# Sente

> Accounts for AI agents — email identity, completed verification, vaulted credentials, kept-alive sessions. Docs for api.sente.run.

## Docs

- [Connections](https://docs.sente.run/api-reference/connections.md): Connect accounts you already own for delegated agent access — write-only credentials, TOTP vaulting, revoke and delete.
- [Identities](https://docs.sente.run/api-reference/identities.md): Create, list, and delete the managed email identities your agent uses.
- [Messages](https://docs.sente.run/api-reference/messages.md): Read an identity's inbox, wait for verification codes, and send email.
- [API overview](https://docs.sente.run/api-reference/overview.md): Base URL, authentication, error format, and rate limits for the Sente API.
- [Registrations](https://docs.sente.run/api-reference/registrations.md): Create and re-login accounts at third-party apps, manage credentials, and open logged-in browser sessions.
- [Runs](https://docs.sente.run/api-reference/runs.md): Observe, resume, take over, and abort browser-driven register/login/connect runs.
- [Webhooks](https://docs.sente.run/api-reference/webhooks.md): Subscribe to inbound mail and run events pushed to your endpoint.
- [Identities and registrations](https://docs.sente.run/concepts/identities-and-registrations.md): The two objects in Sente's data model — the identity (an email address and inbox) and the registration (an account at one app) — and why they are separate.
- [Runs and human takeover](https://docs.sente.run/concepts/runs-and-human-takeover.md): A run is one browser-agent job — register, login, or connect. When it hits something only a human should do, it blocks, pages you, and waits.
- [Connect an account you own](https://docs.sente.run/guides/connect.md): Owner-authorized delegated access — supply credentials for an account you already have, and Sente keeps it logged in, re-loginable, and revocable.
- [Receive verification codes](https://docs.sente.run/guides/receive-verification-codes.md): Block until an OTP or magic link arrives in an identity's inbox — extracted server-side from every inbound email.
- [Register a new account](https://docs.sente.run/guides/register.md): Sente creates an account at a third-party app under the identity's email — driving signup, completing email verification from the identity's own inbox, and vaulting the credentials.
- [Sessions](https://docs.sente.run/guides/sessions.md): Get a browser already logged into an account — as a CDP URL for your own automation, a portable Playwright storageState, or an interactive live view for a human.
- [Webhooks](https://docs.sente.run/guides/webhooks.md): Push notifications for inbound mail and run state changes, verified with a per-org secret header.
- [Sente — Accounts for AI agents](https://docs.sente.run/index.md): Managed email identities and real accounts for AI agents. Sente completes email verification from the identity's own inbox, vaults credentials, and keeps sessions alive.
- [Quickstart](https://docs.sente.run/quickstart.md): Give an agent its own email identity in a few minutes — one pasted line for a coding agent, or four CLI commands by hand.
- [CLI](https://docs.sente.run/sdks/cli.md): The sente command-line tool — identities, inbox, verification codes, register/connect, sessions, and webhooks from the terminal.
- [Python SDK](https://docs.sente.run/sdks/python.md): The sente-sdk client — stdlib-only, import sente, talks to https://api.sente.run/v1.
- [TypeScript SDK](https://docs.sente.run/sdks/typescript.md): The @sente-labs/sdk client — typed, zero runtime dependencies, talks to https://api.sente.run/v1.
- [Acceptable use policy](https://docs.sente.run/trust/acceptable-use.md): What Sente is for, what's allowed, and what we refuse — legitimate managed accounts for AI agents, not bot-farming.
- [Limits and plans](https://docs.sente.run/trust/limits.md): Per-organization caps, plan tiers, and what each limit error code means.
- [Security model](https://docs.sente.run/trust/security.md): Where credentials and session state live, who can read them, and what the AI models involved do and don't see.