What Sente is for
Sente gives an AI agent a managed email identity and gets it real, durable accounts at third-party apps — two ways: create (Sente registers a new account, completing email verification from the identity’s own inbox) and connect (you authorize access to an account your organization already owns by supplying its credentials, which Sente vaults and uses to log the agent in and keep the account alive). Either way it registers or logs in, vaults credentials, and keeps the account reachable and re-loginable. It is infrastructure for agents to hold legitimate accounts their operator is accountable for — not a tool for creating disposable or fraudulent accounts at scale.Acceptable use
- Registering and operating accounts your organization is accountable for and will use in good faith.
- Targets whose terms permit or do not prohibit automated registration, or where you have the site’s consent (your own app, a partner’s app, staging, or a ToS-permitting service).
- Using the identity’s inbox for the account’s own verification, notifications, and correspondence.
- Connecting accounts you are authorized to operate — supply credentials only for accounts your organization owns or is expressly authorized to access on the owner’s behalf. You can revoke a connection (Sente stops using it) or delete it (Sente purges the stored credentials) at any time.
- Reasonable volumes consistent with real product use (per-organization caps apply — see Limits).
Prohibited use
You may not use Sente to:- Farm accounts — create accounts in bulk to evade rate limits, free-tier limits, bans, or one-account-per-person rules; create disposable or throwaway identities at scale.
- Register at targets that prohibit it — apps whose terms ban automated or “bot” registration. We do not build for, demo, or market this use.
- Circumvent anti-abuse controls — solve or evade CAPTCHAs, spoof human traffic, or defeat bot-detection. A CAPTCHA is a stop-and-hand-to-human event, not something we clear for you.
- Spam or abuse the outbound channel — unsolicited bulk email, phishing, or anything that would damage the shared sending domain (which would break verification for every customer).
- Commit fraud or impersonation — create accounts under false pretenses to deceive a provider or a third party, or impersonate a real person or organization without authorization.
- Handle credentials you aren’t entitled to, or use vaulted credentials outside the account’s intended, authorized operation. With connect, supplying someone else’s credentials without their authorization is prohibited.
- Violate applicable law or the usage policies of our providers — including Anthropic’s usage policy, which governs the model that drives our browser and is binding on all use of Sente.
How we support compliance
- Confirm-before-submit: for targets where a human should form the terms-of-service agreement, the agent fills the entire signup and stops before the final submit; a person reviews it in a live browser view and clicks — settling who accepts the terms and owns the account. See Register.
- Honor blocks and revocations: as a matter of operating policy, if a target blocks us or objects, we stop running against it. (Code-level target allow/blocklisting is a scoped fast-follow; today this is policy.)
- Per-organization caps on identities, sends, and runs limit blast radius on the shared domain.
- Audit trail: registrations, credential access, and runs are logged per organization.