Evaluation record · mcp-server-mongodb

MCP MongoDB Server

v1.13.0

MongoDB (Official)

MCPdatabasenosqlmcpmodel-context-protocol

Strong

About This MCP

Official MongoDB MCP server (mongodb-js/mongodb-mcp-server), generally available since v1.9.0 (2026-03-24). Enables AI models to query, insert, update, and delete documents, manage collections and indexes, run aggregations, and administer MongoDB Atlas clusters (50+ tools spanning database, Atlas, Atlas Local, and Assistant tools). Supports a read-only mode, elicitation-based user confirmation for destructive operations, and disables server-side JavaScript by default.

Last Evaluated: July 9, 2026

Official Website

Trust Vector Analysis

Dimension Breakdown

🚀Performance & Reliability

query reliability

Query success rate testing

Evidence

MongoDB Driver — Built on official MongoDB drivers with stable query execution

highVerified: 2026-07-09

aggregation accuracy

Aggregation testing

Evidence

MongoDB Aggregation — Accurate aggregation pipeline execution

highVerified: 2026-07-09

large dataset handling

Scalability testing

Evidence

Implementation Review — Performance degrades with very large result sets; pagination needed

mediumVerified: 2026-07-09

connection stability

Connection stability testing

Evidence

MongoDB Connection Pooling — Stable connections with pooling and automatic reconnection

mediumVerified: 2026-07-09

error recovery

Error handling testing

Evidence

Implementation Review — Handles database errors with retry logic and error reporting

mediumVerified: 2026-07-09

🛡️Security

authentication security

Authentication mechanism review

Evidence

MongoDB Authentication — Supports MongoDB authentication mechanisms (SCRAM, X.509)

highVerified: 2026-07-09

connection string exposure

Credential security analysis

Evidence

MCP Security Model — Connection string with credentials stored locally; accessible to AI

highVerified: 2026-07-09

query injection risk

Injection vulnerability testing

Evidence

Security Analysis — AI can construct arbitrary queries; NoSQL injection risk if not validated

highVerified: 2026-07-09

data modification control

Operation authorization testing; raised from 62 due to the documented read-only mode in the official server

Evidence

MongoDB MCP Server Repository — AI can insert, update, and delete documents within user permissions, but the official server offers --readOnly / MDB_MCP_READ_ONLY=true to restrict tools to read, connect, and metadata operation types

highVerified: 2026-07-09

database drop risk

Destructive operation testing; raised from 70 following elicitation-based confirmation added in the GA releases

Evidence

MongoDB MCP Server Winter 2026 Release Notes — Destructive operations (drop-database, drop-collection, delete-many) now use MCP elicitation to prompt for explicit user confirmation before executing

highVerified: 2026-07-09

audit logging

Audit logging review

Evidence

MongoDB Audit — Auditing available in MongoDB Enterprise; limited in Community Edition

mediumVerified: 2026-07-09

🔒Privacy & Compliance

document data exposure

Data flow analysis

Evidence

MCP Data Flow — Query results and document data sent to LLM provider

highVerified: 2026-07-09

pii protection

PII protection assessment

Evidence

Privacy Analysis — No built-in PII detection or filtering; all queried data exposed

highVerified: 2026-07-09

field level security

Field security assessment

Evidence

MongoDB Field-Level Encryption — Supports field-level encryption but requires application-level implementation

mediumVerified: 2026-07-09

third party data sharing

Data sharing analysis

Evidence

LLM Provider Policies — Database content shared with LLM provider per their privacy policy

highVerified: 2026-07-09

query pattern exposure

Query privacy assessment

Evidence

Privacy Analysis — Query patterns may reveal data structure and business logic

mediumVerified: 2026-07-09

👁️Trust & Transparency

documentation quality

Documentation completeness review; raised from 75 as official first-party docs replaced community-maintained coverage

