MCP Tool Langgraph Integration

Description

Example project of how to integrate MCP endpoint tools into a Langgraph tool node

The graph consists of only 2 nodes, agent and tool.

Prerequisites

To use this project, make sure you have Python 3.11.

uv is recommended

Linux and Mac

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

Windows

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

MCP Server requirements

• This example uses the MCP Server sample @modelcontextprotocol/server-brave-search to add Brave Search tools. This requires that you have node and npx installed.

API Keys

• The MCP Server sample used is for Brave Search, you can get a free API key from https://brave.com/search/api/
• You will need and API key for the chosen AI provider which defaults to Anthropic but can be changed by editing the __main__.py file
• Put all api keys in a .env file in the repository root.

From source Usage

uv run mcp_langgraph_tools

Multiple MCP servers at one time

Check the multi_server branch for a more advanced example of how to use multiple MCP servers at once.

Whats New

• Version 0.1.0:
  • Initial release

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

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

Author

Paul Robello - probello@gmail.com