googleapis/mcp-toolbox

MCP Toolbox for Databases is an open source MCP server for databases.

What it solves

MCP Toolbox for Databases connects AI agents, IDEs, and applications directly to enterprise databases. It eliminates the need to write boilerplate code for database interactions, allowing users to query data in plain English, explore schemas, and generate database-aware code through an AI assistant.

How it works

It operates as a Model Context Protocol (MCP) server with two primary modes of operation:

  1. Prebuilt Tools: Users can instantly connect to supported databases (such as PostgreSQL, BigQuery, MySQL, and MongoDB) using generic tools like list_tables and execute_sql via a simple configuration.
  2. Custom Tools Framework: Developers can define specialized tools using a tools.yaml configuration file. This allows for the creation of structured queries, semantic search, and NL2SQL capabilities with restricted access for production security.

The server can be deployed as a binary, container image, or via NPM, and integrates with MCP-compatible clients (like Claude Code or Gemini CLI) and various SDKs (Python, JS/TS, Go, Java).

Who it’s for

  • Developers using AI-powered IDEs or CLIs who want to interact with their databases without switching contexts.
  • AI Agent Developers building production-ready agents that require secure, structured access to enterprise data sources.

Highlights

  • Broad Database Support: Compatible with a wide range of databases including Google Cloud (AlloyDB, Spanner, Firestore) and others like Snowflake, Redis, and Neo4j.
  • Production-Ready Features: Includes built-in connection pooling, IAM authentication, and OpenTelemetry for end-to-end observability.
  • Framework Integration: Provides SDKs for LangChain, LlamaIndex, and core libraries in multiple languages.
  • Flexible Configuration: Supports dynamic reloading of tools and the ability to group tools into toolsets.