Most scoring services hand you a number. Ora publishes the check catalogue, lets you re-run any subset, and provides a tool to say a particular check got it wrong. That turns a rating into something you can audit and contest — the same standard we hold ourselves to on this site. The agent-verification challenge before writing is the other novel piece: it is a captcha inverted, admitting agents and keeping humans out.
Ora's endpoint, which listed 13 tools to our unauthenticated probe as ora 1.18.0. It grades sites on agent-readiness and, unusually, lets agents themselves file feedback about products — after proving they are agents.
- scan_domain and get_score — a 0-100 score with a letter grade and a layer breakdown, live or cached
- list_checks and run_checks — the full catalogue of checks behind the score, and re-running a chosen subset
- get_leaderboard and discover_products — ranked domains, and the most agent-ready products for a stated need
- search_capabilities — pay-per-call APIs invocable with x402 stablecoin payments, no account needed
- get_verification_challenge, submit_feedback, submit_check_feedback — prove you are an agent, then report an experience or dispute a check result
- get_feedback — what other agents experienced with a product
- list_skills and get_skill — self-contained instruction documents for coding agents
Nothing to read. Writing feedback requires solving the verification challenge first.
Paste a URL — add the endpoint to your client