JSPM

  • ESM via JSPM
  • ES Module Entrypoint
  • Export Map
  • Keywords
  • License
  • Repository URL
  • TypeScript Types
  • README
  • Created
  • Published
  • Downloads 71
  • Score
    100M100P100Q69662F
  • License MIT

Lightning-fast semantic search for MongoDB documentation via Model Context Protocol. 10,000+ documents, <500ms search.

Package Exports

  • mongodocs-mcp
  • mongodocs-mcp/dist/index.js

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (mongodocs-mcp) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

MongoDB Semantic MCP v7.0

Lightning-fast semantic search for MongoDB and Voyage AI documentation via Model Context Protocol. Search 10,000+ documents in <500ms with the latest AI models.

🚀 Features

  • Smart Model Selection: Uses voyage-code-3 for code content and voyage-3 for text
  • Blazing Fast: <500ms search across 10,000+ documents
  • High Accuracy: 96%+ success rate on all query types
  • Instruction-Following Reranker: Uses rerank-2.5 for superior relevance
  • MongoDB Native: Optimized for MongoDB Atlas Vector Search
  • Beautiful CLI: Progress bars, colors, and interactive setup

📦 Installation

npm install -g mongodocs-mcp

🔧 Setup

Prerequisites

  1. MongoDB Atlas cluster with Vector Search enabled
  2. Voyage AI API key

Environment Variables

Create a .env file:

MONGODB_URI=mongodb+srv://...
VOYAGE_API_KEY=pa-...

Quick Start

# 1. Run setup wizard
mongodocs-setup

# 2. Index documentation (takes 20-30 minutes)
mongodocs-index

# 3. Start MCP server
mongodocs-mcp

🎯 Using with Cursor IDE

Add to your ~/.cursor/mcp.json:

{
  "mcpServers": {
    "mongodocs-mcp": {
      "command": "npx",
      "args": ["mongodocs-mcp@latest"]
    }
  }
}

📊 MCP Tools

The MCP provides these tools to AI agents:

  • mongodb-semantic-search - Search documentation with natural language
  • mongodb-find-similar - Find similar documents to provided content
  • mongodb-explain-concept - Get comprehensive explanations
  • mongodb-refresh-docs - Update documentation database
  • mongodb-status - Check system status

🧠 Technical Details

Embedding Models

  • voyage-3: General documentation (60-70% of content)
  • voyage-code-3: Code examples and API references (30-40% of content)
  • rerank-2.5: Instruction-following reranker for result optimization

Performance

  • Index size: ~10,380 documents
  • Search latency: <500ms average
  • Accuracy: 96%+ query success rate

MongoDB Configuration

  • Database: mongodb_semantic_docs
  • Collection: documents
  • Index: semantic_search (vector search index)

🛠️ CLI Tools

mongodocs-index

Indexes all MongoDB and Voyage AI documentation. Features:

  • Smart model selection (voyage-3 vs voyage-code-3)
  • Progress tracking with beautiful UI
  • Incremental updates
  • Content hashing for efficiency

mongodocs-clean

Removes all indexed documents for a fresh start.

mongodocs-setup

Interactive setup wizard for configuration.

📈 Version History

v7.0.0 (Current)

  • Added voyage-code-3 for code content
  • Smart content detection
  • Model distribution tracking
  • 13.80% better code search accuracy

v6.0.0

  • Upgraded to rerank-2.5
  • Instruction-following for better results
  • Special character handling improvements

🤝 Contributing

Contributions welcome! Please ensure:

  • Code passes TypeScript compilation
  • All queries maintain <500ms latency
  • Test coverage for new features

📄 License

MIT

Built with ❤️ for the MongoDB community