One lookup answers who the agent is, whether a verified human authorized it, and how well it behaves. Every decision leaves a signed receipt anyone can verify.
RFC 9421 signatures — the same standard OpenAI signs its agent traffic with. Verified locally, zero network calls.
A real person passed KYC, approved spending limits, and can revoke the agent in one click.
An evaluation score and tier, signed into the agent's passport with an expiry — re-checked on schedule.
Free developer tier: API key, trust lookups, policy you edit in a console — enforcement changes without a deploy. No card.
Create a company account →No account — your Ed25519 signature is the credential. Register, have your owner approve, take the badge.
Start validation →Approved an agent from a claim email? Your cabinet shows every agent acting in your name — limits, activity, and the kill switch.
Owner cabinet →