How do I get the text of a TikTok video?

Paste a TikTok link, including the vm.tiktok.com and vt.tiktok.com short links the app generates, and this resolves it to the video and returns the spoken words as paragraphs with timestamps. TikTok generates captions for most videos and the creator can edit them, so where a caption track exists this reads it; where the account or the video is private there is nothing public to read and the tool says so rather than guessing.

No signup. The result is yours before we ask for anything.

Run it on your own domain

Paste captions instead

Any caption file works: SubRip, WebVTT, a player payload, or plain lines copied from the platform. The cleaning, paragraphing and exports all run on whatever you paste.

How this works

1A vm.tiktok.com or vt.tiktok.com link is a redirect with no video id in it, so it is followed before anything else can happen.
2The caption track is read where the video publishes one.
3Cues are joined into paragraphs on punctuation and pauses, and the rolling duplication automatic captions carry is removed.
4Timestamps are kept against each paragraph, so a quote can be found again in the video.
5Nothing is stored.

What it cannot tell you

TikTok does not reliably expose a caption track to a plain request. Checked on 12 September 2026: the page is reachable and its oEmbed endpoint answers without a key, giving the title and the author, and no subtitle track was present on the sample tested. A tool that promises every TikTok is describing a paid provider rather than the platform.
A private account or a removed video has nothing public to read.
Captions transcribe speech. Text burned into the video as part of the picture is not in them.

Questions people ask

How do I get the transcript of a TikTok video?

Paste the video link into a transcript tool, or in the TikTok app open the video, tap the share arrow and look for the caption or transcript option where the creator has enabled it. TikTok generates captions automatically for most videos.

Do short vm.tiktok.com links work?

They work here. A vm.tiktok.com or vt.tiktok.com link is a redirect that carries no video id, so it has to be followed before the video can be identified, which is why some tools reject them.

Can you get the text of a TikTok with no captions?

Only by running speech recognition over the audio, which is a different and more expensive job than reading a caption track. Any tool offering it for free is either running at a loss or reading captions and calling it transcription.

Does this work on private TikTok videos?

No. A private account or a removed video publishes nothing to read, and no tool can see it without being logged in as somebody who can.

The short version, if you would rather not run it

TikTok is the hardest of the three platforms here to read from a server, and the reason is worth knowing before choosing a tool. The video page is reachable and the official oEmbed endpoint answers without a key, so the title and the author are easy. The caption track is the part that is not reliably there, and the page carries machinery for challenging automated requests.

What that means practically is that a free TikTok transcript tool is doing one of three things: reading a caption track on the videos that publish one, paying a provider that maintains residential addresses and speech recognition, or quietly failing on most input. The first is honest and partial, the second is a monthly bill, and the third is the most common.

The short links deserve a mention of their own because they trip up more tools than anything else. When somebody shares from the app they get a vm.tiktok.com or vt.tiktok.com link, which contains no video id whatsoever, only a token that redirects. A tool that parses the link with a regular expression and never follows it will reject the exact form that most people paste.

Where a caption track can be read, the processing matters as much as it does on any platform. TikTok captions are short, fast and heavily fragmented, since the format rewards dense speech, and a raw dump of them is close to unreadable. Joining them into sentences and keeping a timestamp against each paragraph is what turns the output into something you can quote from.