langchain

This MCP server exposes LangChain-compatible tool calling through the Model Context Protocol. It uses a LangChain MCP Toolkit that integrates with an mcp.ClientSession. After initializing the toolkit, you can call toolkit.get_tools() to retrieve the list of LangChain tools (BaseTool instances) available for the current session. Use these tools to perform schema-aware, model-context-driven actions within your application. The included demo shows how to execute a LangChain-based workflow against a Groq-backed environment by running a Python script through uv, which demonstrates file-reading and summarization capabilities via MCP-tooling integration.

Prerequisites:

• Python 3.8+ installed on your system
• uv (the Python execution runner used in the example) installed and available in PATH
• Internet access to install Python packages

Installation steps:

1. Install the MCP package from PyPI (or install from source if you have a clone):

   • pip install langchain-mcp

2. (Optional) If you want to run the included demo exactly as shown, ensure you have a Groq API key for testing. Set it in your environment:

   • export GROQ_API_KEY=your_key_here

3. Run the demo using uv (as shown in the README example):

   • uv run tests/demo.py "Read and summarize the file ./LICENSE"

4. If you’re developing locally, install dependencies and run the script directly with Python as an alternative:

   • python -m pip install -r requirements.txt # if a requirements file is provided
   • python tests/demo.py "Read and summarize the file ./LICENSE"

Notes and tips:

• The demo relies on GROQ for the data source. You can either provide a valid GROQ_API_KEY or adapt the demo to a different data source.
• The MCP setup shown uses the uv runner to execute a Python script that exercises the LangChain MCP toolkit; ensure your environment has access to the script path used in the command.
• If you encounter issues with tool discovery, verify that the LangChain MCP Toolkit is correctly installed and that the Python path includes the project modules.
• Environment variables used in the example are placeholder suggestions; replace or extend with variables relevant to your environment (e.g., API keys, endpoint URLs).
• This package is aligned with the LangChain MCP ecosystem; for newer adapters or official implementations, consider alternative packages such as langchain-mcp-adapters.