Relay by Empyre

Verifiable identity for AI agents

Relay Identity gives each AI agent its own identity instead of asking a team to share one permanent API key. An agent receives an Agent ID and one-time Agent Secret, then authenticates for scoped, revocable access under the permission mode its owner selected.

Owners can rotate credentials, tighten permissions, review usage and revoke an agent. Tightening a permission mode invalidates derived access rather than leaving older credentials alive. The lifecycle is designed for machine identities that still remain accountable to a human organization.

How Relay works

Relay gives AI agents delegated access without handing them a user's password or a permanent credential. Integrators add a hosted OAuth 2.1 consent flow with S256 PKCE, exact redirect matching, short-lived scoped access tokens, refresh rotation, revocation and introspection.

Relay Platform powers Continue with Relay for applications that need an agent to act for a person. Relay Identity gives an agent a verifiable identity and revocable credentials of its own. Both products keep an audit trail, share one Empyre account, and are designed to fail closed when an approval, token or client is no longer valid.

Public pages

• Relay home
• Relay developer documentation
• Simple pricing for agent auth
• Talk to the Relay team
• Relay terms of service
• Relay privacy policy
• Verifiable identity for AI agents
• Pricing for AI agent identities

JavaScript is required to sign in and use the Relay dashboard. The product information above and the linked public pages remain readable without it.