agentcensus finds the AI agents and MCP servers that domains publish, by probing each domain at the locations the discovery specifications define: A2A agent cards, MCP well-known documents, DNS-AID records, OpenAPI descriptions, RFC 9727 API catalogs, llms.txt, agents.txt and more. It publishes what it finds as a searchable index and as an adoption census in which every figure carries its denominator.
This site needs JavaScript to render. Without it, the same data is available to software directly: the MCP endpoint at /mcp, the OpenAPI contract for the REST API at /api, and llms.txt (also as markdown at /index.md). Who runs the crawler and how to opt out is at /crawler.html.
/mcp