What is Agent discovery?
Updated May 2026
Agent discovery is how AI tools find and understand other agents, usually via standard files like agent.json, llms.txt, and OpenAPI documents.
The infrastructure layer that makes agentic commerce possible. An AI client given just a URL should be able to discover everything it needs to call the agent's tools: agent card, MCP endpoint, OpenAPI doc, llms.txt summary.
Every withlove agent ships all four discovery surfaces by default at /<slug>/.well-known/agent.json, /<slug>/mcp, /<slug>/openapi.json, and /<slug>/llms.txt.
Frequently asked questions
What does agent discovery look like in practice?
An AI client fetches /.well-known/agent.json, reads the endpoints, then calls /mcp or /openapi.json to learn the tools. The whole handshake is automatic and standards-based.
Related terms
Your own AI assistant in five minutes
Skip the jargon. withlove gives you a working AI agent at a shareable URL — voice, calendar, prices, leads — without any of the setup.
Try withlove free