TikTok Captions: How to Add Them Automatically

The short answer: TikTok makes the captions, you fix them

TikTok auto-generates captions from your video's spoken audio inside the editing screen, and you edit the text before you post. Record or upload the clip, open the captions option in the editing panel, give it a few seconds to transcribe, correct whatever it got wrong, position the text, post.

That's the whole answer for most people. The built-in tool is free, it runs in seconds, and the transcription is good enough that reaching for anything else only makes sense in a few specific cases. The rest of this is the detail that separates deliberate captions from sloppy ones.

Captions, text overlays, and stickers are three different things

Three separate features get called "captions," and they behave differently once the video is live. Confusing them is the most common reason a clip looks fine in the editor and wrong in the feed.

That last column is the part that matters. Captions are a layer the viewer controls: switched off by anyone who finds them distracting, switched on by anyone who needs them. An overlay is burned into the frame forever. Neither is better. Captions carry dialogue; overlays carry titles, labels, prices, and anything the audio doesn't say.

One more thing muddies the vocabulary: the post description you type before publishing is also called a caption. That one sits under your username, not on the video.

How to add TikTok captions in the app, step by step

TikTok redesigns its editor regularly and the labels move, so this is the shape of the flow, not a promise about where a button sits this month. If a name has changed, look for the word "caption" in the tool strip on the right of the edit screen.

1. Record in the app or upload a finished clip. Both land on the same editing screen, and captions work either way.
2. Open the captions option in the editing panel. It usually sits in the vertical tool strip alongside text, stickers, effects, and voiceover.
3. Let TikTok transcribe. It reads the audio track and generates timed caption lines. Check which language it detected: if it assumed a language you didn't record in, the transcription comes back as nonsense, and the fix is to set the language before you correct a single word.
4. Read every line and edit it. Tap a caption line to open the text. Don't skip this one. It's the whole ballgame, and it's the next section.
5. Position and style the text. Drag the caption block where you want it and pick a style. Where it can safely go is below.
6. Preview, then post. Watch it once at normal speed with the sound off. If you can follow it, the captions work.

Viewers can toggle captions on and off themselves, usually from the share panel, so someone with them switched off won't see yours. Put anything essential, a name or a URL, into a text overlay too.

Editing the generated text is the step everyone skips

Automatic transcription is good at ordinary conversational speech and predictably bad at the words you most want right. A wrong word on screen doesn't read as a technical limitation, it reads as carelessness.

The categories that break, roughly in the order they'll bite you:

• Proper nouns. Names of people, companies, products, streets, cities. The transcriber has no way to know your co-founder is Siobhan rather than Shavonne. If your video names anyone, check every spelling.
• Brand names with deliberate misspellings. Anything that drops a vowel, swaps a letter, or fuses two words comes back as the nearest dictionary word.
• Jargon and acronyms. Industry vocabulary, model numbers, and spoken acronyms get flattened into whatever sounds closest. "SOC 2" becoming "sock two" is the genre.
• Homophones. There/their, your/you're, its/it's, principal/principle. These look sloppiest, because a viewer notices them instantly and blames you, not a machine.
• Numbers, units, and currency. Spoken figures come back inconsistently formatted. Pick a style and apply it to all of them.
• Speech over music or cross-talk. Wherever two sounds compete, expect a dropped or invented word.

Two habits make this fast. Read the captions cold, audio off, because your brain patches a wrong word when it can hear the right one. And keep a short list of the terms you say in every video, since they're the ones that break every time. A custom dictionary that learns those terms saves you the same corrections forever.

Styling and positioning that survives the feed

TikTok's interface sits on top of your video and it always wins. The username and post description run along the bottom of the frame; the like, comment, save, and share buttons run down the right edge. Anything you place there gets covered. Verify it yourself: play any video and look at where the chrome sits.

Keep caption text out of the bottom strip and away from the right-hand column. The safe area is the middle band of the frame, offset slightly left. Then check it on a phone rather than in a desktop editor, because the desktop preview doesn't show you the same overlay.

The rest of the rules are dull and they work. High contrast always: light on dark or dark on light, with a solid or semi-transparent box behind the words if the background changes mid-clip. Big enough to read at arm's length on a phone, which is bigger than looks right on a laptop. Short lines. One typeface. Legibility over personality.

