Use cases
Businesses need checkpoints. Agents need passports.
It starts at the concrete moment an AI agent wants to act — and someone has to decide whether to trust it.
Where it fits
Six places agent trust breaks down today.
Businesses with websites
Add the free checkpoint before unknown agents submit forms, access accounts, or hit paid routes.
Hermes users
Passport an assistant before asking it to shop, book, or submit requests for you.
OpenClaw and Odysseus users
Prove which AI is acting, who owns it, and what it's allowed to do.
Engineering agents
Require approval before repo writes, deploys, incident commands, or infra changes.
Ops and finance agents
Put human review in front of refunds, vendor actions, and payment changes.
Support agents
Let AI draft and summarize freely — but gate account changes, refunds, and exports.
Pilot workflow
One website action. One passported agent.
Register
Create the passport and assign owner, purpose, scope, and expiry.
Protect
Route sensitive actions through an Gardien check.
Decide
Allow low-risk work, pause risky actions for approval, deny unknown agents.
Review
Pull the decision trail when anyone asks what happened.
Next step