TikTok Transcript Scraper

Get the words spoken in TikTok videos as text. Paste video links and get a timestamped transcript for each video, plus plays, likes, shares, author, caption and music. The actor uses TikTok's own caption track when a video has one, and transcribes the audio with Whisper AI when it doesn't, so videos without captions still get a transcript. No login, no cookies, no API keys.

Who it is for

• Marketers and agencies: study what top creators actually say, and the hooks that open their best videos.
• Brand and social-listening teams: find product mentions in the spoken content, not only in captions.
• AI and RAG pipelines: feed TikTok speech into LLMs, summarizers and search indexes as clean text.
• Researchers: build spoken-content datasets from public videos.
• Creators: repurpose your own videos into posts, scripts and subtitles.

What you get

Field	Example	Notes
url, videoId	https://www.tiktok.com/@renreports/video/7307726351291960622
authorUsername, authorNickname, authorVerified	renreports
createdAt	2023-12-01T20:30:54Z	UTC
playCount, likeCount, commentCount, shareCount, collectCount	66900, 2703, 54, 228, 2927
durationSeconds	139
transcript	[{"start": 0.08, "end": 1.28, "text": "I work in influencer marketing"}, …]	timed segments, or one string (see Transcript format)
transcriptSource	tiktok_captions or ai_whisper	where the text came from
hook	I work in influencer marketing and I look at media kids all day long.	the opening line (segments starting in the first 3 seconds)
language, isAutoGenerated, isTranslated	en, true, false	isTranslated is true when the caption track is TikTok's translation
availableCaptionLanguages	["en", "es"]	TikTok caption tracks on the video
wordCount, hasSpeech, audioSeconds
description, hashtags, musicTitle, musicAuthor, isOriginalSound		the post text and sound
locationCreated, isAd, coverUrl
error, errorReason	no_speech	set when a video could not be transcribed

Videos without speech (music-only dances, silent clips), removed or private videos and photo slideshows are listed with an error code so you can see what happened to every input, and you are not charged for them.

How to use it

1. Paste TikTok videos: full links (tiktok.com/@user/video/…), short links (vm.tiktok.com/…, tiktok.com/t/…), mobile links or bare video IDs.
2. Pick the Transcript source:
   • auto (default): TikTok's caption track when there is one, Whisper AI otherwise.
   • captions: TikTok's captions only, never AI. Videos without captions come back as no_captions rows.
   • ai: always transcribe the audio with Whisper, for example to translate everything to English.
3. Run, then export the dataset as JSON, CSV or Excel, or use the API.

Input options

Option	Default	What it does
videoUrls	–	TikTok video links or IDs
transcriptSource	auto	auto, captions or ai
languages	["en"]	preferred caption languages, in order (for TikTok's caption tracks)
fallbackToAnyLanguage	true	use the video's original-language captions when none match
translateToEnglish	false	English output for non-English speech in AI transcripts
outputFormat	segments	segments, text or both
maxVideos	100	cap for the run
maxAudioMinutes	20	longer videos without captions are skipped (error: too_long) instead of AI-transcribed
proxyConfiguration	none	not needed; the actor switches to Apify proxies on its own when TikTok refuses an IP

Example input

{
  "videoUrls": [
    "https://www.tiktok.com/@renreports/video/7307726351291960622",
    "https://m.tiktok.com/v/7488736374602927402.html",
    "7477698081920748830"
  ],
  "transcriptSource": "auto",
  "outputFormat": "both"
}

Example output

{
  "videoId": "7307726351291960622",
  "url": "https://www.tiktok.com/@renreports/video/7307726351291960622",
  "authorUsername": "renreports",
  "createdAt": "2023-12-01T20:30:54Z",
  "playCount": 66900,
  "likeCount": 2703,
  "durationSeconds": 139,
  "transcriptSource": "tiktok_captions",
  "language": "en",
  "isAutoGenerated": true,
  "hook": "I work in influencer marketing and I look at media kids all day long.",
  "transcript": [
    {"start": 0.08, "end": 1.28, "text": "I work in influencer marketing"},
    {"start": 1.28, "end": 3.36, "text": "and I look at media kids all day long."},
    {"start": 3.36, "end": 5.08, "text": "And here is one mistake that I see"}
  ],
  "wordCount": 512,
  "error": null
}

Scheduling and integrations

Run it from a schedule, Make, Zapier or the API whenever you collect new TikTok links, and send the dataset to Google Sheets, Slack or a webhook. A run interrupted mid-way resumes where it stopped when started again with the same input, without duplicate rows.

Use with AI assistants (MCP)

Apify's MCP server exposes this actor as a tool, so Claude, ChatGPT, Cursor and other MCP clients can call it directly: "Get the transcripts of these 20 TikToks and list the products mentioned". Use outputFormat: "text" for compact input.

FAQ

Do I need a TikTok account or cookies? No. Only public videos are read, logged out.

Where do the transcripts come from? Many TikTok videos carry an auto-generated caption track; the actor returns it as is (transcriptSource: tiktok_captions). For videos without one, the actor runs OpenAI's Whisper (base model) on the audio (ai_whisper). Clear speech transcribes well; music, slang or crosstalk lower accuracy, and sung lyrics are transcribed like speech.

Which languages? TikTok caption tracks exist in many languages and you choose the preferred order. Whisper detects ~100 spoken languages and can translate them to English.

Why does a video say no_speech? Dance clips with only music, silent videos and ASMR without talking have nothing to transcribe. These rows are free.

Can it list all videos of a creator or a hashtag? Not yet: paste video links (or export them from a TikTok profile or search scraper) and this actor transcribes them.

Is this legal? The actor reads publicly available videos without logging in. You are responsible for how you use the data, in particular personal data under GDPR/CCPA and TikTok's terms.

Something failed? Rows carry an error code and reason. If you see many blocked rows, re-run later or set a residential proxyConfiguration, and open an issue on the actor page.