TikTok Transcript Scraper — no AI, no waiting

TikTok Transcript Scraper — no AI, no waiting

Transcripts and subtitles for TikTok videos: plain text, timestamped segments, VTT and SRT. Give it video URLs, or a creator's handle to transcribe their recent videos in bulk.

No login, no cookies, no browser — and no speech-to-text.

Why there is no AI in this

TikTok already publishes a caption track for most videos. This Actor reads that — the real subtitle file, the one the app shows — instead of running audio through a model.

That changes three things you can measure:

speed	a fetch and a parse, not a transcription queue
cost	no per-minute or per-second inference meter
fidelity	it is TikTok's own text, not a second-hand guess at it

Where AI transcription is genuinely needed is the videos with no caption track — and this Actor does not pretend otherwise. It returns a row saying so.

Coverage, stated honestly

Captions are common but not universal, and coverage varies by creator:

sample	videos with a caption track
26 videos across search results and one creator	21 — 80%
@nasa's 10 most recent	6 — 60%

Of the tracks found, most were auto-generated by TikTok (19 of 21) and a few were written by the creator; the row tells you which via isAutoGenerated.

Videos with no captions are not charged for. You still get a row — hasTranscript: false with a note — so you can see which inputs produced nothing rather than finding a short dataset and wondering. Charging for a missing transcript would be charging for the absence of the product.

How to use

1. Click Try for free at the top of this page. Apify creates a free account if you do not have one — no card is needed to run this.
2. Put what you want scraped into Video URLs — one per line.
3. Press Start.
4. Rows show up in the Dataset tab while the run is still going. Export to JSON, CSV, Excel or XML from there, or read them over the API.

Videos with no captions still get a row — hasTranscript: false and a note — and are not charged for.

Input

Use either or both.

field
videoUrls		Full video URLs or bare numeric IDs
profiles		@handles, profile URLs or user IDs — transcribes their recent videos
maxVideosPerProfile	20	Only used with profiles
languages	all	Two-letter codes. Empty means the video's original track
includeSegments	on	{start, end, text} per cue
includeSrt	on	The same track as SRT
includeRawVtt	off	TikTok's original file, unmodified

Every field, with its default — paste the whole block and delete what you do not need:

{
  "videoUrls": [
    "https://www.tiktok.com/@tiktok/video/7283509500307442987"
  ],
  "profiles": [],
  "maxVideosPerProfile": 20,
  "languages": [],
  "includeSegments": true,
  "includeSrt": true,
  "includeRawVtt": false
}

Transcribing a whole creator is the usual reason people want this — hook analysis, script study, or feeding a model something better than a caption you paid to regenerate.

Output

field
videoId, url, authorUsername, authorNickname, description, createdAt, durationMs
hasTranscript	the field to filter on
text	the whole transcript as one string
segments[]	{start, end, text} — timestamps as HH:MM:SS.mmm
srt	ready to save next to a video file
vtt	TikTok's original, when asked for
language, isAutoGenerated, isOriginalLanguage, availableLanguages[]
wordCount, segmentCount
source	url, or creator:<handle> — so mixed runs stay readable
note	why a transcript is missing, when one is

One row, exactly as it lands in your dataset (long signed CDN links shortened here for readability):

{
  "videoId": "7666060020596968717",
  "url": "https://www.tiktok.com/@velvetgloww1/video/7666060020596968717",
  "source": "url",
  "authorUsername": "velvetgloww1",
  "authorNickname": "beauty tips !",
  "description": "✨ Green Tea for Your Face 😍 I Tried This Unexpected Beauty Hack!",
  "createdAt": "2026-07-24T11:48:17.000Z",
  "durationMs": 12534,
  "availableLanguages": [
    "en"
  ],
  "hasTranscript": true,
  "language": "en",
  "isAutoGenerated": true,
  "isOriginalLanguage": true,
  "segmentCount": 5,
  "wordCount": 50,
  "text": "I just wanted to wash my dirty face when I spilled my green tea instead of wasting it I rinsed my face with the cooled green tea later my skin felt refreshed and looked calmer ever since then I've bee …",
  "segments": [
    {
      "start": "00:00:00.020",
      "end": "00:00:03.180",
      "text": "I just wanted to wash my dirty face when I spilled my green tea"
    },
    {
      "start": "00:00:03.181",
      "end": "00:00:06.901",
      "text": "instead of wasting it I rinsed my face with the cooled green tea later"
    },
    {
      "start": "00:00:06.902",
      "end": "00:00:08.781",
      "text": "my skin felt refreshed and looked calmer"
    },
    "…"
  ],
  "srt": "1\n00:00:00,020 --> 00:00:03,180\nI just wanted to wash my dirty face when I spilled my green tea\n\n2\n00:00:03,181 --> 00:00:06,901\ninstead of wasting it I rinsed my face with the cooled green tea later\n …",
  "note": null
}

