ujjawalptdr/Mongo-MCP-Server

🧠 MongoDB MCP Server

A Model Context Protocol (MCP) server built with Node.js, MongoDB, and the MCP SDK, enabling Large Language Models (LLMs) to perform real-time CRUD operations on MongoDB databases through structured, validated, and modular tools.

🚀 Overview

The MongoDB MCP Server acts as a middleware layer that bridges AI models (LLMs) and MongoDB databases.
It provides a set of dynamic MCP tools that allow LLMs to directly create, read, update, delete, and manage database documents — all through a structured, schema-validated protocol.

This project demonstrates how AI can interact safely and intelligently with databases using the Model Context Protocol (MCP).

✨ Key Features

• 🔗 LLM Integration: Enables language models to perform real-time CRUD operations on MongoDB.
• ⚙️ Modular MCP Tools: Each operation is defined as a reusable, independent tool.
• 🧩 Zod-Based Validation: Ensures all inputs are structured, safe, and schema-compliant.
• ⚡ Asynchronous Handlers: Designed for performance, scalability, and error resilience.
• 🔁 Persistent MongoDB Connection: Maintains a single client instance for efficient access.
• 🧠 AI-Driven Operations: Allows AI systems to manage data contextually and intelligently.

🧰 Available Tools

🛠️ Tech Stack

• Node.js – Runtime environment
• MongoDB – NoSQL database
• Model Context Protocol (MCP) SDK – Server framework for LLM integration
• Zod – Input validation library
• dotenv – Environment variable management

⚙️ Installation

# Clone the repository
git clone https://github.com/ujjawalptdr/mongodb-mcp-server.git

# Navigate into the project
cd Mongo-MCP-Server

# Install dependencies
npm install

🧾 Configuration

Create a .env file in the project root:

MONGODB_URI=mongodb://localhost:27017 (your_MongoDB_URI)
DB_NAME=<your_database_name>

▶️ Usage

Run the MCP Server

node src/server.js

You’ll see:

✅ Connected to MongoDB: <your_database_name>
🧠 MongoDB MCP Server running...

The server exposes a suite of tools that can be consumed by LLMs or MCP-compatible clients such as AI assistants, Copilot, Claude or custom applications.

🧠 How It Works

1. MCP SDK handles communication between the LLM and the server.
2. Each database operation (like createDocument or listDocuments) is defined as a tool with strict Zod validation.
3. LLMs can query, insert, update, or delete data by calling these tools through the protocol.
4. The server ensures data integrity and returns formatted responses that the LLM can interpret.

💡 Example Tool Definition

server.tool(
  "createDocument",
  "Insert a document into MongoDB",
  {
    collectionName: z.string(),
    data: z.record(z.any()),
  },
  mongoHandlers.createDocument
);

🧱 Project Structure

📦 mongo-MCP-Server
├── 📦 src          # MCP Server configuration and tool registration
    ├── 📜 server.js             # MCP Server configuration
    ├── 📜 mongoHandlers.js      # CRUD logic and MongoDB connection
    ├── 📜 mcpServerMongo.js      # MCP Server configuration and tool registration
├── 📜 package.json
├── 📜 package-lock.json
├── 📜 .gitignore
├── 📜 .env                  # Environment variables
└── 📜 README.md             # Documentation

⚡ Example Operations

🧩 Future Enhancements

• 🧠 Add support for advanced query filters and aggregation pipelines.
• 🔄 Implement automatic reconnection and connection health monitoring.
• ⚙️ Introduce caching for frequent document retrievals to improve performance.
• 🧰 Extend MCP tools for multi-database or cross-collection operations.
• 🔐 Integrate optional authentication for secure AI–database communication.

👨‍💻 Creator Info

Ujjawal Patidar
Full Stack Developer | Problem Solver 🔗 GitHub

📜 License

This project is open-source and available under the MIT License.

MongoDB MCP Server – bridging AI intelligence with real-time database control through the Model Context Protocol.