YouTube caption file for a Short: with timing or without timing?

Pick With timing if you already have a caption file whose cues carry start and end times; pick Without timing if you have only the words and want YouTube to sync them. That is the choice YouTube's Upload file option presents. Sume can produce the words and sentence segments from your clip, but it does not create YouTube caption tracks, so the file step stays on YouTube's side.

The YouTube steps are from Add subtitles & captions, and the Shorts caption note is from Enhance your Shorts, both read 2026-10-02. Sume's are from the video inspect and video captions docs.

What does YouTube's page say about the two options?

To add captions by uploading a file you select Upload file, choose With timing or Without timing, select a file and click Save. The page also lists auto-sync, typing manually and auto-translate as other ways. It says transcripts are not recommended for videos over an hour long or with poor audio quality, and that automatic captions are limited to the video's default language.

The page points to a separate article for the supported subtitle file formats and does not list them, so check that article rather than assuming a format.

With timing or without: how do I choose?

With timing means the file carries cue times, so YouTube shows each line when you said so. Without timing means plain text that YouTube aligns itself. For a Short under three minutes with clear speech, Without timing is the quicker path. If your wording must break at exact moments, such as a rapid joke, use With timing.

Poor audio is the one case YouTube warns about directly: do not count on a plain transcript to sync on a muffled recording.

Where does Sume help?

Run video inspect with transcribe: true and segmentation.mode: "sentence". The docs say that returns word timings and gapless sentence segments that are caption-line shaped. Speech-to-text is $0.01 per audio minute (confirm in GET /v1/catalog), and a clip with no audio fails as inspect_source_has_no_audio, so probe has_audio first with frames: false. Add language_code as a hint, for example en.

You then turn those segments into the file format YouTube accepts. That conversion is yours to write; this post does not supply a format because YouTube's supported-format list was not on the page read.

curl -X POST https://api.sume.com/v1/video-inspect \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: short-transcript-001" \
  -d '{"video_url":"https://media.sume.com/artifacts/artf_demo/talk.mp4",
       "frames":false,"transcribe":true,
       "language_code":"en","segmentation":{"mode":"sentence"}}'

What about burned-in captions instead?

Burned-in captions skip the file question entirely, at the cost of not being toggleable. Sume's video captions endpoint accepts no SRT uploads; it takes a video URL, optional script_text, or authored cues with text, start and end. If you want both, burn the visual hook and upload a track too.

Sume does not translate caption tracks and does not touch your channel's settings.