Sender protection

See sender trouble early.
See the way back.

Production traffic runs behind verified domains, warm-up limits, rolling health checks, and project-level pauses. Metrics explains the signal and the recovery path.

Get an API keyRead autopilot docs

Complaint and repeat incidents can require operator review; recovery is never forced past the observed evidence.

Deliverability autopilotExample project healthHealthy
01ObserveRolling bounce and complaint signals
02ProtectThrottle or pause the affected project
03RecoverCooldown, controlled probes, restored capacity
Metrics shows the active signal and next evaluation.
Authenticated domainsAccount-wide suppressionsSigned event deliveryProject-scoped recovery
Built into the path

Guardrails you do not have to remember.

The safest default is part of the send path, not another checklist someone has to keep open.

Recipient

Unsafe addresses stop early

Hard bounces, complaints, provider blocks, and unsubscribes create suppression state before another provider handoff.

Sender

Production starts verified

Domain ownership, SPF, and DKIM records are checked before the sender moves into production.

Event

Callbacks arrive signed

Webhook payloads include an HMAC-SHA256 x-noticeapi-signature so your app can verify the raw body.

Access

Keys stay inside their grant

Project-scoped keys can be narrowed for a job, shown once, and revoked without replacing every application key.

Project isolation

Trouble pauses one project, not the whole workspace.

Routine deliverability incidents stay with the project that caused them. The account-level pause remains an operator emergency control.

Read the recovery states
  1. 01WatchMeasure the rolling signalBounce and complaint outcomes are evaluated against the current project and plan thresholds.
  2. 02ReduceThrottle before a hard stopWarning-range traffic can move into a lower production budget while health is re-evaluated.
  3. 03IsolatePause the affected projectA threshold breach stops that project's production traffic while sandbox testing remains available.
  4. 04RecoverEarn capacity through probesA first bounce incident can progress from cooldown into controlled sends, expanding only when outcomes stay healthy.
Documented boundaries

Read the policy behind each control.

The public guides explain what runs automatically, what your application must verify, and when an operator has to review the next step.

Unsubscribe handlingOne-click List-Unsubscribe and how opt-outs carry across later sends.Security practicesAPI keys, webhook signatures, admin access, logs, and retention.Domain setupOwnership, SPF, DKIM, DMARC guidance, and per-record verification.SuppressionsBounce, complaint, provider, unsubscribe, and manual recipient blocks.Autopilot docsMonitoring, warm-up limits, pauses, cooldowns, probes, and review paths.Acceptable useConsent rules and the sending behavior NoticeAPI does not support.
Test the stop path too

Rehearse the outcome before production.

Use simulator recipients to trigger delivery, bounce, complaint, and suppression behavior without sending real mail.

Get an API keyExplore the simulator