MCP Documentation Server 📚

A production-ready MCP (Model Context Protocol) server that provides intelligent documentation access for AI coding assistants. Seamlessly integrates with Claude and GitHub Copilot to provide context-aware assistance.

📋 Features

• Local Documentation Index: Indexes docs from multiple git repositories
• Semantic Search: Finds relevant context based on meaning, not just keywords
• AI Assistant Integration: Works with Claude, GitHub Copilot, and other MCP-compatible tools
• Automatic Repository Sync: Keeps documentation up-to-date
• Vector-based Search: Fast and accurate retrieval using Qdrant
• Low Resource Usage: Efficient local processing with minimal overhead
• Secure: Input validation, path sanitization, authentication options
• Production Ready: Robust error handling, logging, and monitoring

🚀 Quick Start

# Clone the repository
git clone https://github.com/username/mcp-server-docs-lookup.git
cd mcp-server-docs-lookup

# Install dependencies
npm install

# Build the project
npm run build

# Configure repositories (edit config/config.json)
cp config/config.example.json config/config.json

# Start the server
npm run api

The server will be running at http://localhost:3001. Access the health endpoint:

curl http://localhost:3001/health

Documentation

• INTEGRATION.md - How to integrate with AI assistants
• ARCHITECTURE.md - Architecture overview
• TROUBLESHOOTING.md - Common issues and solutions
• OLYMPUS_CLOUD.md - Olympus Cloud specific information
• CONTRIBUTING.md - Contribution guidelines

Prerequisites

• Node.js 18+ - Required for running the MCP server
• Docker - Required for Qdrant vector database
• Git - Required for cloning documentation repositories
• 8GB+ RAM - Recommended for processing large documentation sets

How it Works

1. Clone & Index: The server clones documentation repositories to your local machine
2. Process: Documents are chunked and converted to vector embeddings
3. Store: Embeddings are stored in a local Qdrant instance
4. Serve: When your AI assistant needs context, it queries the MCP server
5. Return: Relevant documentation chunks are returned to enhance AI responses

Configuration

Configuration is stored in config/config.json in your project directory:

{
  "project": {
    "name": "my-project",
    "description": "Project documentation"
  },
  "repositories": [
    {
      "name": "react-docs",
      "url": "https://github.com/reactjs/react.dev.git",
      "branch": "main",
      "paths": ["/src/content/learn", "/src/content/reference"],
      "category": "framework",
      "priority": "high",
      "syncInterval": 60
    }
  ],
  "contextGeneration": {
    "strategies": ["hybrid"],
    "maxChunks": 20
  },
  "vectorStore": {
    "type": "qdrant",
    "qdrant": {
      "url": "http://localhost:6333",
      "collectionName": "documentation"
    }
  }
}

Integration with AI Coding Assistants

For complete integration instructions with different AI assistants, refer to INTEGRATION.md.

Quick Integration Commands

# For Claude integration
npm run integrate:claude

# For GitHub Copilot integration
npm run integrate:copilot

# For VS Code integration
npm run integrate:vscode

Troubleshooting

For common issues and solutions, refer to TROUBLESHOOTING.md.

Common troubleshooting commands:

# Test connection
npm run tools:search

# Fix embedding issues
npm run tools:fix-embeddings

Development

# Clone repository
git clone https://github.com/username/mcp-server-docs-lookup.git
cd mcp-server-docs-lookup

# Install dependencies
npm install

# Start Qdrant locally
docker run -p 6333:6333 qdrant/qdrant

# Run in development
npm run dev

# Run tests
npm test

# Build for production
npm run build

License

MIT - See LICENSE file