Google AI ships Model Context Protocol (MCP) servers to data sharing, giving AI agents first-class access to public statistics

Google has released a Model Context Protocol (MCP) server for data sharing, which reveals public datasets (Census, Health, Climate, Economics) that projects are interconnected by a standard-based interface that can be queried in natural language. Data is now available using the Gemini CLI and Google’s Agent Development Kit (ADK).

What was released

• MCP Server This allows any client or AI proxy with MCP to discover variables, resolve entities, get time series, and generate reports from data sharing without manually encoding API calls. Google positioned it as “from initial discovery to generating reports” where example hints cover exploratory, analytical, and generative workflows.
• Developer ramp: PYPI packages, Gemini CLI streams and ADK samples/Colabs are embedded in data sharing queries within the proxy pipeline.

Why is MCP now?

MCP is an open protocol for connecting LLM proxy to external tools and data with consistent features (tools, tips, resources) and transport semantics. By shipping first-party MCP servers, Google makes data sharing addressable through the proxy that has been used for other sources, reducing per-integrative glue code and enabling registry-based discovery with other servers.

What can you do?

• Exploration: “What health data do you have for Africa?” → List variables, coverage and sources.
• analyze: “Compare life expectancy, inequality and GDP growth in BRICS countries.” → Search series, normalize GEOS, align years and return to table or chart payloads.
• generate: “Create a concise report on income and diabetes in U.S. counties.” → Extract measures, calculate relevance, including source.

Integrate surfaces

• Gemini CLI/Any MCP client: Install the data CONSONS MCP package, point the client to the server, and issue an NL query; the client coordination tool calls behind the scenes.
• ADK Agent: Use Google’s sample agent to form a data-sharing call using your own tools (e.g., visualization, storage) and return the output from the source.
• Document entry point: MCP – Interactively query data with AI agents Link to QuickStart and User Guide.

Real-world use cases

Google Highlights A data proxybuilt by a campaign’s data Commons MCP server. It allows policy analysts to query tens of millions of sanitary financing data points through natural language, visualize results and export clean datasets for downstream work.

Summary

In short, Google’s Data CONSONS MCP server transforms a wide range of public statistics into a top-notch, protocol-local data source for proxy-reduces custom glue code, retains source, and cleanly pastes MCP customers on existing MCP customers like Gemini Cli and Adk.

Check GitHub repository and try Gemini CLI. Check out ours anytime Tutorials, codes and notebooks for github pages. Also, please feel free to follow us twitter And don’t forget to join us 100K+ ml reddit And subscribe Our newsletter.

Michal Sutter is a data science professional with a master’s degree in data science from the University of Padua. With a solid foundation in statistical analysis, machine learning, and data engineering, Michal excels in transforming complex data sets into actionable insights.