Agent Identity
A first-class digital identity belonging to an AI agent — separate from any human — composed of an email address, credential vault, and authorization chain.
What is an agent identity?
An agent identity is the set of credentials and addresses that let an AI agent operate on the internet as itself, not as a human pretending to be the agent. At minimum it contains a routable email address (so other systems can send mail to the agent), a place to store API keys and OAuth tokens (so the agent can authenticate to other services), and a way to handle one-time codes (so the agent can pass 2FA challenges).
Crucially, an agent identity is not a service account. Service accounts are owned by an organization and shared between humans and machines, with no record of which human a given action traces back to. An agent identity is single-purpose: it belongs to exactly one agent, was created by exactly one human, and every action it takes can be traced back through a delegation chain to that human.
Why agents need their own identity
Most agents today borrow credentials from the humans who built them — a developer's Gmail, a personal Stripe key in a .env file, a phone number that receives SMS codes meant for a human. This works for one agent and one developer. It does not work when you have ten agents, or when the developer leaves the company, or when an agent needs to do something the developer doesn't want sitting in their personal sent-mail folder.
Giving an agent its own identity solves three concrete problems. First, isolation: a compromised agent doesn't compromise its operator's personal accounts. Second, attribution: every email sent and every API call made is recorded under the agent's identity, not buried in the operator's logs. Third, revocation: when an agent is retired, you delete its identity and every downstream credential dies with it — no orphaned access.
What an agent identity contains
A complete agent identity has four parts: an email address with full inbox semantics (send, receive, thread, label), an encrypted vault for storing secrets the agent needs to authenticate to other systems, a TOTP store for 2FA challenges from sites the agent logs into, and a delegation chain that records which human authorized this agent and what scopes were granted.
The email address is the most important piece — almost every workflow on the internet falls back to email for verification, password resets, receipts, and notifications. An agent without its own email is an agent that cannot complete any flow that involves a confirmation link, which is most of them.
How Loomal implements agent identity
Loomal provisions a complete agent identity in one API call. Each identity gets a real DKIM-signed address at agent@loomal.ai (or a custom domain on paid plans), an AES-256 encrypted vault, a TOTP store, and a delegation chain rooted at the human or organization that created it. The agent authenticates with a single API key and gets MCP tools for mail, vault, and TOTP operations.
The whole identity is revocable in one operation. Deleting an agent identity invalidates its API key, locks its inbox, wipes its vault, and severs every downstream OAuth grant — the audit trail remains so you can reconstruct what the agent did, but it can do nothing further.
Loomal primitives
identity.whoamiidentity.signSee it in production
More from the glossary
Agent 2FA (TOTP)
Time-based one-time password generation that lets an AI agent complete two-factor authentication on services that require it — without borrowing a human's phone or authenticator app.
Agent Email
A routable email address that belongs to an AI agent — not a forwarding alias, not a shared team inbox, but a first-class mailbox the agent sends from and receives into on its own.
Agent-to-Agent Email
Using standard SMTP email as the transport between AI agents — one agent sends a message to another agent's mailbox, and the recipient reads and acts on it. The simplest interoperable A2A protocol that already exists.
DKIM for Agents
DomainKeys Identified Mail signing applied to email sent by AI agents, so recipients can verify the message actually came from the agent's authorized domain — not a spoofed lookalike.
Build agents with their own identity.
Email, vault, and TOTP — provisioned in 30 seconds.
Get API Key — FreeLast updated: 2026-04-14