MCPDoc

Provides documentation sources and a fetch_docs tool to audit llms.txt-based context for MCP-enabled clients.

typescript

GitHub Stars

typescript

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "swesmith-repos-langchain-ai__mcpdoc.538df6d0": {
      "command": "uvx",
      "args": [
        "--from",
        "mcpdoc",
        "mcpdoc",
        "--urls",
        "LangGraph:https://langchain-ai.github.io/langgraph/llms.txt LangChain:https://python.langchain.com/llms.txt",
        "--transport",
        "stdio"
      ]
    }
  }
}

You deploy an MCP server to give host applications like Cursor, Windsurf, and Claude access to a user-defined set of llms.txt sources and a simple fetch_docs tool. This lets you audit tool calls, control which domains are accessed, and fetch documentation from trusted sources on demand.

How to use

Set up an MCP server to provide your selected llms.txt sources and the fetch_docs tool to client applications. Start the server locally, verify it runs, then connect your MCP client (Cursor, Windsurf, or Claude Desktop/Code) to use the tool flow for querying languages model document sources.

How to install

Prerequisites: you need a runtime to execute MCP tooling. Install the MCP runtime (uv) and prepare to run the MCP server via stdio I/O or HTTP as configured.

Additional options and usage notes

The MCP server enables you to specify documentation sources and to run the fetch_docs tool against those sources. You can test tools locally, connect client apps, and tune security by explicitly allowing domains when using local files.

Tools and endpoints

{
  "mcpServers": {
    "langgraph-docs-mcp": {
      "command": "uvx",
      "args": [
        "--from",
        "mcpdoc",
        "mcpdoc",
        "--urls",
        "LangGraph:https://langchain-ai.github.io/langgraph/llms.txt LangChain:https://python.langchain.com/llms.txt",
        "--transport",
        "stdio"
      ]
    }
  }
}

Connecting MCP to Cursor, Windsurf, and Claude Desktop

Follow these steps to connect your MCP server to client applications and enable tool-assisted prompts.

Examples to try

Try a sample prompt after connecting the tools to see how it fetches and reflects on llms.txt sources.

Available tools

list_doc_sources

Tool to enumerate available llms.txt sources configured for the MCP server and expose their names and URLs.

fetch_docs

Tool to read and retrieve documentation from the specified llms.txt URLs and related sources for auditing and context.

Built by

VeilStrat

AI signals for GTM teams