MCP Documentation Server

Integrates LLM applications with documentation sources using the Model Context Protocol.

GitHub

MCP Documentation Server

A customized version of the MCP documentation server that enables integration between LLM applications (like Cursor, Claude Desktop, Windsurf) and documentation sources via the Model Context Protocol.

Overview

This server provides MCP host applications with:

  1. Access to specific documentation files (langgraph.txt and mcp.txt)
  2. Tools to fetch documentation from URLs within those files

Supported Documentation

Currently set up for:

Quick Start

Setup and Run

# Clone the repository
git clone https://github.com/esakrissa/mcp-doc.git
cd mcp-doc

# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install the package in development mode
pip install -e .

Running the Server

You can run the server using the installed command:

# Run the server with the config file
mcpdoc \
    --json config.json \
    --transport sse \
    --port 8082 \
    --host localhost

Or if you prefer using UV:

# Install uv (if not already installed)
curl -LsSf https://astral.sh/uv/install.sh | sh

# Run the server with UV
uvx --from mcpdoc mcpdoc \
    --json config.json \
    --transport sse \
    --port 8082 \
    --host localhost

IDE Integration

Cursor

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "mcp-doc": {
      "command": "uvx",
      "args": [
        "--from",
        "mcpdoc",
        "mcpdoc",
        "--urls",
        "LangGraph:https://raw.githubusercontent.com/esakrissa/mcp-doc/main/docs/langgraph.txt",
        "ModelContextProtocol:https://raw.githubusercontent.com/esakrissa/mcp-doc/main/docs/mcp.txt",
        "--allowed-domains",
        "*",
        "--transport",
        "stdio"
      ]
    }
  }
}

Then add these instructions to Cursor's Custom Instructions:

for ANY question about LangGraph and Model Context Protocol (MCP), use the mcp-doc server to help answer -- 
+ call list_doc_sources tool to get the available documentation files
+ call fetch_docs tool to read the langgraph.txt or mcp.txt file
+ reflect on the urls in langgraph.txt or mcp.txt 
+ reflect on the input question 
+ call fetch_docs on any urls relevant to the question
+ use this to answer the question

To test if the integration is working, ask Cursor a question about LangGraph or MCP, and check if it uses the documentation server tools to fetch information.

Security Note

For security reasons, strict domain access controls are implemented:

  • Remote documentation files: Only the specific domain is automatically allowed
  • Local documentation files: No domains are automatically allowed
  • Use --allowed-domains to explicitly add domains or --allowed-domains '*' to allow all (use with caution)

References

This project is based on the original mcpdoc by LangChain AI, modified to provide focused documentation access for LangGraph and MCP.

Máy chủ liên quan