Auto captions
Also known as: automatic captions, auto subtitles, AI captions
Captions generated automatically from the spoken audio of a video, timed to the speech and usually editable before export or publishing.
Updated
What it means
Auto captions turn speech into on-screen text without anyone typing it. A speech-to-text model transcribes the audio, the software splits the transcript into short lines and times each line to when the words are spoken. The result can be burned into the picture or delivered as a separate track.
They matter because a large share of short videos are watched with the sound off, and captions also make videos accessible to deaf and hard-of-hearing viewers. Platforms such as YouTube create automatic captions on upload, while editing apps let you style them as big animated words, a common look in vertical video.
How it works in practice
Captions only work when there is speech worth captioning: a vlog intro, a guide explaining a route, a coach talking through a drill. Pure action footage with wind and engine noise produces nonsense lines or nothing at all.
Clean audio is the single biggest factor. A lavalier microphone or a camera with the mic facing the speaker beats a helmet camera at speed.
Always proofread. Names of spots, tricks and brands are often misheard, and line breaks may split phrases awkwardly. Keep lines short, two lines at most, and leave them on screen long enough to read. See subtitles and captions for burned-in versus sidecar files.
What to watch out for
Burned-in captions are permanent. If they sit low in the frame they can collide with platform buttons and descriptions, so respect the safe zones of the app you publish to.
Auto captions on music produce garbage. Lyrics from a track get transcribed partially and look like errors, so caption only the spoken parts.
Do not rely on a platform caption alone if accuracy matters. Automatic tracks are a starting point; upload a corrected file when the video carries instructions, safety information or sponsor messages that must be exact, and keep the corrected file for every future upload.
How RawClip handles it
RawClip does not add captions or text: the finished highlight is music and footage only. If you need captions, open the MP4 in an editor or app that generates them, or let the platform create them when you publish.