Official MCP server for MongoDB databases and Atlas clusters with query, schema and administration tools.

MCP Toolbox for Databases overview and key facts (October 2026)
Open-source MCP server from Google that connects AI agents to Google Cloud databases and many other databases.
What is MCP Toolbox for Databases?
MCP Toolbox for Databases is an open-source MCP server maintained under Google's googleapis organization that gives AI agents and IDEs controlled access to Google Cloud databases and many other SQL and NoSQL systems.
Supported sources include AlloyDB, BigQuery, Cloud SQL, Spanner and Firestore as well as PostgreSQL, MySQL, MariaDB, SQL Server, Oracle, MongoDB, Redis, Elasticsearch, ClickHouse, Neo4j, Snowflake and Trino. Prebuilt tools give instant access, while tools.yaml defines restricted, structured queries for production agents. The server handles connection pooling and IAM authentication and emits OpenTelemetry metrics and traces. SDKs exist for Python, JavaScript and TypeScript, Go and Java. The GitHub API showed about 16,500 stars and a push on September 28, 2026.
MCP Toolbox for Databases AI features and use cases
- Lets developers query databases from the IDEPrebuilt tools such as list_tables and execute_sql let a coding assistant explore a database during development. Developers check data and schemas without switching to a SQL client.
- Gives agents only approved queriesIn tools.yaml, teams define named tools with fixed, parameterized queries. Production agents can then answer customer or staff questions without free SQL access.
- Monitors agent database accessOpenTelemetry metrics and traces show which tools agents call and how often. Platform teams can audit and tune agent workloads on shared databases.
What sets MCP Toolbox for Databases apart
- Supports Google Cloud databases and many third-party systems in one server, including Oracle, Snowflake and Neo4j.
- Separates quick prebuilt tools for development from restricted, config-defined tools for production agents.
Where MCP Toolbox for Databases is strongest
Open-source software with no country restriction that runs wherever the connected databases are reachable.
Questions to ask before you buy MCP Toolbox for Databases
Do not give production agents the generic execute_sql tool; define restricted tools in tools.yaml and use database accounts with minimal rights. Query results pass to the model provider of the agent, so check whether personal data may leave your environment. You operate the server yourself, including updates and monitoring.
MCP Toolbox for Databases reviews
No reviews yet. Be the first to share how your team uses it.
Every review shows its verification level. We read each one before it goes live. How we check reviews
Questions about MCP Toolbox for Databases
Does it only work with Google Cloud?
No. Besides AlloyDB, BigQuery, Cloud SQL, Spanner and Firestore it supports PostgreSQL, MySQL, SQL Server, Oracle, MongoDB, Snowflake and more.
How can I restrict what agents can query?
Define custom tools with fixed queries in tools.yaml instead of exposing generic SQL execution.
Is it free?
Yes, it is open source under the Apache 2.0 license.
MCP Toolbox for Databases update historyno updates yet
No changes recorded for MCP Toolbox for Databases yet.
Changes the vendor made and confirmed user reports, each checked against the vendor's own pages. We review every tool monthly.
Alternatives to MCP Toolbox for Databases
Tools in the same category (Databases & data sources) with the most overlapping markets.
Supabase's official MCP server that lets AI assistants query tables, apply migrations, deploy Edge Functions and read logs.
Qdrant's official MCP server that gives AI assistants a semantic memory based on the Qdrant vector database.


