TikTok Transcript & Hook Extractor: Video Hook Analyzer avatar

TikTok Transcript & Hook Extractor: Video Hook Analyzer

Pricing

from $30.00 / 1,000 video analyzeds

Go to Apify Store
TikTok Transcript & Hook Extractor: Video Hook Analyzer

TikTok Transcript & Hook Extractor: Video Hook Analyzer

TikTok transcript plus what a transcript misses: the text written on the frames, the opening hook and its type, the beat-by-beat structure with timestamps and the call to action. Works on silent and music-only videos. Paste video URLs, get one row per video. No login.

Pricing

from $30.00 / 1,000 video analyzeds

Rating

0.0

(0)

Developer

Jura Moshkov

Jura Moshkov

Maintained by Community

Actor stats

0

Bookmarked

3

Total users

2

Monthly active users

14 hours ago

Last modified

Features

Limited permissions

Share

TikTok Transcript & Hook Extractor: On-Screen Text & Structure

Get the TikTok transcript and everything a transcript misses. For each video you get the spoken words, the text written on the frames, the opening hook and its type, the beat-by-beat structure with timestamps, and the call to action. It watches the video and listens to it in one pass, so it also works on silent and music-only videos. No TikTok account, no cookies, no proxies.

  • Transcript: everything said, verbatim, in the original language.
  • On-screen text: titles, labels, list items and prices burned into the frames, with the second each one appears. Subtitles that only repeat the speech are left out.
  • Hook: the exact opening line (or, for a wordless opening, what the first shot shows), when it ends, and what kind of hook it is.
  • Structure: hook → context → proof → value → call to action, with start and end seconds.
  • Call to action: the exact words and the type (follow, comment, link, buy...).
  • Stats: plays, likes, comments, shares, saves and engagement rate.
  • Flat rows: ready for CSV, Excel, Google Sheets or your LLM pipeline.

What is a TikTok script extractor?

A transcript tells you what was said. It does not tell you that the first frame carried the title "3 wines under $20", that the hook ended at second 3, or that the video closed by asking for a comment. This actor returns the script of a TikTok the way a content strategist would write it down: words, on-screen text, hook, structure and call to action, one row per video.

Use it as a TikTok transcript scraper that also reads the frames, as a TikTok hook extractor for swipe files, or as a video-to-text step in front of an LLM.

What you get per video

FieldExample
Hook"In one of the most star studded GTA radio lineups in its history" · bold_claim · ends at 3.8 s
Transcript"In one of the most star studded GTA radio lineups in its history, Bad Bunny..."
On-screen text"GTA 6 confirms Bad Bunny, Burna Boy and Lana Del Rey | COCOTEO FM | CHAMBER 106.6"
Structurehook 0-3.8s → value 3.8-40.3s → proof 40.3-58.9s → cta 58.9-60.1s
Call to action"What station are you gonna be listening to?" · comment
Format and topicsvoiceover b-roll · GTA 6, Gaming, Music
Summaryone or two sentences
Creator, description, posting date, durationbbcnews, 61 s, 2026-10-08
Plays, likes, comments, shares, saves, engagement rate867400, 85200, 1445, 9109, 5490, 0.1167

Example output row

{
"videoUrl": "https://www.tiktok.com/@bbcnews/video/7694397093380853015",
"hookText": "In one of the most star studded GTA radio lineups in its history",
"hookType": "bold_claim",
"hookEndSeconds": 3.8,
"transcript": "In one of the most star studded GTA radio lineups in its history, Bad Bunny, Burner Boy, Lana Del Rey and Robin will host some of the newly revealed radio stations in GTA 6...",
"transcriptSource": "tiktok_captions",
"language": "eng-US",
"wordCount": 194,
"onScreenText": "GTA 6 confirms Bad Bunny, Burna Boy and Lana Del Rey | VOICE OF MEL RAMSAY BBC Journalist | COCOTEO FM | CHAMBER 106.6",
"onScreenCards": [
{ "start": 0, "text": "GTA 6 confirms Bad Bunny, Burna Boy and Lana Del Rey" },
{ "start": 14, "text": "COCOTEO FM" }
],
"hasBurnedInCaptions": true,
"structureOutline": "hook 0-3.8s → value 3.8-40.3s → proof 40.3-58.9s → cta 58.9-60.1s",
"structure": [
{ "beat": "hook", "start": 0, "end": 3.8, "summary": "The video introduces the star-studded radio lineups in GTA 6." },
{ "beat": "value", "start": 3.8, "end": 40.3, "summary": "Detailed overview of specific hosts and radio stations confirmed for the game." },
{ "beat": "proof", "start": 40.3, "end": 58.9, "summary": "Teasers of featured songs and the new gameplay mechanic of listening on foot." },
{ "beat": "cta", "start": 58.9, "end": 60.1, "summary": "A question to the viewer about their preferred station." }
],
"ctaText": "What station are you gonna be listening to?",
"ctaType": "comment",
"format": "voiceover b-roll",
"topics": ["GTA 6", "Gaming", "Music"],
"summary": "This video details the newly revealed radio stations, celebrity hosts, and upcoming features in GTA 6.",
"authorUsername": "bbcnews",
"durationSeconds": 61,
"postedAt": "2026-10-08T20:30:08.000Z",
"plays": 867400,
"likes": 85200,
"comments": 1445,
"shares": 9109,
"saves": 5490,
"engagementRate": 0.1167
}

