Skip to main content
Glama
MongoDB MCP Server

MongoDB MCP Server

by Sanjeev4523

MongoDB MCP Server

npm version License: MIT

A read-only Model Context Protocol (MCP) server for querying MongoDB databases. Designed to be used with MCP-compatible clients like Claude Code, Claude Desktop, etc.

Quick Start

Run directly with npx — no installation needed:

MONGODB_URI="mongodb://localhost:27017" npx mongodb-mcp

Or install globally:

npm install -g mongodb-mcp

Related MCP server: mongo-mcp

Environment Variables

Variable

Required

Description

MONGODB_URI

Yes

MongoDB connection string (e.g. mongodb://localhost:27017 or mongodb+srv://...)

Configuration

Claude Code

Add to your .mcp.json:

{
  "mcpServers": {
    "mongodb": {
      "command": "npx",
      "args": ["mongodb-mcp"],
      "env": {
        "MONGODB_URI": "mongodb+srv://user:pass@cluster.mongodb.net/dbname"
      }
    }
  }
}

Claude Desktop

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "mongodb": {
      "command": "npx",
      "args": ["mongodb-mcp"],
      "env": {
        "MONGODB_URI": "mongodb+srv://user:pass@cluster.mongodb.net/dbname"
      }
    }
  }
}

Tools

list_databases

Lists all databases on the connected MongoDB instance.

list_collections

Lists all collections in a given database.

Parameter

Type

Description

database

string

The database name

list_indexes

Lists all indexes on a collection.

Parameter

Type

Description

database

string

The database name

collection

string

The collection name

run_aggregation

Runs an aggregation pipeline on a collection.

Parameter

Type

Description

database

string

The database name

collection

string

The collection name

pipeline

object[]

The aggregation pipeline stages

run_aggregation_to_file

Runs an aggregation pipeline and writes results to a JSON file on disk instead of returning them over MCP. Returns only metadata (document count, file path, file size), which avoids loading large result sets into the agent's context window. Use CLI tools like jq, head, or the Read tool to selectively inspect the output file.

Parameter

Type

Description

database

string

The database name

collection

string

The collection name

pipeline

object[]

The aggregation pipeline stages

output_path

string

Absolute file path where JSON results are written

Safety

  • Read-only mode — Write stages ($out, $merge) in aggregation pipelines are rejected with an error.

  • Query timeout — Aggregation queries have a 50-second server-side timeout (maxTimeMS) to prevent runaway queries.

  • Read preference — The connection uses secondaryPreferred read preference as defense in depth.

Development

npm run dev    # watch mode
npm run build  # compile TypeScript
npm start      # run the compiled server

License

MIT

Available Tools

5 tools
list_collectionsA

List all collections in a given database

Inputs: database (required).

list_databasesA

List all databases on the connected MongoDB instance

Inputs: no named parameters.

list_indexesB

List all indexes on a collection

Inputs: database (required), collection (required).

run_aggregationC

Run an aggregation pipeline on a collection

Inputs: database (required), pipeline (required), collection (required).

run_aggregation_to_fileA

Run an aggregation pipeline on a collection and write results to a JSON file on disk. Returns only metadata (document count, file path, file size) instead of the actual documents, which avoids loading large result sets into the context window. Use CLI tools like jq, head, or the Read tool to selectively inspect the output file.

Inputs: database (required), pipeline (required), collection (required), output_path (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.6/5.0

    Scored across 5 tools

    Disambiguation5/5

    Each tool targets a distinct operation: listing databases, collections, indexes, running aggregations, and running aggregations to file. The two aggregation tools are clearly differentiated by output method and purpose.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun pattern (list_* for enumeration, run_* for execution). No mixing of conventions or vague verbs.

    Tool Count5/5

    Five tools is well-scoped for a focused MongoDB utility server. Each tool earns its place without redundancy or excessive breadth.

    Completeness2/5

    The set covers listing and aggregation but lacks common MongoDB operations such as find/query, insert, update, and delete. This leaves significant gaps for typical database usage, forcing agents to work around missing functionality.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      An MCP server that enables querying, inserting, updating, deleting, aggregating, and listing collections in MongoDB via natural language commands.
      23 npm
      MIT
    • F
      license
      Not graded
      quality
      D
      maintenance
      A secure, read-only MCP server that enables AI agents to connect to MongoDB, perform queries and aggregations, and manage connections.
      -