News

Google's Data Commons Launches Server to Connect AI Systems with Public Datasets

• By John K. Waters
• 09/26/2025

Google's Data Commons project has released a Model Context Protocol (MCP) Server, designed to provide artificial intelligence systems with direct access to public datasets, the company announced on Tuesday.

The new server allows AI developers to integrate Data Commons' statistical information into their applications without requiring specialized knowledge of the underlying application programming interfaces (APIs), according to the announcement.

Data Commons, which aggregates public data from government agencies and international organizations, said the server aims to reduce AI "hallucinations" - instances where large language models generate incorrect or fabricated information - by grounding responses in verified statistical data.

The release includes a partnership application with the ONE Campaign; a global advocacy organization focused on Africa. The collaboration produced "The ONE Data Agent," which allows users to search health financing data using natural language queries across what the organizations describe as "tens of millions" of data points.

The tool addresses challenges in global health policy research, where relevant data is often scattered across multiple databases and reporting systems, according to the announcement.

Google said the server integrates with its Agent Development Kit and Gemini CLI tools, as well as other AI development platforms. The company has made the server available through the Python Package Index (PyPI) and provided sample code through Google Colab notebooks.

The Data Commons project, which launched several years ago, aggregates statistical information from sources including the World Bank, United Nations, and various national statistical offices. The new server represents the project's latest effort to make this information more accessible to AI systems and developers.