How much does it cost?

Pay per result: $0.04 per analyzed video on the Apify Free plan, $0.036 on Bronze, $0.033 on Silver and $0.03 on Gold and above. That is $30-40 per 1,000 videos. One flat price per video: no per-minute meter, no separate charge for the transcript.

You pay only for videos that come back analyzed. A private or removed video, a photo or slideshow post, a video that is too long, or a link that is not a TikTok video costs nothing. The same video pasted twice is billed once. Set Max spend and the run stops on its own before it bills past it.

The Free plan analyzes up to 5 videos per run. Paid plans have no cap.

Only need the words? The TikTok Transcript Scraper returns the transcript alone for $6-10 per 1,000 videos.

How to use it

  1. Paste one or more TikTok video URLs, e.g. https://www.tiktok.com/@bbcnews/video/7694397093380853015. Short share links work too.
  2. Run. A video takes about 10 seconds; two are analyzed side by side.
  3. Export as JSON, CSV or Excel, or pull it over the API.

How do I use it with an AI agent (MCP)?

This actor is available as a tool through the Apify MCP server. An agent can call it with a plain input filled straight from the schema:

{"videoUrls": ["https://www.tiktok.com/@bbcnews/video/7694397093380853015"]}

The output is one flat JSON object per video, with nothing to unwrap before handing it to a model.

Common use cases

  • Hook swipe files: collect the opening lines of high-performing videos, with their type and length.
  • Script writing: reverse-engineer how a format is built (how long the hook runs, where the proof sits, when the ask comes) and write to that outline.
  • Silent and text-only videos: recipes, checklists, price lists and tutorials where the content is on the frames, not in the audio.
  • UGC and ad creative research: see what is actually said and shown, and what the viewer is asked to do.
  • Competitor and niche research: break down the top videos of a niche into rows you can sort and filter.
  • Content repurposing: turn a TikTok into a blog post, newsletter or thread with the on-screen points included.
  • AI and LLM pipelines: feed structured video content into summarizers, classifiers and RAG.

It reads publicly available videos: what anyone sees in the TikTok player without logging in. It does not log in anywhere or use your accounts. The words and images in a video belong to its creator; you are responsible for how you use the output.

FAQ

How do I get the transcript of a TikTok video?

Paste the video URL and run. The full text is in the transcript field. When TikTok has captions for the video they are used word for word; otherwise the audio is transcribed.

How is this different from a TikTok transcript scraper?

A transcript scraper returns the spoken words. This actor also reads the frames and returns the on-screen text, the hook, the structure and the call to action, so a video with no speech still comes back with its content.

Does it work on videos with no speech?

Yes. A music-only or silent video comes back with its on-screen text, hook, structure and summary; the transcript field is empty and transcriptSource is none. When the opening has no words at all, hookType is visual and hookText describes the opening shot.

What are the hook types?

question, bold_claim, curiosity_gap, story, problem, result_first, list, news, visual and other. The label is a fixed set so you can filter and count across videos.

Are the burned-in subtitles included in the on-screen text?

No. Subtitles that repeat the speech are left out, so onScreenText holds only the text the creator designed into the video. hasBurnedInCaptions tells you whether the video had such subtitles.

Which languages are supported?

Any language. The transcript and on-screen text stay in the original language; the summary, format and topics are in English.

How long can a video be?

Up to about 4 minutes. A longer video comes back as a free row that says so.

Do I need a TikTok account, cookies or proxies?

No. There is nothing to log into and no proxy to configure.

Can I analyze every video of a profile or a hashtag?

This actor takes video URLs. Collect the URLs with any TikTok profile or hashtag scraper and paste them here.

Why does a video come back without a breakdown?

It is private, removed, region-locked, a photo post, or longer than the limit. The row says which, and it is not billed.