open-docs-mcp

MCP.Pizza Chef: askme765cs

open-docs-mcp is an open-source MCP server that provides comprehensive document management capabilities including crawling, indexing, and full-text search across multiple document formats. It exposes a resource-based API and tool-based interfaces to manage documents efficiently within the MCP ecosystem, enabling real-time document access and search functionality.

Use This MCP server To

Crawl and index documentation from multiple sources automatically Perform full-text search across indexed documents Enable or disable document crawling dynamically Build and update search indexes for documentation Access and manage documents via MCP server API Integrate document search into AI workflows and agents

README

open-docs-mcp MCP Server

smithery badge License Node Version TypeScript

An open-source MCP implementation providing document management functionality. 中文文档

Features

Document Management

  • Crawl and index documentation from various sources
  • Support for multiple document formats
  • Full-text search capabilities

MCP Server API

  • Resource-based access to documents
  • Tool-based document management

Available Tools

  1. enable_doc - Enable crawling for a specific doc
  2. disable_doc - Disable crawling for a specific doc
  3. crawl_docs - Start crawling enabled docs
  4. build_index - Build search index for docs
  5. search_docs - Search documentation
  6. list_enabled_docs - List enabled docs
  7. list_all_docs - List all available docs

Cursor @Docs Compatibility

This project aims to replicate Cursor's @Docs functionality by providing:

  1. Document Indexing:

    • Crawl and index documentation from various sources
    • Support for multiple document formats (HTML, Markdown, etc.)
    • Automatic re-indexing to keep docs up-to-date
  2. Document Access:

    • Search across all indexed documentation
    • Integration with MCP protocol for AI context
  3. Custom Docs Management:

    • Add new documentation sources via enable_doc tool
    • Manage enabled docs via list_enabled_docs tool
    • Force re-crawl with crawl_docs tool

Architecture

┌───────────────────────────────────────────────────────┐
│                    open-docs-mcp Server                    │
├───────────────────┬───────────────────┬───────────────┤
│   Crawler Module  │  Search Engine    │  MCP Server   │
├───────────────────┼───────────────────┼───────────────┤
│ - Web crawling    │ - Full-text index │ - Resources   │
│ - Doc conversion  │ - Relevance score │ - Tools       │
│ - Storage         │ - Query parsing   │ - Prompts     │
└───────────────────┴───────────────────┴───────────────┘

Usage

npx -y open-docs-mcp --docsDir ./docs

Installing via Smithery

To install Document Management Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @askme765cs/open-docs-mcp --client claude

Configuration

To use with Claude Desktop, add the server config:

On MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json On Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "open-docs-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "open-docs-mcp",
        "--docsDir",
        "/path/to/docs"
      ]
    }
  }
}

Configuration Options:

  • command: Node.js executable
  • args: Array of arguments to pass to the script
    • --docsDir: Required, specifies docs directory path
  • disabled: Set to true to temporarily disable the server
  • alwaysAllow: Array of tool names that can be used without confirmation

Development

npm run watch  # Auto-rebuild on changes
npm run inspector  # Debug with MCP Inspector

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

MIT

open-docs-mcp FAQ

How do I enable document crawling for a specific document?
Use the 'enable_doc' tool provided by the server to activate crawling for that document.
Can I disable crawling for certain documents?
Yes, the 'disable_doc' tool allows you to stop crawling specific documents as needed.
What document formats does open-docs-mcp support?
It supports multiple common document formats, enabling broad document management and search capabilities.
How does the server handle search indexing?
The 'build_index' tool builds and updates the full-text search index for all enabled documents.
Is the open-docs-mcp server compatible with different LLM providers?
Yes, it is designed to work with various LLMs including OpenAI, Claude, and Gemini by exposing structured document data.
What Node.js version is required to run open-docs-mcp?
The server requires Node.js version 16.0.0 or higher.
Is open-docs-mcp open source and under what license?
Yes, it is open source and licensed under the MIT license.
How can I integrate open-docs-mcp into my existing MCP client?
You can connect via its resource-based API and use the provided tools to manage document workflows.