Google has released MCP Toolbox for Databases, an open-source MCP server that connects AI agents directly and securely to enterprise databases. Built on the Model Context Protocol (MCP), it supports more than 17 databases.
Google Cloud · Open Source
MCP Toolbox for Databases: one config file connects AI agents to 30+ enterprise databases
Google's open-source MCP server gives AI agents — Claude, Gemini CLI, LangChain and LlamaIndex clients — secure access to production data, with connection pooling, IAM auth and observability built in.
<10
Lines of code to embed
1
Config file (tools.yaml ) to wire up
4
Distribution channels: binary · Docker · npm · pypi
How an agent reaches the data
A standardized JSON-RPC path replaces custom per-source integrations — a common connector, like USB-C.
AI Agent
Claude · Gemini CLI · LangChain
→
MCP Toolbox
prebuilt tools + tools.yaml
→
Databases
AlloyDB · BigQuery · Postgres · 30+
Enterprise-grade by default
Connection pooling
Handled out of the box.
IAM authentication
Secure access to production data.
OpenTelemetry
Built-in observability.
Pre-approved tools
Limits arbitrary SQL generation.
What developers like
Pooling, auth and observability handled for them — leaving focus on logic. Pre-approved custom tools reduce the risk of agents running arbitrary SQL, a plus for security-minded enterprises.
The caveats
It is not zero-setup — some configuration is required. Others argue that building your own MCP server gives more reliable control over the data.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…