Change caption style without transcribing again: source_caption_id

To change a caption style without transcribing again, send source_caption_id instead of video_url on POST /v1/video-captions. Sume reuses the earlier caption's source video and the word timings it already has, so no second speech-to-text runs. The restyle is still a render and is billed as one.

Descript's 2026-09-17 changelog added 18 caption presets that apply when you click them. Trying looks through the API works the same way, one request per look, using Video captions.

What does the request look like?

Send the id of the finished caption job and the new style. This sample burns black-outline over the earlier result.

curl -X POST https://api.sume.com/v1/video-captions \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: caption-restyle-001" \
  -d '{ "source_caption_id": "vc_123", "style": "black-outline" }'

What is reused and what is billed?

When should I pass words?

Only to correct the wording. Pass words alongside source_caption_id and Sume burns your corrected text instead of the transcript. script_text, words, cues and segments are mutually exclusive, so pick one.

Can I change colors and motion too?

Yes, with design overrides on styles that support them; punch and tiktok-green do not. See customize burned-in captions for the fields.