Overview

The MongoDB MCP Server is the official Model Context Protocol server from MongoDB that enables AI assistants to interact directly with MongoDB databases and MongoDB Atlas. It provides natural language access to database operations, schema inspection, and cloud resource management for both local and Atlas deployments.

Official Server:

Developed and maintained by MongoDB

Key Features

🍃

MongoDB Atlas Integration

Manage Atlas clusters, projects, users, and network access directly through AI

📄

Document Operations

Query, insert, update, and delete documents using natural language commands

🔍

Schema Intelligence

Automatic schema inspection and pattern discovery across collections

🔐

Read-Only Mode

Built-in read-only mode for safe data exploration and analysis

Available Tools

Quick Reference

Tool	Purpose	Category
`connect`	Connect to MongoDB instance	Connection
`find`	Query documents in collection	Read
`aggregate`	Run aggregation pipelines	Read
`count`	Count documents in collection	Read
`insert-one`	Insert single document	Write
`insert-many`	Insert multiple documents	Write
`update-one`	Update single document	Write
`update-many`	Update multiple documents	Write
`delete-one`	Delete single document	Write
`delete-many`	Delete multiple documents	Write
`create-index`	Create collection index	Schema
`list-databases`	List all databases	Schema
`list-collections`	List collections in database	Schema
`collection-schema`	Describe collection schema	Schema
`collection-indexes`	List collection indexes	Schema
`collection-storage-size`	Get collection size in MB	Stats
`db-stats`	Get database statistics	Stats
`rename-collection`	Rename a collection	Admin
`drop-collection`	Remove a collection	Admin
`drop-database`	Remove a database	Admin

Detailed Usage

connect▶

Establish connection to a MongoDB instance using a connection string.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "connect",
  arguments: {
    connectionString: "mongodb://localhost:27017/myDatabase"
  }
});

Supports MongoDB Atlas, Community Edition, and Enterprise Advanced.

find▶

Query documents from a collection with optional filters and projections.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "find",
  arguments: {
    database: "myDatabase",
    collection: "users",
    filter: { status: "active" },
    limit: 10
  }
});

aggregate▶

Execute aggregation pipelines for complex data transformations and analytics.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "aggregate",
  arguments: {
    database: "myDatabase",
    collection: "orders",
    pipeline: [
      { $match: { status: "completed" } },
      { $group: { _id: "$customerId", total: { $sum: "$amount" } } },
      { $sort: { total: -1 } }
    ]
  }
});

count▶

Count documents in a collection matching optional filter criteria.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "count",
  arguments: {
    database: "myDatabase",
    collection: "users",
    filter: { status: "active" }
  }
});

insert-one▶

Insert a single document into a collection.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "insert-one",
  arguments: {
    database: "myDatabase",
    collection: "users",
    document: {
      name: "John Doe",
      email: "[email protected]",
      status: "active"
    }
  }
});

Disabled when using --readOnly flag.

insert-many▶

Insert multiple documents into a collection in a single operation.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "insert-many",
  arguments: {
    database: "myDatabase",
    collection: "users",
    documents: [
      { name: "Alice", email: "[email protected]" },
      { name: "Bob", email: "[email protected]" }
    ]
  }
});

update-one▶

Update a single document matching the filter criteria.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "update-one",
  arguments: {
    database: "myDatabase",
    collection: "users",
    filter: { email: "[email protected]" },
    update: { $set: { status: "inactive" } }
  }
});

create-index▶

Create an index on a collection to improve query performance.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "create-index",
  arguments: {
    database: "myDatabase",
    collection: "users",
    keys: { email: 1 },
    options: { unique: true }
  }
});

list-databases▶

List all databases accessible through the current connection.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "list-databases",
  arguments: {}
});

list-collections▶

List all collections in a specified database.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "list-collections",
  arguments: {
    database: "myDatabase"
  }
});

collection-schema▶

Inspect the schema structure of a collection by analyzing document samples.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "collection-schema",
  arguments: {
    database: "myDatabase",
    collection: "users"
  }
});

Useful for understanding data patterns and field types in schema-less documents.

collection-indexes▶

List all indexes defined on a collection.

use_mcp_tool({
  server_name: "mongodb",
  tool_name: "collection-indexes",
  arguments: {
    database: "myDatabase",
    collection: "users"
  }
});

MongoDB Atlas Tools

The server also provides tools for managing MongoDB Atlas resources:

Organization & Project Management: List and create Atlas organizations and projects
Cluster Operations: Create free tier clusters, list clusters across projects
User Management: Create database users with specific roles
Network Security: Manage IP access lists
Monitoring: Access alerts and performance advisories
Atlas Local: Deploy and manage local Atlas development environments

Installation

Prerequisites

Node.js 20.19.0+, v22.12.0+, or v23+
MongoDB connection string (local or Atlas)

Configuration

{
  "mcpServers": {
    "mongodb": {
      "command": "npx",
      "args": [
        "-y",
        "mongodb-mcp-server@latest",
        "--readOnly"
      ],
      "env": {
        "MDB_MCP_CONNECTION_STRING": "mongodb://localhost:27017/myDatabase"
      }
    }
  }
}

Read-Only Mode:

The --readOnly flag is recommended for safe data exploration. Remove it to enable write operations.

Atlas Connection String Example

{
  "env": {
    "MDB_MCP_CONNECTION_STRING": "mongodb+srv://username:[email protected]/myDatabase"
  }
}

Common Use Cases

Data Analysis

Explore data patterns and schemas using natural language:

"Show me the schema of the users collection and find the most common user types"

Query Generation

Generate complex aggregation pipelines:

"Create an aggregation to group orders by customer and calculate total spending,
sorted by highest spenders first"

Database Administration

Manage indexes and performance:

"Create a unique index on the email field in the users collection,
then show all indexes for that collection"

Additional Deployment Options

Docker

Run the server in a container without local Node.js:

docker run -e MDB_MCP_CONNECTION_STRING="mongodb://..." mongodb-mcp-server

HTTP Server

Deploy as an HTTP service:

npx mongodb-mcp-server --transport http --port 3000

HTTP Security:

When using HTTP transport, implement authentication, TLS encryption, and firewall protection. Never expose directly to the internet.

MongoDB MCP Server