Evidence

MongoDB MCP Server Documentation — Official MongoDB documentation site with get-started guides, tool reference, and configuration options, maintained alongside the GA product

highVerified: 2026-07-09

query visibility

Query logging assessment

Evidence

MCP Protocol — All queries logged in MCP transaction logs

highVerified: 2026-07-09

open source transparency

Source code review

Evidence

GitHub Repository — Open source under Apache 2.0, officially maintained by MongoDB (1.1k+ stars)

highVerified: 2026-07-09

operation coverage clarity

API documentation review

Evidence

MCP Server Documentation — Clear but incomplete documentation of supported MongoDB operations

mediumVerified: 2026-07-09

⚙️Operational Excellence

ease of setup

Setup complexity assessment

Evidence

Setup Documentation — Simple setup requiring MongoDB connection string

highVerified: 2026-07-09

query performance

Performance benchmarking

Evidence

MongoDB Performance — Performance depends on query complexity and indexing (typically 10-500ms)

mediumVerified: 2026-07-09

reliability

Reliability analysis

Evidence

MongoDB Stability — Built on mature MongoDB drivers with high reliability

highVerified: 2026-07-09

operation coverage

Feature coverage assessment; raised from 80 reflecting the expanded GA tool surface including Atlas management

Evidence

MongoDB MCP Server Repository — 50+ tools: CRUD, aggregations, indexing, and collection management plus Atlas cluster/user management, Atlas Local tools, GA lexical and vector search, and sample dataset loading

highVerified: 2026-07-09

community support

Community support assessment; raised from 75 reflecting first-party GA maintenance

Evidence

MongoDB MCP Server Repository — Officially maintained by MongoDB with frequent releases (latest stable v1.13.0, 2026-06-16), 1.1k+ stars, and a companion mongodb/agent-skills package

highVerified: 2026-07-09

Strengths

+Official MongoDB implementation, GA since March 2026, with 50+ tools
+Comprehensive CRUD, aggregation, and Atlas cluster management capabilities
+Read-only mode (--readOnly) and elicitation-based confirmation for destructive operations
+Server-side JavaScript disabled by default as a security measure
+GA lexical and vector search tools for AI retrieval workloads
+Open source (Apache 2.0) with first-party MongoDB maintenance and official docs

Limitations

!Query results and document data exposed to LLM provider
!No built-in PII detection or sensitive data filtering
!Risk of NoSQL injection without proper query validation
!AI can modify or delete data within permission scope unless read-only mode is set
!Connection strings with credentials stored in local configuration
!Limited audit logging in MongoDB Community Edition

Metadata

license: Apache 2.0

supported platforms

0: All platforms with Node.js; Docker

programming languages

0: TypeScript

mcp version: 1.0

github repo: https://github.com/mongodb-js/mongodb-mcp-server

github stars: 1100

docs: https://www.mongodb.com/docs/mcp-server/

api dependency: MongoDB Node.js Driver, Atlas Administration API

authentication: MongoDB connection string (MDB_MCP_CONNECTION_STRING) or Atlas API service account (client ID/secret)

security controls

0: read-only mode (--readOnly / MDB_MCP_READ_ONLY)

1: elicitation confirmation for destructive operations

2: server-side JavaScript disabled by default

ga date: 2026-03-24

first release: 2025 (public preview)

maintained by: MongoDB

Use Case Ratings

code generation

Good for generating MongoDB queries and database schemas

customer support

Useful for querying customer data and support ticket analysis

content creation

Moderate applicability for content management systems using MongoDB

data analysis

Excellent for analyzing NoSQL data with aggregation pipelines

research assistant

Good for researching data patterns and document structures

legal compliance

High risk of exposing sensitive legal documents; requires strict controls

healthcare

High risk of HIPAA violations; not recommended without strong data controls

financial analysis

Moderate risk; financial data exposure concerns

education

Excellent for teaching NoSQL databases and data modeling

creative writing

Useful for managing story databases and character information