MCP MongoDB Integration
@the-sukhsingh
This project demonstrates the integration of MongoDB with the Model Context Protocol (MCP) to provide AI assistants with database interaction capabilities.
Overview
What is MCP MongoDB Integration?
MCP MongoDB Integration connects MongoDB with the Model Context Protocol (MCP) to give AI assistants direct database interaction capabilities. It consists of an MCP server that exposes MongoDB operations as tools and a client-side terminal chatbot that uses Google’s Gemini AI to invoke those tools.
How to use MCP MongoDB Integration?
Set up the server by navigating to mcp-mongo-project, install dependencies, create a .env file with MONGODB_URI and MONGODB_NAME, then start with npm start. Set up the client by navigating to client-side, install dependencies, create a .env with GEMINI_API_KEY, and run node index.js. The client connects to the server via Server-Sent Events (SSE) and lets you chat with Gemini to perform MongoDB operations.
Key features of MCP MongoDB Integration?
- Full CRUD operations exposed as MCP tools
- Gemini AI integration for natural language queries
- Terminal-based chatbot interface
- Supports find, insert, update, delete, and aggregation
- Count documents, list and create collections
- Communicates via Server-Sent Events (SSE)
Use cases of MCP MongoDB Integration
- Query collections and documents using natural language
- Insert, update, or delete documents through a chat interface
- Run aggregation pipelines via AI-driven tool calls
- List or create new collections interactively
FAQ from MCP MongoDB Integration
What MongoDB tools are available?
The server exposes 12 tools: findDocuments, findOneDocument, insertOneDocument, insertManyDocuments, updateOneDocument, updateManyDocuments, deleteOneDocument, deleteManyDocuments, aggregateDocuments, countDocuments, listCollections, and createCollection.
What are the prerequisites?
You need Node.js v14 or higher, a MongoDB instance (local or remote), and a Google Gemini API key.
How does the integration work?
The MCP server connects to MongoDB and exposes operations as tools. The client sends user queries to Gemini AI, which decides when to call a tool. The tool call is forwarded to the MCP server, executed, and the result is included in Gemini’s response.
What environment variables are required?
For the server: MONGODB_URI (connection string) and MONGODB_NAME (database name). For the client: GEMINI_API_KEY (Google Gemini API key).
What license does this project use?
It is licensed under ISC.