Notes

• url is TikTok's own canonical link. A photo post therefore points at /photo/ and a video at /video/, and a row whose author handle is missing gets null rather than a link that would 404.
• A video with several caption tracks (original plus translations) returns the original unless you name a language in languages.
• Duplicate inputs are only fetched once, including a video that appears both in videoUrls and in a creator's recent posts.
• Only public data. There is no login, so nothing hidden from a logged-out visitor is reachable.

Run it from your own code

Every Actor is an API endpoint. This starts a run, waits for it, and returns the rows in one call — swap in your token from Settings → Integrations:

curl -X POST 'https://api.apify.com/v2/acts/mu0i~tiktok-transcripts/run-sync-get-dataset-items?token=YOUR_TOKEN' \
  -H 'Content-Type: application/json' \
  -d '{
    "videoUrls": [
      "https://www.tiktok.com/@tiktok/video/7283509500307442987"
    ],
    "profiles": [],
    "maxVideosPerProfile": 20,
    "languages": [],
    "includeSegments": true,
    "includeSrt": true,
    "includeRawVtt": false
  }'

Pricing

Pay per result, and here "result" means a transcript, not a row — a video that turned out to have no captions still gets a row explaining that, and it is free. Nothing is billed for time or compute.

Event	Per 1,000	What one unit is
transcript	$4.00	one transcript in the dataset — rows with hasTranscript: false cost nothing

Paid Apify plans pay less than this. The table above is the free-plan rate. Starter, Scale and Business each get a lower per-unit price — up to 55% off on Business — and the Actor's Pricing tab shows every tier before you run anything.

Starting a run costs $0.00001 — a cent per thousand runs — and there is no run minimum. You are never billed for a run that found nothing. If an input has nothing to collect the run says so in a row of the dataset — never a silently empty result, and never a charge. A run that fails because the data could not be reached fails loudly, because that one is worth retrying.

Related Actors

Fourteen Actors covering the whole public TikTok surface. Same billing rules and the same guarantees throughout: billed once for each row you receive, never billed for a run that found nothing, and told in the dataset why it found nothing.

Actor	What it does
TikTok Search Scraper — videos, users, hashtags	Keyword search across every tab, deduplicated.
TikTok Hashtag Scraper — videos by hashtag	Every video under a hashtag, walked deep.
TikTok Trending Scraper — the real Explore feed	The app's real Explore feed, not Creative Center.
TikTok Profile Scraper — followers, likes, bio	Followers, likes, bio for any public account.
TikTok User Posts Scraper — every video + stats	Every video an account posted, with stats.
TikTok Comments Scraper — replies, no duplicates	Every comment and reply, no duplicate pages.
TikTok Video Downloader — no watermark MP4/MP3	No-watermark MP4, MP3, covers and subtitles.
TikTok Sound Scraper — videos using a sound	Videos that actually use a sound, filtered.
TikTok Live Scraper — rooms, viewers, streams	Live rooms with playable stream links.
TikTok Places Scraper — locations and reviews	Locations with addresses, coordinates and reviews.
TikTok Shop Scraper — Products, Sellers & Reviews	Shop search, product detail, seller catalogue.
TikTok Shop Category Scraper — every product	Every product in a Shop category and its subcategories.
TikTok Shop Reviews Scraper — full archive	The full review archive for a Shop product.