Gemma
Google's family of open-weight AI models you can download, run locally, and self-host
Google's open-source MCP server that connects AI agents to enterprise databases.
This is a serious, enterprise-minded piece of MCP infrastructure: YAML-defined tools, connection pooling, OAuth2/OIDC, and OpenTelemetry observability, all open source and backed by Google. For safely exposing databases to agents, it's a much better default than hand-rolling a server. Caveats: it's focused on data sources rather than general tools, self-hosting means you own security and ops, and while Google-maintained it's an open-source project (not a supported paid product), so treat SLAs accordingly. Verify supported databases and current releases.
MCP Toolbox for Databases is Google's open-source MCP server that connects AI agents to databases with connection pooling, OAuth2/OIDC auth, and OpenTelemetry observability, using YAML-defined tools. It supports AlloyDB, Cloud SQL, Spanner, Bigtable, self-managed Postgres/MySQL, and community databases like Neo4j. It is free and self-hosted, with SDKs across multiple languages.
MCP Toolbox for Databases is an open-source MCP server maintained under the googleapis organization that lets developers connect gen-AI agents to enterprise data securely. Originally named Gen AI Toolbox for Databases (its development predated MCP), it was renamed to reflect MCP support. It reduces boilerplate by letting you define tools in YAML, and it handles connection pooling, authentication (OAuth2/OIDC), and end-to-end observability via OpenTelemetry. Toolbox supports a broad set of databases: AlloyDB for PostgreSQL (including AlloyDB Omni), Spanner, Cloud SQL for PostgreSQL/MySQL/SQL Server, Bigtable, and self-managed MySQL and PostgreSQL, plus third-party contributions such as Neo4j and Dgraph. It is fully open source and free; Google added a Java SDK in March 2026 alongside existing language support. Because it's self-hosted infrastructure, you run and secure it yourself, which is exactly what many enterprises want for database access.
MCP Toolbox for Databases is Google's free, open-source MCP server for connecting agents to databases with pooling, OAuth2/OIDC, and observability via YAML-defined tools.
MCP Toolbox for Databases is maintained under Google's googleapis organization, originally as Gen AI Toolbox for Databases and renamed for MCP support.
It targets enterprises and developers who need to expose databases to AI agents securely and observably.
Toolbox sits between agents and data sources, providing connection pooling, OAuth2/OIDC auth, and OpenTelemetry observability, with tools defined in YAML to reduce boilerplate.
It supports a broad set of databases and offers SDKs across multiple languages, including a Java SDK added in March 2026.
Enterprises and developers, especially Google Cloud database users, who need controlled, observable agent access to relational and other databases.
Developers and data engineers exposing DBs to agents.
Platform/data engineering leads at enterprises.
Google Cloud community and database teams.
An enterprise data team, often on Google Cloud, that needs to give agents safe, observable, pooled access to production databases via MCP.
This is an open-source project maintained by Google (googleapis org); it is not a separately funded startup and has no license cost.
Yes. It is fully open source and free; you only pay for the infrastructure and databases you run.
AlloyDB, Spanner, Cloud SQL (Postgres/MySQL/SQL Server), Bigtable, self-managed Postgres/MySQL, and community databases like Neo4j and Dgraph.
Through a YAML configuration that specifies queries and parameters, reducing boilerplate.
No. It is a self-hosted open-source MCP server, so you run and secure it yourself.
Yes. It supports OAuth2/OIDC authentication and OpenTelemetry-based end-to-end observability, plus connection pooling.
Side-by-side pages for pricing, features, and best-fit use cases.
Google's family of open-weight AI models you can download, run locally, and self-host
Meta's open-weight 30B agentic model that runs local, multimodal AI agents on a single GPU
Open-source autonomous coding agent for VS Code that runs on your own model API keys.
Microsoft's multi-agent conversation framework (now in maintenance mode).