TikTok Video to Text

Every processed video exposes a Download menu with three files, and all of them are on the free tier — there is no paywall on export formats.

• Plain text (.txt) — the spoken words as one continuous block, no cue numbers, no timings. This is the one you want for blog drafts, newsletters, research notes, and LLM prompts.
• SRT subtitles (.srt) — numbered, timed cues in the universal subtitle format. Imports into CapCut, Premiere, DaVinci Resolve, and the subtitle uploaders on LinkedIn, Facebook, and YouTube.
• WebVTT subtitles (.vtt) — the same cues with a WEBVTT header, which is the format HTML5 video requires for a caption track on your own site.

A timed cue looks like this, so you can see exactly what lands in the file:

1
00:00:01,360 --> 00:00:03,040
Here's the hook that got me 2 million views

2
00:00:03,120 --> 00:00:06,480
and the three-part structure behind it

The practical split: copy the on-page text when you are writing, download the SRT when you are editing video, download the VTT when you are publishing to the web.

Most TikTok-to-text workflows end in a writing tool, not a video editor, and timings actively get in the way there. Pasting a timestamped file into ChatGPT means the model spends its attention on cue numbers; pasting it into a CMS means stripping them out by hand.

The on-page output is clean UTF-8 with no hidden markup, so it survives a paste into Notion, Obsidian, Google Docs, or a prompt box without dragging formatting along with it. Most single TikToks fit comfortably in one LLM turn with room left for your instructions — a 60-second video is usually 150–200 words.

That is also why the AI summary exists alongside the transcript. When you batch several videos, the summary synthesises themes across the whole set rather than making you read every transcript in full, which is the difference between reviewing ten competitor videos in two minutes and in twenty.

URL formats. Standard /video/ links, /@username/video/ links, and the shortened share-links the TikTok app copies all resolve correctly — you do not need to expand a share-link first.

Batching. Up to 2 videos per request on the free tier and 10 on Pro, with one combined AI summary across the batch. Free usage is capped at 15 transcriptions per week on videos up to 10 minutes; Pro raises that to 100 per day.

Honest limits. The video has to be public — private and friends-only posts are not reachable. Accuracy tracks audio quality: a creator talking clearly to camera transcribes very well, while heavy background music, several people talking over each other, or thick ambient noise will produce mistakes. Speaker labels are not included, so a duet or a two-person clip comes back as continuous text rather than labelled turns. And a video with no speech at all — a dance clip over a licensed track — has nothing to transcribe.