Know which agent called you.
Give each agent a verifiable identity with an owner behind it. Your API can tell who sent the call and who it acts for.
Definition
Agent identity is a reusable, verifiable record of which agent is acting, who owns it, and which runtime it acts through.
How the identity travels
The identity is issued
The agent gets an identity proof bound to its owner and its wallet or account.
The identity travels
The same identity proof is presented in every runtime the agent works from.
The receiving system checks it
A gate confirms the identity is current before the call is trusted.
Verify the agent before the call
agent-gate.ts
When to use it
Use it when
- Your API serves agents you did not build.
- Actions need a named, accountable caller.
- The same agent works across more than one runtime.
Not when
- Anonymous public traffic with no privileged access.
- A single-tenant script that only you run.
Identity boundaries
- The identity names the agent and its owner. It never carries the owner secrets.
- An identity can be revoked. Old identity proofs fail the check.
- The owner controls which runtime presents the identity.

Know which agent called you.
Give each agent a verifiable identity and check it before the call runs.