Hi PH 👋 I'm James. Quick origin story: Google, Microsoft and Hugging Face published a standard this year for how AI agents discover websites: one JSON file at a well-known path. I scanned 20 of the biggest tech sites this morning and not one publishes a valid catalog (Hugging Face's passed two days ago, then an update broke it). That felt like the right moment to build the tool that closes the gap. Agent Ardvark (an aardvark detective, my own design) does three things: 🔍 Investigates any domain against the official ARD v1.0 schema 📝 Drafts a valid ai-catalog.json from plain-English answers 🏅 Issues a Cleared badge for passing sites: it links to a live re-check, so it can't go stale Everything is free and runs in your browser (the only server call fetches the target file). No signup, no tracking, no stored data. Would love to hear whether you think agent discovery becomes table stakes, and what the tool should check next. :)

Agent Ardvark

Check if AI agents can find your site under ARD

Check if AI agents can find your site under ARD