Mcp Mongodb
Let your agent run read-only MongoDB finds, aggregates, counts, and stats while you debug data shapes and queries.
Overview
io.github.infoinlet-marketplace/mcp-mongodb is a Build-phase MCP server that lets AI agents perform read-only MongoDB find, aggregate, count, and stats operations.
What is this MCP server?
- Read-only find and aggregate pipelines with no write stages
- Count documents and surface collection statistics for agents
- Safer agent-assisted query tuning than handing out mongo shell write access
- @infoinlet/mcp-mongodb v0.1.1 stdio package on npm
- Useful when schema drift or bad aggregates block a solo builder ship window
- npm @infoinlet/mcp-mongodb version 0.1.1
- stdio MCP transport per 2025-12-11 server schema
- Read-only: find, aggregate (no write stages), count, stats
What problem does it solve?
You waste cycles exporting BSON or sharing Atlas access when the agent only needs to verify a filter, pipeline, or document count.
Who is it for?
Solo builders on MongoDB-backed services who want in-editor,data-grounded answers during backend and debugging work.
Skip if: Projects on SQL-only stacks, or anyone who needs inserts, updates, index builds, or shard administration from the agent.
What do I get? / Deliverables
After install, the agent can run read-only queries and aggregates so you fix data logic faster without granting destructive database permissions.
- Agent-executed find and read-only aggregate results
- Document counts and collection statistics for validation
- Faster iteration on backend queries without full GUI tooling
Recommended MCP Servers
Journey fit
Document database exploration happens while implementing and validating backend data models—the canonical shelf is Build backend work. Find, aggregate without writes, count, and collection stats are developer data-inspection tasks, not production error paging alone.
How it compares
Read-only MongoDB MCP adapter, not Mongoose code generation or a full database migration skill.
Common Questions / FAQ
Who is mcp-mongodb for?
Developers using MCP-enabled agents who store application data in MongoDB and want safe read access for exploration and query validation.
When should I use mcp-mongodb?
Use it in Build backend phases when designing aggregates, checking document shapes, or confirming counts before you ship API changes.
How do I add mcp-mongodb to my agent?
Install @infoinlet/mcp-mongodb from npm, configure stdio in your MCP host, and provide a read-scoped MongoDB connection string the server documentation describes.