Captions by platform: sidecar file or burned in, and what Sume does

Sume burns captions into the video pixels; LinkedIn, YouTube and X each also offer a separate caption file route. So the choice is per platform: a sidecar file stays editable and toggles on and off, while a burned-in caption survives every platform and every muted autoplay. Sume's video captions API produces the second kind.

How Sume burns them

The request needs video_url; style, font, language, script_text, words and cues are optional. With no style, Latin text defaults to slam. Speech-to-text needs audible speech: a silent clip fails with caption_no_speech, and the fix is to send authored cues with text, start and end. The API says it does not take SRT uploads, so an SRT you already have must be converted to cues.

Restyling does not transcribe twice: send source_caption_id instead of video_url and Sume reuses the word timings. Credits are flat per accepted job for videos up to 60 seconds; the rate is on the API rate card.

Choosing

Pick burn-in when the same file goes to several platforms, or when the platform's own caption route cannot show your style. Pick a sidecar file when the platform needs a text track, for example LinkedIn ads, where caption processing must succeed for the ad to serve.

If you need timings to build a sidecar file, video inspect can return a transcript with words[] when you set transcribe: true; you convert those to the format the platform wants. Sume does not document a caption-file output.