How to transcribe a TikTok video to text

Transcribing means turning a video’s speech into text. It’s useful to save a tip, study a class, draft the script for your next video or add captions to your own content. With the link to a public TikTok video, the TikTok transcript page gives you the text plus SRT and VTT captions, with no sign-up.

Step by step on your phone

1. In the TikTok app, open the video and tap Share (the arrow on the right side of the screen).
2. Tap Copy link.
3. Open the transcript page in your browser, tap Paste, then Transcribe.
4. Wait a few seconds. A 3-minute video takes close to a minute.
5. Tap Copy text to paste it anywhere, or SRT captions to download the file.

If the share button doesn’t show Copy link, see how to copy a TikTok link.

On a computer

On the TikTok website, open the video and copy the address from the browser bar, or use Copy link next to the like and comment buttons. Paste it on the transcript page. The SRT file goes to your Downloads folder.

Captions in CapCut

The SRT has the text of each line with the moment it starts and ends. To add captions to your video:

1. open the project in CapCut on a computer and look for the option to import captions (an .srt file);
2. pick the downloaded SRT. The segments show up on the timeline, already in place;
3. adjust the font, size and position, and export the video.

Option names change from one CapCut version to another. If you can’t find caption import, you can copy the text and paste it into the editor’s own captions.

The most common mistakes

• Loud music over the speech. The AI mixes up words or skips parts. Videos with a clean voice come out almost error-free.
• Wrong language. If the video opens with a song or a phrase in another language, the AI may think the whole video is in that language. Tap Other language and pick the right one.
• Sentences nobody said. In silent stretches, the AI sometimes writes phrases like “Thanks for watching”. Delete them before using the text.
• Names and slang. Brand names, people’s names and new slang tend to come out spelled the way they sound.
• On-screen text. The AI doesn’t read what’s written on the video, it only listens. Anything that only appears on screen is left out.

Free usage limits

Public videos up to 3 minutes long, with up to 4 transcriptions every 15 minutes per connection. Asking for the same video in the same language again doesn’t use quota, and neither does a failed attempt. Videos from private or friends-only accounts can’t be transcribed.

Use it responsibly

Transcribing to study or keep a reference is personal use. To publish an excerpt, credit the speaker and link the video. Publishing the full transcript, or presenting someone else’s words as yours, needs the creator’s permission.