A Python framework and runtime for building, serving, and managing AI agent platforms, with a web UI for administration.
googleapis/mcp-toolbox
An open source MCP server that connects AI agents, IDEs and apps to databases with prebuilt tools or custom tool definitions.
- Stars
- 16,681
- Forks
- 1,765
- Open issues
- 396
- Last commit
- 3 days ago
What is it?
MCP Toolbox for Databases is an open source Model Context Protocol server that gives AI agents, IDE assistants and applications access to databases. Run it with a prebuilt flag for a given database, such as --prebuilt=postgres, and MCP clients like Gemini CLI, Claude Code or Codex get generic tools for listing tables, running SQL and exploring schemas.
For production use, the project also works as a framework for custom tools. Developers define sources, tools, toolsets, prompts and resources in a tools.yaml file, with parameterized SQL statements and restricted access. Client SDKs for Python, JavaScript/TypeScript, Go and Java load these tools into frameworks such as LangChain, LlamaIndex, Genkit and ADK.
The server supports a broad set of databases, including PostgreSQL, MySQL, BigQuery, Spanner, MongoDB, Redis, Elasticsearch and ClickHouse. It handles connection pooling, IAM-based authentication and OpenTelemetry metrics and tracing, and includes a built-in UI for testing tools.
Key features
- 01Prebuilt generic tools for popular databases, usable from MCP-compatible IDEs and CLIs
- 02Custom tool definitions in YAML with parameterized SQL, toolsets and prompts
- 03Client SDKs for Python, JavaScript/TypeScript, Go and Java
- 04IAM-integrated authentication and connection pooling
- 05OpenTelemetry metrics and tracing export
- 06Interactive UI for testing tools and toolsets
Use cases
- Letting an AI coding assistant explore a PostgreSQL or MySQL schema and write database-aware code
- Giving a production agent a few restricted, parameterized SQL tools instead of raw database access
- Connecting LangChain or LlamaIndex agents to BigQuery or Spanner data
- Exposing database lookups to multiple agents through named toolsets
Quick start
- 1
Add the Toolbox server to your MCP client config
- 2
Run the Toolbox server with your config file
npx @toolbox-sdk/server --config tools.yaml
Things to know
- Connecting to a database requires setting connection environment variables or a configuration file.
- Running via npx is optimized for convenience rather than performance; binaries or containers are recommended for production.
Topics
FAQ
- What is mcp-toolbox?
- An open source MCP server that connects AI agents, IDEs and apps to databases with prebuilt tools or custom tool definitions. MCP Toolbox for Databases is an open source Model Context Protocol server that gives AI agents, IDE assistants and applications access to databases. Run it with a prebuilt flag for a given database, such as --prebuilt=postgres, and MCP clients like Gemini CLI, Claude Code or Codex get generic tools for listing tables, running SQL and exploring schemas.
- Is mcp-toolbox free to use? What license does it have?
- mcp-toolbox is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is mcp-toolbox still maintained?
- Actively maintained. The last commit to mcp-toolbox was on Oct 17, 2026.
- What can I use mcp-toolbox for?
- Letting an AI coding assistant explore a PostgreSQL or MySQL schema and write database-aware code Giving a production agent a few restricted, parameterized SQL tools instead of raw database access Connecting LangChain or LlamaIndex agents to BigQuery or Spanner data Exposing database lookups to multiple agents through named toolsets
- How do I install mcp-toolbox?
- Add the Toolbox server to your MCP client config Run the Toolbox server with your config file: npx @toolbox-sdk/server --config tools.yaml
- Where can I get mcp-toolbox?
- The source code and installation instructions are on GitHub: https://github.com/googleapis/mcp-toolbox
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source embedding database for AI apps, storing documents and vectors to power semantic, hybrid and full-text search.
A JavaScript library that embeds a GUI agent in any web page, letting users control the interface with natural language.
Open-source SDK for building and running AI agents in Python and TypeScript, with tools, MCP, memory, and model portability.
A local MCP server that routes prompts from Cursor, VS Code, Windsurf or Claude Desktop to ChatGPT, Claude, Gemini, Perplexity, BYOK APIs or local Ollama models.
A self-improving AI agent that creates and refines skills from experience, runs on many model providers, and works through CLI and messaging apps.