Webhooks

Rather than poll Status, supply a callbackUrl when scheduling an online interview recording. Scout Select will POST a signed event to that URL once the report is ready (or fails to generate) - the exact same delivery mechanism as assessment webhooks.

Events

• live_interview.completed - the report is ready. result is populated.
• live_interview.failed - either the bot never made it into the meeting after all retries, transcript processing failed, or report generation errored. result is null.

Payload shape

{
  "eventId": "8f14e45f-...",
  "eventType": "live_interview.completed",
  "occurredAt": "2026-08-12T15:52:10.120Z",
  "liveInterviewSessionId": "clv9d...",
  "externalReference": "your-internal-id-12345",
  "result": {
    "reportId": "clva1...",
    "overallScore": 7.5,
    "summary": "Strong grasp of indexing strategy and replication tradeoffs...",
    "recommendation": "HIRE"
  }
}

For live_interview.failed, result is null - check Status for botJoinStatus/status to understand what happened, since this event fires for both a failed bot join and a failed report generation.

Verifying the signature

Every callback carries an x-scout-signature header of the form sha256=<hex-hmac>, computed the same way as assessment webhooks - using the signing secret tied to the API key you used to schedule the session.

Delivery notes

• Same retry policy as assessment webhooks: up to 5 attempts with exponential backoff (roughly 1 min, 5 min, 30 min, 2 hr, then 12 hr) before being marked permanently failed. Deduplicate on eventId.
• callbackUrl requires API-key authentication (rejected with BAD_USER_INPUT for org-admin session callers) and must satisfy your organisation's webhook domain allowlist if one is configured.
• Keep polling Status as a fallback for critical flows, same as with assessments.