01 / BEFORE THE LINK
What you need before you start.
WatchToText works from a public TikTok video link. It takes the words a video already carries in a caption track, and where there is no caption track it runs speech recognition on the video's audio instead. Both routes end in the same place: text in the result card that you can copy or save.
Speech recognition writes down what it hears. When the result says Speech recognition, the words come from the audio and can differ from any text the creator placed on the video.
That distinction matters more than any step in the walkthrough below. A caption track is text the video already carries: written by the creator, or generated by TikTok's own automatic captions. Speech recognition is a machine listening to speech, and machines mishear. Knowing which of the two produced your transcript tells you how much checking it needs.
You do not need an account, a sign-up or an extension to use the page.
02 / GET THE LINK
Copy the link from the app or the browser.
Tap Share, then Copy link.
In the TikTok app, open the video and tap Share, then Copy link. On the web, copy the address from the browser's address bar.
Paste it and press Get the text.
Paste the link into the video field on the TikTok transcript page and press Get the text. If you started from the homepage, pick the TikTok tab first.
Wait for the result card.
The result opens below the input, with the words in order and a source line that names the route that produced them.
These link forms are accepted: a full video link, a vt.tiktok.com short link, a similar vm.tiktok.com link, a tiktok.com/t/ share link, or the numeric video ID on its own. Short links from the app are followed to the video automatically, so you can paste what Share gives you without editing it.
One thing the field will not accept: a photo post or a profile link. Neither is a video, and the page says so instead of returning filler text.
03 / WHAT THE TOOL DOES
Caption track first, speech recognition second.
When the video already has a caption track, WatchToText reads it. The words come from that track, and the timings are the ones attached to it.
When the video has no caption track, the page does not stop there. It works from the video's audio and sends that audio to speech recognition, and the timings in the result come from the recognised words rather than from a caption file. This second route takes longer on a long video, and it is limited: speech recognition handles up to 20 minutes per video, while a caption track has no length limit here.
A TikTok video with no captions on screen is not a dead end on this page.
Paste a link and the result tells you which route ran. Read that line before you quote anything.
This is the short-video path. A YouTube video with no captions at all is a different job, covered in how to transcribe a video with no captions.
04 / CHECK THE RESULT
Read the result before you keep it.
Speech recognition and caption tracks can both contain mistakes, especially with names, accents and background noise. Skim the first few lines and compare them with the video: if a name or a technical term is wrong once, it is probably wrong everywhere it appears.
Music is the other common weak spot. When a clip is mostly music, the model may return a repeating line or a stock phrase instead of real words. Clips that are mostly talking work best, and a music-heavy clip is worth treating as a sample rather than a transcript.
If you are keeping the text for a document, this is the moment to fix what the machine got wrong: correct the names, restore the punctuation you need, and remove the passages you have no use for. WatchToText does not edit the words for you, so the version you save is the version you checked.
05 / SAVE THE WORDS
Copy, TXT, SRT or VTT.
Four ways out of the result card, and the right one depends on what happens next.
- Copy puts the text on your clipboard and creates no file. Good for pasting into a note, a message or a draft.
- TXT is the reading copy: the words in order, with no clock.
- SRT gives numbered cues with a start and end time for each one, which is what a subtitle editor expects.
- VTT carries the same cues in WebVTT form for a web player.
If a result arrives without timings, SRT and VTT stay switched off and the text can still be copied.
06 / WHEN IT FAILS
Read the message, then choose the next step.
The video will not load
Private, removed and restricted videos may not load at all. The status line says the video could not be loaded right now, and the only useful next step is another public video or a second attempt later.
The field rejects the link
That message means the input does not look like a TikTok video link. Go back to the app, use Share then Copy link, and paste that. A profile link or a photo post will not pass either.
The audio has no recognisable speech
A music-only or silent clip can come back with a message instead of text: no caption track, and no speech the model could use. There is nothing to fix on your side.
The video is longer than 20 minutes
That message applies to the speech recognition route, which handles up to 20 minutes per video. A video with a caption track is not limited in that way.
07 / QUESTIONS
Do I need a TikTok account?
No. There is no account, sign-up or extension, and the video only has to be public.
What if the video has no captions?
Then there is no caption track to read, and the page transcribes the audio instead. That route handles up to 20 minutes per video.
Is the transcript the same as the captions on screen?
Only when the result says caption track. When it says speech recognition, the words came from the audio and can differ from text the creator placed on the video.
Can I save the result?
Yes. Copy it, or save it as TXT, SRT or VTT. SRT and VTT need timings; a result without timings can still be copied.
READY WHEN YOU ARE
Keep the part worth quoting.
One sentence you want to use. One link you want to remember. Paste it and take the words with you.
Open the TikTok transcript tool ↗