → END

---

📁 Project Structure

├── langgraph_agent.py   # Main agent — graph, nodes, tools
├── config.py            # All settings loaded from .env
├── requirements.txt     # Dependencies
├── .env.example         # Environment variables template
├── docs/
│   ├── architecture.md  # Deep-dive architecture docs
│   └── api_reference.md # API reference
├── tests/
│   └── test_agent.py    # pytest test suite
└── README.md

---

🚀 Quick Start

1. Clone & install

git clone https://github.com/Pokemon455/langgraph-multi-agent.git
cd langgraph-multi-agent
pip install -r requirements.txt

2. Set up environment

cp .env.example .env
# Open .env and fill in your API keys

3. Run

import asyncio
from langgraph_agent import build_and_run

result = asyncio.run(build_and_run("python version check karo", thread_id="1"))
print(result)

---

🔧 Required Environment Variables

Variable	Description	Get it from
NVIDIA_API_KEY	LLM inference	build.nvidia.com
PINECONE_API_KEY	Vector database	pinecone.io
DATABASE_URL	PostgreSQL memory	neon.tech
APIFY_API_TOKEN	Web search	apify.com
LANGCHAIN_API_KEY	Tracing (optional)	smith.langchain.com
MCP_SERVER_URL	FastMCP Python REPL	fastmcp-python-repl-server

---

🔗 Related Projects

• fastmcp-python-repl-server — The MCP Python REPL server used by this agent
• langchain-ai/langgraph — The graph framework powering this project
• langchain-mcp-adapters — MCP integration for LangChain

---

🤝 Contributing

Contributions, issues and feature requests are welcome! See CONTRIBUTING.md for guidelines.

---

📄 License

MIT © Arbaz