DocsMCP

What is DocsMCP about?

DocsMCP enables Large Language Models to retrieve and query documentation from configured sources, supporting both remote URLs and local file paths.

How to use DocsMCP?

• Cursor: Add a .cursor/mcp.json file with a server entry that runs npx docsmcp and specifies a --source argument.
• VS Code: Add a .vscode/mcp.json configuration defining a stdio server that runs npx docsmcp with the appropriate --source flag.
• CLI: Invoke the server directly with npx docsmcp --source=Name|URL_or_path.

Key Features

• getDocumentationSources: Lists all configured documentation sources.
• getDocumentation: Fetches and parses documentation from a given URL or local file.
• Simple configuration via JSON files for popular editors (Cursor, VS Code).
• Works out‑of‑the‑box with npx, no global installation required.

Use Cases

• Supplying LLMs with up‑to‑date API reference material during code generation.
• Allowing AI assistants to answer product‑specific questions by querying internal docs.
• Providing a uniform interface for multiple documentation sources across projects.

FAQ

Q: Do I need to install anything globally? A: No. The server runs with npx, which downloads and executes the package on demand.

Q: How are spaces handled in source definitions? A: Wrap the entire --source string in single quotes, e.g., '--source=Model Context Protocol (MCP)|https://example.com/doc.txt'.

Q: Can I add multiple sources? A: Yes. Include additional --source arguments when launching the server.

Q: What file formats are supported? A: Any text‑based format that can be fetched via HTTP(S) or read from the filesystem.

Q: Is there a license? A: The project is released under the MIT license.

DocsMCP

A Model Context Protocol (MCP) server that provides documentation access to LLMs.

Overview

DocsMCP enables Large Language Models (LLMs) to access and query documentation from specified sources, whether from local files or remote URLs. It uses the Model Context Protocol (MCP) to facilitate communication between the LLM and documentation sources.

Usage

Cursor MCP Configuration

You can also configure DocsMCP in your Cursor project by creating a .cursor/mcp.json file:

{
  "mcpServers": {
    "docs-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "docsmcp",
        "'--source=Model Context Protocol (MCP)|https://modelcontextprotocol.io/llms-full.txt'"
      ]
    }
  }
}

This configuration allows Cursor AI to use the documentation MCP server automatically when you open your project.

Note:

When specifying a source that contains spaces, ensure to wrap the entire string in quotes. For example: '--source=Model Context Protocol (MCP)|https://modelcontextprotocol.io/llms-full.txt'

VS Code MCP Configuration

You can configure DocsMCP in VS Code by adding a configuration to your .vscode/mcp.json file:

{
  "servers": {
    "documentation-mcp-server": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "docsmcp",
        "--source=Model Context Protocol (MCP)|https://modelcontextprotocol.io/llms-full.txt"
      ]
    }
  }
}

This configuration allows VS Code extensions that support MCP to use the documentation server automatically.

Available Tools

The MCP server provides two main tools:

getDocumentationSources

Lists all available documentation sources that have been configured.

getDocumentation

Fetches and parses documentation from a given URL or local file path.

Parameters:

• url: The URL or file path to fetch the documentation from

License

MIT