YouTube subtitle files: SRT has no styling, so burn captions with Sume

On YouTube, an SRT file carries only basic captions: the help page says only basic versions are supported and no style info (markup) is recognized (read 2026-10-02). TTML supports styling and positioning, and WebVTT supports positioning with limited styling. If you want a specific look on every viewer's screen, burn the captions into the picture with Sume's video captions, and upload a sidecar file as well for the caption track.

What each YouTube format keeps

YouTube groups its supported files into basic, advanced and broadcast formats. The ones most creators meet are below.

Where burned-in captions fit

A sidecar caption track is separate data: viewers can switch it off and YouTube treats it as text. A burned-in caption is part of the video, so the font, colour and placement you chose are the ones everyone sees. You can have both, and for accessibility you should keep the track.

• Burned captions cannot be turned off or translated by the viewer.
• A sidecar track keeps the text available to the player as a caption track.
• Burn for look; upload a track for access.

The Sume side

POST /v1/video-captions takes a public HTTPS video_url and returns a captioned video. Pick a style such as slam, punch or tiktok-green. Per-request design overrides change colours, typography, placement, phrasing and motion, but are not supported on punch or tiktok-green.

To restyle without a second transcription, pass source_caption_id instead of video_url; Sume reuses the source video and its word timings.

What Sume does not give you

The docs say SRT uploads are unsupported as an input, and I found no SRT or WebVTT export from the caption job. If you need a sidecar file, take the words from video inspect with transcribe: true (word and sentence timings) and write the file in your own code. Review the wording before publishing; speech-to-text can mishear names.