How to get a TikTok transcript without an account

You can get a TikTok transcript without a TikTok account. For one video, paste its link into a free transcript page. For many videos, have your agent call the looot endpoint scrapecreators-tiktok-video-transcript with the link. It returns the captions as timestamped WEBVTT text for public videos under two minutes.

People search for "tiktok transcript free no sign up" and "tiktok to text" because they want the words of a video, not another app to log in to. This guide shows the exact command, a real output from a run on 2026-10-13 at 23:18 UTC, what that run cost, how to turn the result into plain text, and where the endpoint stops working.

How do you get a TikTok transcript without logging in?

You send the public video link to a service that reads the captions for you. Nobody logs in to TikTok, so you need no TikTok account. Pick the route by how many videos you have.

The second and third routes need a looot account, which is a different thing from a TikTok account. looot is one key for paid data and action tools, and the transcript call is one of them.

Step by step: a TikTok transcript with one command

Install the CLI and sign in once. `looot init` logs you in if needed and offers to connect Claude Code, Cursor or Codex.

npx looot init

Read the endpoint before you spend. Inspecting is free and shows the inputs, the price and the limits.

looot inspect scrapecreators-tiktok-video-transcript

Then run it with the video link. This is the exact command behind the output below.

looot run scrapecreators-tiktok-video-transcript --input '{"url":"https://www.tiktok.com/@financebronextdoor/video/7670588824619142414"}' --wait=60

The only required input is `url`. Two optional inputs exist: `language`, and `use_ai_as_fallback`. The catalog notes that the fallback costs the provider 10 more credits per call, so leave it off unless the plain call comes back empty. We did not test the fallback for this post.

What the TikTok transcript output looks like

Here is the run, shortened. The transcript is one string in WEBVTT format: a time range, then the words spoken in it.

{
  "status": "completed",
  "endpointId": "scrapecreators-tiktok-video-transcript",
  "result": {
    "id": "7670588824619142414",
    "success": true,
    "transcript": "WEBVTT\n\n\n00:00:00.060 --> 00:00:03.060\nNo.1. Take your monthly income and multiply it by point five.\n\n00:00:03.061 --> 00:00:06.381\nFive. That's the maximum you should aim to spend on essential expenses\n\n...\n\n00:00:44.902 --> 00:00:45.613\nSave for later.\n",
    "credits_charged": 1
  },
  "actualCost": 0.00188
}

The provider answered in under four seconds once the run started. The last cue ends at 45.6 seconds.

Read the numbers in a transcript by ear before you quote them. In this run the captions split one figure across two lines: "point five." ends a cue and "Five." starts the next. A model summarizing the text could read that as two numbers. The endpoint returns caption text, and caption text carries mistakes.

How do you turn a TikTok transcript into plain text?

Drop the timestamps and keep the words. The CLI prints the run as JSON on standard output, so `jq` pulls the transcript out and `grep` removes the header, the time ranges and the blank lines.

looot run scrapecreators-tiktok-video-transcript --input '{"url":"<video link>"}' --wait=60 \
  | jq -r '.result.transcript' \
  | grep -v -e '^WEBVTT' -e '-->' -e '^$' > tiktok.txt

Keep the WEBVTT form when you want a subtitle file. Save the transcript string as it comes to a file ending in `.vtt`. If you work through an agent, ask it for "the transcript as plain text, one sentence per line" and it does the same cleanup.

Which TikTok transcript endpoints are in the catalog?

These three catalog endpoints return TikTok transcripts. They differ in what they read and what they say about misses. Each line below comes from `looot inspect`, read on 2026-10-14.

What does a TikTok transcript cost?

You pay per call, one call per video, from a prepaid balance. The run above reported its cost in `actualCost`. The table below reads the live price of each provider from the looot catalog when the page loads.

To price a batch, multiply by the number of links and add room for misses. What 1,000 lookups cost walks through that arithmetic.

Run it with an agent

Paste this into Claude Code, Cursor or any agent that reads a skill file. It shows the price first and stops at three videos.

For the rest of what an agent can read from TikTok, profiles, video stats and search included, see the TikTok scraper guide and the TikTok platform page.

What this post does not cover

• Videos of two minutes or longer. The Scrape Creators endpoint states a two-minute limit, so long videos need another endpoint or a speech-to-text step.
• Private videos. The endpoints read public videos by link. Nothing here logs in to TikTok as you.
• Downloading the video file. A transcript call returns text, not the video.
• Checking a creator on a timer. That needs scheduled runs, not available yet. Your agent runs the call when you ask.
• Videos without speech or captions. We did not test one for this post, so we cannot say what each provider returns or bills for it.

Last checked 2026-10-14. The output above is from one real run on 2026-10-13 at 23:18 UTC. Endpoint details come from `looot inspect` on 2026-10-14, and prices in the table are read from the live catalog when the page loads.