Captioning for an audience that doesn't speak your language

TikTok has shipped translation features for captions and post text, but be realistic about what they give you. Availability varies by region, account, and language pair, and the output is machine translation layered on machine transcription. Errors compound: a name the transcriber mangled gets faithfully translated in its mangled form, and idioms travel badly.

For a casual international audience, that's a reasonable free bonus. For an audience you're trying to convert, it isn't enough. Get the source transcript right first, have a speaker of the target language check the translation, then add the result as a text overlay or a separate language cut.

When it's worth captioning outside the app

For a single short clip, the in-app tool is the right answer. Going outside TikTok earns its keep in three cases: a long edit, where correcting lines one tap at a time on a phone gets miserable fast; one cut going to several platforms, where you want a single corrected text version instead of five editing sessions; and any time you need the words as reusable text. On that last one, our explainer on what a transcript is and its formats covers the shapes it comes in.

This is where a transcription tool is honestly relevant rather than a detour. Laxis takes uploaded audio and video files and returns a timestamped transcript, in more than 100 languages with automatic language detection; the free tier covers 300 transcription minutes a month, with a 45-minute cap per recording. The feature that matters here is the Personal Dictionary: teach it the brand names auto-captions mangle and you stop re-fixing the same words. Plenty of other tools transcribe too. The point is working from corrected text, not a phone keyboard.

From there it depends what you're making. For subtitle timings rather than plain text, our walkthrough on how to build an SRT subtitle file explains the format. If you're still weighing options, we compared the realistic methods for turning video into text, plus a plainer version for turning audio recordings into text.

Why bother captioning at all

Three reasons, none of which needs a statistic propping it up.

Accessibility. Deaf and hard-of-hearing viewers can't use your video without text. That's the whole argument, and it doesn't get stronger with a percentage attached.

Sound-off viewing is normal. People watch in bed next to someone asleep, on a bus without headphones, at a desk in an open office. A muted video with no text is a video nobody watched.

Text holds attention through the first second. A feed is a scroll, and a legible line of words says what a clip is about before the audio has. Captions also let someone who lost the thread catch up instead of swiping away.

Add them, spend the extra minute fixing the words, and keep the text out of the bottom third.

Frequently asked questions

Does TikTok add captions automatically?

Yes. TikTok transcribes your video's spoken audio and generates timed captions inside the editing screen, at no extra cost. Open the captions option after recording or uploading, wait for the transcription, then correct the text before posting. Viewers can toggle captions on or off themselves, so treat them as an accessibility layer, not decoration.

What's the difference between TikTok captions and a text overlay?

Captions are the transcribed spoken-word track TikTok generates and times to your audio; a text overlay is text you type and place yourself. The practical difference is control: a viewer can switch captions off, but an overlay is burned into the video and always visible. Use captions for dialogue, overlays for titles and anything the audio doesn't say.

Why are my TikTok captions wrong?

Automatic transcription reliably struggles with proper nouns, brand names, jargon, spoken acronyms, strong accents, fast speech, and words that sound alike. It has no way to know you meant your company's name rather than the ordinary word that sounds like it. That's why the review step exists: read the lines once with the sound off, fix the wrong words, then post.

Where should I put captions so they don't get covered?

Keep captions in the middle band of the frame, clear of the bottom strip and the right-hand edge. TikTok's own interface sits in exactly those places, the username and post description along the bottom, the like, comment, save, and share buttons down the right side, and it covers anything you put there. Check the result on a phone.

Can TikTok translate my captions into other languages?

TikTok has translation features for captions and post text, but availability varies by region, account, and language pair, and the output is machine translation applied to machine transcription, so errors compound. That's a fine free bonus for a casual international audience. For an audience you're seriously trying to reach, translate the corrected transcript deliberately and have a native speaker check it.

When should I caption a video outside the TikTok app?

Use an outside tool when the edit is long, when the same cut is going to several platforms, or when you want the words as reusable text rather than only on screen. Uploading the file to a transcription tool gives you a timestamped transcript you correct once and reuse everywhere. For a single short clip, TikTok's built-in captions are free and good enough.