How to transcribe a TikTok video
Start with the individual video rather than a creator profile or a search-results page. Open it and listen briefly: a slideshow with written instructions is a different input from a person explaining those instructions aloud. transcript1 converts the spoken audio into text.
Copy the public video URL
Use TikTok’s share action to copy the link. Both full post URLs and the supported short-link forms work. Check that the video is available without signing into the creator’s account.
Submit it in the dashboard
Create an account, verify your email and paste the URL into the link transcription form. Leave language on auto unless you have a reason to supply a language hint. Start the transcription and open the completed result.
Review the script with the video
Check names, product terms and numbers against the recording. Fast speech, a loud backing track and rapid edits can affect recognition. Timestamps help you return to the uncertain passage instead of restarting the clip.
Copy or download the result
Copy the text into your notes, or download TXT, SRT, VTT or JSON. All four are views of the completed transcript; exporting another format does not require another paid transcription.
Which TikTok links and posts work?
A full TikTok video link identifies a specific post. A share link redirects to that post. Both can be submitted, but a full link is preferable for a repeated API or agent workflow because it can be matched to an earlier transcription when reuse is requested. A short link has no post identity at submission and does not get that reuse match.
The post must be public and contain a video with an audio track. Profiles, live pages, stories and TikTok photo posts are not supported. A public-looking URL can still fail if TikTok requires a login to fetch the underlying media. An upload of your original recording is another way to transcribe content you have locally.
| Input | What to expect |
|---|---|
| tiktok.com/@creator/video/… | A full video-post link; suitable for repeat workflows. |
| vm.tiktok.com or vt.tiktok.com | A supported short share link; no recent-result reuse match at submission. |
| tiktok.com/t/… | Another supported TikTok share-link form. |
| Photo post or creator profile | Choose an individual supported video instead. |
TikTok audio to text: speech, music and screen text
A TikTok often contains several kinds of text: the creator’s post description, words printed over the image, automatic captions and speech in the soundtrack. The transcript here comes from speech-to-text. It does not scrape the post description or read text from video frames. If a recipe quantity appears only on screen, add it from the video when preparing your own notes.
Music can make the spoken script harder to recover. Listen to any line that seems implausible, particularly where a voiceover overlaps lyrics. A silent audio track may yield little useful text; a video with no audio track fails with no_audio. A transcript should not be treated as evidence that an inaudible word was definitely said.
For a duet or a conversation, request speaker identification if distinguishing voices matters. The result uses labels such as Speaker 1 and Speaker 2. These are audio labels, not the names of the accounts visible on screen, and overlapping voices still need review.
Use the script to plan your next piece
For your own tutorials, a transcript gives you a draft checklist: separate the setup, each action and the result. Watch the video again to fill in steps demonstrated silently. That produces a more useful written tutorial than publishing the unedited narration alone.
For a spoken hook, keep its opening line beside the timestamp and your notes about the visual setup. Compare how the opening promise is answered later in the clip. The transcript supplies wording and timing; it does not measure audience retention or explain why a video performed well.
When turning your recording into a longer post, use the script as source material. Supply examples, links and context that were absent from the short clip. If you are studying someone else’s video, retain the source link and distinguish a quotation from your interpretation before sharing your notes.
Choose text for writing or subtitle cues for editing
Download TXT when your next step is an outline, a script revision or a searchable set of notes. Choose SRT or WebVTT when you need timed cues for a player or editor. Those files do not add styled captions to the TikTok itself; you import and format them in your editing tool.
If you cut the opening of a video or change its speed, the original timestamps will no longer match your edit. Review subtitle timing after the final cut. For animated word highlighting, the API can return word timings with include=words; use those timings in your own editing workflow rather than assuming each segment is a single subtitle-ready line.
Automate TikTok transcription through the API
Submit the public video URL to POST /v1/transcriptions using the create-transcription API. The platform is detected from the URL, so an explicit platform field is optional. TikTok uses speech-to-text; source=captions is reserved for YouTube. Save the returned job id and read its status rather than submitting the same clip again while it is processing.
Each successful job costs one credit per media minute, rounded up. A 45-second clip therefore uses one credit, while a 90-second clip uses two. New accounts receive 60 free minutes once after email verification, without a card. Failed and canceled transcriptions are not charged. These rules also apply when an MCP-connected assistant requests the video.