Adding TikTok’s built-in captions
Record or upload in the TikTok app, then on the edit screen tap Captions in the right-hand toolbar. TikTok transcribes the speech; tap the text to fix words and choose one of TikTok’s caption styles. Tap Done and post. On videos already published you cannot add captions afterwards; you would re-upload.
If TikTok’s captions are missing or wrong
The Captions button only appears when TikTok detects speech in a supported language; music-only clips and quiet audio get no button. Wrong words are common on names, slang and any language but English. Fix them by tapping the text, or turn captions off and burn your own in — the styles and safe zones for that route are on TikTok captions. If you caption inside CapCut instead, its own failure modes are on CapCut auto captions.
Can viewers turn burned-in captions off?
No, and that is the point for feed video: most people watch on mute, and a caption that can be switched off is a caption that is off. For long-form YouTube, where viewers may want the choice, use an .srt file instead; Titro exports one from the same project. The same two-way choice exists on Facebook and Reels.
Captions on TikTok ads
The ad feed is as muted as the organic one, and TikTok’s own captions do not follow a video into Ads Manager the way burned-in text does. A Titro export is a normal video file at the original resolution, so it uploads as ad creative like any other asset. Two placement habits pay off: keep the text in the centre band, because ad placements overlay call-to-action elements along the bottom edge; and put the claim in the first caption on screen, since the mute viewer meets your text before your voice. The same creative re-exports with German or Albanian subtitles in one click when the campaign splits by audience.
Keeping one look across your TikToks
TikTok’s built-in captions always look like TikTok. Burned-in captions can look like you — and the difference compounds over a profile grid. Start from Titro’s TikTok style, adjust the font and the highlight colour once, save the result under a name, and pick it at upload from then on. Keep the line length consistent too: two or three words per line is the setting that reads at feed speed, and switching it clip to clip is as noticeable as switching fonts. A profile where every video carries the same caption voice reads as a channel, not a collection of clips.
