What are auto captions?

1. Speech recognition turns the audio into text. Modern models, like OpenAI's Whisper, give every word a start and end time, which is what lets captions appear word by word.
2. The words are grouped into lines. Subtitles on a long video show a sentence or two at a time; captions on a short clip show two to four words, so they can be read at a glance on a phone.
3. The lines are shown one of two ways: as closed captions the app draws and viewers can switch off, or burned into the picture, so they look the same on every app (see burned-in captions).

The apps each have their own version, under their own names:

The catch for clips: an app's auto captions or closed captions exist only on that app. A clip cross-posted to all three needs its captions in the picture. Step-by-step instructions for each app are in how to add captions on TikTok, Reels and Shorts.

YouTube's own help page puts it plainly: automatic captions “might misrepresent the spoken content due to mispronunciations, accents, dialects, or background noise,” so creators should always review them. It adds that they may not appear at all when several people talk over each other, when languages are mixed, or when a video opens on a long silence. In clips, the usual misses are:

• Names of guests, products, games and places, especially unusual spellings.
• Slang, jargon and numbers, which can come out as the nearest everyday word.
• Crosstalk and music, where two voices or a backing track blur the words.

The fix is a read-through before posting: one wrong word in big captions is what viewers notice, and comment on. Fix names first, then anything that changes the meaning. Clean audio helps more than any setting, which is why podcast and interview clips caption best.

• Short lines, two to four words, that change as the words are said, so viewers read along without pausing the clip.
• Timed to the word. Captions that run ahead of the voice or lag behind it feel off, even when the words are right. A highlight on the spoken word, as in karaoke captions, makes the timing visible.
• Big, bold and outlined, readable over any background on a phone.
• Placed where nothing covers them: above the apps' captions and username, inside the safe zone.

Sizes, fonts and placement in more depth are in how to caption short videos, and the difference between captions and translated subtitles in captions vs subtitles.