mongodb-js/mongodb-mcp-server

What it solves

This project provides a Model Context Protocol (MCP) server that allows AI agents (like Claude, Cursor, and VS Code) to interact directly with MongoDB databases and MongoDB Atlas. It bridges the gap between LLMs and your data, enabling agents to query, manage, and analyze your MongoDB collections without requiring manual data exports.

How it works

The server acts as a standardized interface between an AI agent and MongoDB. It can be configured to run locally via Node.js or Docker, or used as a managed plugin. It supports multiple authentication methods, including MongoDB connection strings and Atlas API credentials. Once connected, the agent can call specific "tools" (functions) to perform database operations.

Who it’s for

• AI Agent Users: People using MCP-compatible clients (Cursor, Claude Desktop, etc.) who want their AI to have real-time access to their MongoDB data.
• Developers: Those who want to integrate MongoDB management and querying capabilities into their AI-driven workflows.
• Database Administrators: Users who need to automate Atlas cluster management via an AI interface.

Highlights

• Broad Client Support: Compatible with Cursor, VS Code, Claude, Codex, GitHub Copilot CLI, and Grok.
• Comprehensive Toolset: Includes tools for standard database operations (find, aggregate, insert, update, delete) and Atlas-specific management (creating clusters, managing access lists, and performance advisor).
• Flexible Deployment: Can be run as a local process, a Docker container, or a managed Atlas plugin.
• Safety First: Includes a --readOnly mode to prevent AI agents from accidentally modifying or deleting data.