mcp-docs-server

dgmulei/mcp-docs-server

3.2

If you are the rightful owner of mcp-docs-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

The MCP Documentation Server provides structured access to Model Context Protocol documentation, examples, and troubleshooting guides for AI coding assistants.

Tools
5
Resources
0
Prompts
0

MCP Documentation Server

An MCP server providing structured access to Model Context Protocol documentation, examples, and troubleshooting guides for AI coding assistants.

Purpose

This server bridges the gap between MCP specifications and working implementations by providing AI coding assistants with:

  • Real-time access to complete MCP documentation
  • Working code examples from proven implementations
  • Troubleshooting guides for common issues
  • Validation tools for compliance checking

Tools

  • search_mcp_docs - Search across all documentation
  • get_transport_examples - SSE vs Streamable HTTP examples
  • troubleshoot_connection - Debug common connection failures
  • validate_oauth_flow - Check OAuth implementation compliance
  • get_working_templates - Get proven server templates

Setup

  1. Install dependencies:
npm install
  1. Prepare documentation:
npm run prepare-docs
  1. Build and run:
npm run build
npm start
  1. For development:
npm run dev

Configuration

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "mcp-docs": {
      "command": "node",
      "args": ["/path/to/mcp-docs-server/dist/index.js"]
    }
  }
}

Data Structure

Documentation is organized in the data/ directory:

  • specifications/ - Official MCP specs
  • implementations/ - Working code examples
  • troubleshooting/ - Problem-solution mappings
  • examples/ - Templates and guides

Testing

Test with MCP Inspector:

npx @modelcontextprotocol/inspector node dist/index.js

Updating Documentation

1. Monitor Documentation Sources

Set up monitoring for MCP documentation updates:

# Official specification repository
git clone https://github.com/modelcontextprotocol/specification.git

# TypeScript SDK examples
git clone https://github.com/modelcontextprotocol/typescript-sdk.git

# Python SDK examples  
git clone https://github.com/modelcontextprotocol/python-sdk.git

# Check for updates periodically
cd specification && git pull
cd typescript-sdk && git pull
cd python-sdk && git pull

2. Adding New Documentation

# Copy new files to appropriate directories
cp new-specification.md data/specifications/
cp new-example.md data/examples/
cp troubleshooting-guide.md data/troubleshooting/

# Restart server to load new content
npm run build
npm start

3. File Format

Use frontmatter for better categorization:

---
title: "Document Title"
tags: ["oauth", "transport", "debugging"]
category: "specification"
---

# Document Content

Categories: specification, implementation, troubleshooting, examples

4. Automated Updates (Future Enhancement)

Consider implementing:

  • GitHub webhooks to monitor spec repository changes
  • Scheduled jobs to fetch latest SDK examples
  • Version detection for schema.ts changes
  • Automated categorization based on content analysis

SDK Integration Strategy

Current Workflow

  1. Use this server first for understanding MCP requirements and debugging
  2. Reference SDK repositories for specific implementation patterns:
  3. Return to this server for troubleshooting and validation

When to Use Each Resource

This Documentation Server:

  • Understanding Claude's specific requirements (SSE transport, OAuth 2.1, Dynamic Client Registration)
  • Debugging connection failures ("Claude was unable to connect")
  • Validating OAuth flow implementation
  • Finding working templates (Cloudflare, SimpleScraper examples)
  • Learning MCP fundamentals and architecture

SDK Repositories:

  • Complete working server implementations
  • Transport layer code (SSEServerTransport, StreamableHTTPServerTransport)
  • Tool definition patterns and best practices
  • Authentication implementation details
  • Error handling and protocol compliance examples

Recommendation for AI Assistants

When helping users implement MCP servers:

  1. Start with requirements: Use search_mcp_docs and troubleshoot_connection tools from this server
  2. Get implementation patterns: Direct users to appropriate SDK examples for their language
  3. Debug issues: Use this server's troubleshooting tools and guides
  4. Validate compliance: Use validate_oauth_flow and other validation tools

Contributing

To add new documentation:

  1. Place markdown files in appropriate data/ subdirectory
  2. Use descriptive filenames and include frontmatter
  3. Run npm run prepare-docs to rebuild index (if script exists)
  4. Restart the server to load new files

Documentation will be automatically categorized and indexed for search and retrieval.

Architecture

The server automatically:

  • Loads all .md files from data/ subdirectories
  • Categorizes based on folder structure and frontmatter
  • Builds searchable index with Fuse.js
  • Provides structured access via MCP tools

File categorization logic:

  • /specifications/ or contains "authorization", "spec" → specification
  • /troubleshooting/ or contains "troubleshoot", "debug" → troubleshooting
  • /examples/ or contains "example", "cloudflare", "simplescraper" → examples
  • Default → implementation

Development Notes

  • Documentation is loaded on server startup (restart required for new files)
  • Search uses fuzzy matching across title, content, tags, and category
  • Tools provide structured access optimized for AI assistant consumption
  • Session management handles multiple concurrent documentation queries