What is AI identity?
AI identity is a verifiable, cryptographic identity for an AI agent, application, robot, device or edge node — alongside humans, organisations, developers and third parties. It establishes who or what is acting. It does not grant permission: authorization is decided separately from capability grants, policy, context and risk.
- Nine identity types supported conceptually.
- Hardware-backed keys where available.
- Rotation and revocation are first-class operations.
Frequently asked questions
Why do AI agents need their own identities?
Shared service credentials make agent actions unattributable. Per-agent identity enables least privilege, revocation and audit.
Does a verified identity mean an agent is trusted?
No. Identity is the first check. Capability, policy, context and risk still decide what it may do.