data object depends on
the event. Treat payloads as forward-compatible: ignore fields you do not use instead of
rejecting the request. Delivery is at least once and unordered, so
deduplicate on id.
Envelope
Events contain no prompt or answer text, cited sources, email addresses, person names
or Stripe identifiers.
All timestamps in the examples are synthetic.
run.completed
A run reached a final successful state, whether it was manual or scheduled. brandId
is the workspace of the run.
Run events are sent once a run can no longer change. While a run holds a quarantined
answer whose provider payload can still be replayed, for up to 7 days, its event waits.
During that time, List runs can already report the
run as final. This applies to run.failed and run.cancelled too.
run.failed
A run failed, including a planning failure. brandId is the workspace of the run.
run.cancelled
A run was cancelled. brandId is the workspace of the run.
pitch_audit.completed
A Pitch Audit report is ready. Pitch Audits belong to the team, so brandId is null.
pitch_audit.failed
A Pitch Audit failed terminally. brandId is null.
site_audit.completed
Every page of a page audit finished. brandId is the workspace of the audit.
prompt_suggestions.ready
A Prompt Studio generation finished. brandId is the workspace of the generation.
recommendations is the value that
Get a suggestion generation
returned for this run when the event was recorded. A later read of the run can
differ, because tracked prompts, competitors and newer generations change it. It
is null when the run stands for no recommendations: a profile preparation, a
pool that is no longer the newest usable pool, or a topic that was deleted. Only customer Prompt Studio runs
produce these events, not the public lead-generation tools.
prompt_suggestions.failed
A Prompt Studio generation failed. brandId is the workspace of the generation.
webhook.test
Sent only by Send a test event, to one endpoint,
even when it is disabled. brandId is null.

