MCP + LangGraph Agent

by hirokiyn

399 downloads
Not rated
GitHub

About

LangGraph agent tutorial adapted to use MCP servers.

Details

Author
hirokiyn
Downloads
399
Categories
Other

- Integrates LangGraph for managing agent state and message routing.
- Uses MCP servers to provide access to tools.
- Includes example MCP servers for math and weather.
- Provides a command-line interface for interacting with the agent.

Clone the repository, install dependencies using Poetry, and set the Anthropic API key in the .env file. Start the MCP servers (python src/mcp_servers/math_server.py and python src/mcp_servers/weather_server.py), then run poetry run main to launch the agent in interactive mode.

MCP + LangGraph Agent

This is a minimal, functional example of an agent powered by LangGraph with tools implemented using MCP (Model Context Protocol) servers instead of traditional LangChain tools. It's based on the official LangGraph tutorial, adapted to demonstrate how to integrate MCP servers as tools.

> 🧠 Use this repo as a skeleton to quickly build your own LangGraph agent with MCP tools!

Features

flowchart TD
    __start__ --> chatbot
    chatbot -.-> tools
    chatbot -.-> __end__
    tools --> chatbot

Integrates LangGraph for managing agent state and message routing.
Uses MCP servers to provide access to tools.
Includes example MCP servers for math and weather.
Provides a command-line interface for interacting with the agent.

Installation

1. Clone the repository:

    git clone <repository_url>
    
2. Install the dependencies using Poetry:
    poetry install
    

Usage

1. Set the Anthropic API key in the .env file. You may need to create this file if it doesn't exist. For example:

    ANTHROPIC_API_KEY=your_api_key
    
2. Start the MCP servers:

Math server: python src/mcp_servers/math_server.py
Weather server: python src/mcp_servers/weather_server.py
2. Run the agent:

    poetry run main
    

This will start the agent in interactive mode. You can then enter prompts, and the agent will respond using the tools provided by the MCP servers.

Example

User: What's (3 + 5) x 12?
Assistant: The result of (3 + 5) Γ— 12 = 96
User: What is the weather in New York?
Assistant: It's always sunny in New York.

MCP Server Configuration

The MCP servers are configured in src/main.py. You can modify the configuration to add or remove servers, or to change the transport mechanism.

LLM Configuration

The LLM used by the agent can be changed in src/common.py.

No reviews yet β€” be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.