Research Tools: US Government Publishing (GPO) Office Begins Public Preview of the GovInfo Model Context Protocol (MCP) Server

From a GovInfo.gov Post:

We are proud to announce the public preview of the GovInfo Model Context Protocol (MCP) server that will enhance access to GovInfo content for large language models (LLMs), specifically artificial intelligence (AI). MCP is an emerging industry standard that enables systems to create specific tools for LLMs to interact with, without requiring those systems to train their own custom models.

[Clip]

For the first time, GPO is providing an officially supported method to allow the use of LLMs and AI agents to “converse” with GovInfo, the world’s only certified trustworthy digital repository. This could benefit legal researchers, policy analysts, data consumers, government users, academics, compliance assistants, and many other users.

[Clip]

Through the GovInfo MCP, GPO is providing tools that enable LLM agents to take user prompts and then access up-to-date GovInfo content and metadata. The GovInfo MCP will provide the same response back to any LLM making the same request, acting as a universal API for AI agents to access live information and perform actions. Depending on the LLM and the user request, the LLM may take additional steps that will require user review.

[Clip]

With the addition of MCP, there are now three primary interfaces that access the same underlying data:

• Web User Interface (UI): humans
• API: programmatic and machine access
• MCP: LLM and AI access

What does the public preview look like? Where can I learn more?

This initial release provides basic search and retrieval capabilities, allowing the development community to provide feedback. More information is available in GPO’s API GitHub repository, including how to integrate the GovInfo MCP into your LLM usage and an example scenario.

Learn More, Read the Complete Announcement

Direct to US GovInfo MCP Documentation (via Github)

Filed under: Data Files, News, Open Access, Patrons and Users, Publishing