Lost an avatar video job id? List avatar videos instead

If your worker crashed after submitting an avatar video and you lost the job id, do not submit again. List your avatar videos with GET /v1/avatar-videos and read one with GET /v1/avatar-videos/:id. The list returns summaries of your workspace's avatar-video jobs, so you can find the render and read its result once it finishes.

This is the recovery path the docs describe, alongside the jobs endpoints for status, events and result.

Which call for which question

Different handles answer different questions.

A recovery routine

List, match by the script or title you remember, read, and download from the Sume URL.

curl https://api.sume.com/v1/avatar-videos \
  -H "Authorization: Bearer $SUME_API_KEY"

curl https://api.sume.com/v1/avatar-videos/avatar_video_123 \
  -H "Authorization: Bearer $SUME_API_KEY"

Prevent it next time

• Write the job id and your own record id to storage before you do anything else with the response.
• Send an Idempotency-Key built from your own record, so a retry returns the original job.
• Prefer a signed webhook for completion, with polling as the fallback.
• If you use previews, store the avatar_video_preview_id too; the final render hangs off it.

What the resource gives you

A video resource is the durable record: it points at the finished artifact, so you can rebuild a page or a message from it without holding on to the original job response. If the list is long, pass limit (1 to 100) and a status filter; ready is accepted as an alias for completed jobs. If you find two near-identical entries, a retry probably double-submitted, so send an Idempotency-Key next time.

Limits

The list route shows what exists, not why a job failed; use the job events for that. Completed results are mirrored to media.sume.com, so store that URL rather than a provider link. Keep your own copy of anything you must keep, since retention is not a promise the docs make on this page.