PUBLICMCP

PUBLICMCP is an open standard designed to make organizations, businesses, training programs, and libraries discoverable and directly queryable by AI agents. It addresses the challenge that AI agents often lack a reliable, structured way to access up-to-date information about organizations, which leads to inaccuracies and reliance on web scraping or outdated summaries. By implementing the PUBLICMCP protocol, organizations can expose a public, read-only endpoint that provides structured data, ensuring that AI agents receive current and authoritative facts directly from the source.

The standard defines a set of required and recommended tools that a compliant server must expose. Core required endpoints include get_info for business identity and metadata, get_services for a full service catalog with pricing and delivery details, and get_location for geographic coverage. Recommended features extend this with tools like get_portfolio for work examples, get_markets for markets served, get_products for live product catalogs (including real-time Shopify integration), and search_site_content for full-text search across indexed pages. Additional endpoints allow agents to retrieve specific page content, list all site pages, and perform project qualification via AI-powered scoring. For libraries and training programs, PUBLICMCP offers endpoints such as search_library for full-text search across a large collection of primary sources, fetch_library for retrieving individual items, list_library_sets for curated topic sets, and get_training_sections for structured curriculum access, including resources for semiconductor engineering.

The platform is intended for a wide range of organizations, including businesses, nonprofits, libraries, and training providers seeking to be accurately represented and easily discoverable by AI assistants like ChatGPT, Claude, Gemini, and others. Deployment is flexible: organizations can self-host their PUBLICMCP-compliant server for free or use hosting services provided by PUBLICMCP. The standard is actively evolving and can be implemented on any technology stack, with FastMCP cited as a common implementation.

Every PUBLICMCP server is automatically indexed in a public directory, enabling AI agents to query organizations directly without scraping or guessing URLs. The approach emphasizes accuracy, control over organizational narratives, and future-proofing as AI-driven search becomes more prevalent.