Agent identity
The verifiable, persistent set of credentials that lets an AI agent prove who it is and act accountably — typically a DID, signing keys, and an email address.
Agent identity is what makes an AI agent a known, accountable actor rather than an anonymous script. At minimum it includes a persistent identifier the agent controls, a cryptographic key it can sign with, and a reachable address like email.
With an identity, an agent can prove it's the same actor across services, sign its requests, and leave an attributable trail — the foundation for letting it operate in the world without a human standing in for it.
What it's made of
A decentralized identifier (DID) anchors the identity, an Ed25519 keypair lets the agent sign and prove control, and capabilities like an inbox, vault, and wallet hang off it. Scopes on the agent's key decide what it may actually do.
Identity is the root of trust: every other capability is granted to, and logged under, the agent's identity.
Why it matters
As agents take real actions — sending mail, paying invoices, accessing systems — anonymity stops being acceptable. A verifiable identity lets you grant least-privilege access, audit what happened, and revoke trust cleanly when needed.
FAQ
Is agent identity the same as an API key?
No. An API key authenticates a request; an identity is a persistent, verifiable presence the agent controls and signs with.
Can an agent prove its identity to a third party?
Yes. With a did:web identity and signing key, the agent can sign requests that others can verify against its published key.
Give your agent an inbox.
A real email address, a vault, 2FA, and an identity in one API call.