Verify that an AI agent acts for a real, unique person, then enforce custom policies for what it can do, how often, and when human approval is required.
Private identity · Custom policies · Verifiable actions
Accounts and API keys were designed for people operating software directly. Agents need something stronger: a way to prove they represent a real person, receive limited authority, and make requests that services can verify.
Signet provides that trust layer while keeping personal identity private.
|
Verify that an agent is backed by a real, unique person using zero-knowledge proofs without exposing passport details. |
Give agents short-lived, scoped permissions with clear limits on what they can do and for how long. |
Bind each request to the agent, its permission, and the exact action to prevent replay and unauthorized use. |
- No identity disclosure: Services do not receive names, passport numbers, or passport images.
- No cross-service tracking: Each service sees a different pseudonym for the same person.
- No permanent agent authority: Permissions can be narrow, limited, and short-lived.
- No blind trust: Registry updates, proofs, delegations, and requests are independently verifiable.
More than auth for your agent.