YouTube Transcript Scraper — Bulk Captions to Text & JSON

What does YouTube Transcript Scraper do?

It downloads the transcript (captions/subtitles) of public YouTube videos, as clean plain text or as timestamped segments, together with the video's title, channel, duration and view count.

• Paste video URLs, Shorts links, youtu.be links, live/embed links or bare 11-character video IDs. Mix them freely; duplicates are removed. Links pasted with trailing punctuation, quotes or several per line (comma- or space-separated) are fine.
• Video links only: playlist and channel links are not expanded; they come back as invalid with a note to paste the video links.
• You only pay for transcripts that are actually returned. Videos without captions, private or deleted videos and invalid links are listed in the results with the reason, free of charge.
• Pick your preferred caption languages in order (e.g. en, de, pt-BR). Human-made captions are preferred over auto-generated ones unless you switch that off.
• No YouTube account, API key or cookies needed.

Why use it?

• Summaries, search and RAG: the text field is a single clean string ready for summarising, embedding or chunking.
• Content research: analyse what competitors or creators say, find quotes, track topics across channels.
• SEO and repurposing: turn videos into blog posts or show notes.
• Academic and media research: study what is said across many videos.

How much does it cost?

Pay per event: $0.0045 per transcript returned, whatever the video's length. No charge for videos without a transcript. Platform usage is included in the price.

Videos with transcripts	Cost
100	$0.45
1,000	$4.50
10,000	$45.00

Set a maximum cost per run in the run options and the actor stops cleanly when it reaches it.

Input

Field	Description
videos (required)	List of YouTube URLs or video IDs.
languages	Preferred caption languages in order. Default ["en"]. If none is available, the first available language is returned (and charged) with languageMatched: false.
onlyPreferredLanguages	Never fall back to another language: a video without your languages comes back as a free language_unavailable row listing what it has. Default false.
preferManualCaptions	Prefer creator-uploaded captions over auto-generated ones. Default true.
outputFormat	both (default), text or segments.
includeMetadata	Add title, channel, duration and view count to each transcript. Default true, no extra cost.

{
  "videos": [
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
    "https://youtu.be/jNQXAC9IVRw",
    "https://www.youtube.com/shorts/oNTaHYXQ4MM"
  ],
  "languages": ["en"],
  "outputFormat": "both"
}

Output

One dataset item per video. Download as JSON, CSV, Excel or HTML, or read it through the Apify API.

{
  "input": "https://youtu.be/jNQXAC9IVRw",
  "videoId": "jNQXAC9IVRw",
  "status": "ok",
  "url": "https://www.youtube.com/watch?v=jNQXAC9IVRw",
  "title": "Me at the zoo",
  "channel": "jawed",
  "channelId": "UC4QobU6STFB0P71PMvOGN5A",
  "durationSeconds": 19,
  "viewCount": 441157415,
  "language": "en",
  "languageMatched": true,
  "isAutoGenerated": false,
  "availableLanguages": ["de", "en"],
  "text": "All right, so here we are, in front of the elephants the cool thing about these guys is that they have really... really really long trunks and that's cool (baaaaaaaaaaahhh!!) and that's pretty much all there is to say",
  "segments": [
    { "start": 1.2, "duration": 2.16, "text": "All right, so here we are, in front of the elephants" },
    { "start": 5.318, "duration": 2.656, "text": "the cool thing about these guys is that they have really..." }
  ],
  "fetchedAt": "2026-10-08T10:00:00.000000+00:00"
}

Videos without a transcript come back like this (not charged):

{ "input": "https://www.youtube.com/shorts/oNTaHYXQ4MM", "videoId": "oNTaHYXQ4MM", "status": "no_captions",
  "url": "https://www.youtube.com/watch?v=oNTaHYXQ4MM", "error": "video has no captions", "text": "" }

status is one of ok, no_captions, language_unavailable (none of your languages, with onlyPreferredLanguages), unavailable (private, deleted, region-locked, members-only, age-restricted, live or upcoming), invalid (not a YouTube video link or ID), blocked or error (temporary YouTube refusal after several retries), or po_token_required (YouTube demands a token this actor doesn't generate). Only ok rows are charged. Rows without a transcript have text: "", so pipelines that read text don't break; filter on status == "ok".

If YouTube refuses more than half the videos in a run of 5 or more, the run ends as Failed (stopping early once that is clear), so a problem never hides behind a "Succeeded" run. A run of 10 or more videos where none has a transcript and almost all got the same answer also ends as Failed, because that usually means YouTube changed something. You are still charged only for the transcripts returned.

Inputs that aren't YouTube links or video IDs are skipped before anything is fetched and listed as invalid (up to 1,000 of them), free.

Runs where almost no video has a transcript (for example a list of music videos without captions) stop early: once more than 100 videos have no transcript and fewer than 1 in 10 checked videos has one, the run ends with a message saying how many videos were not processed. You are never charged for videos without a transcript.

Tips

• Run via the API or schedule it, and connect it to Make, Zapier, n8n or Google Sheets through Apify integrations.
• For many videos, put them all in one run rather than one run per video: it is faster.
• Lists over about 300 videos: start the run asynchronously and fetch the dataset when it finishes. The synchronous API endpoint (run-sync-get-dataset-items, used by some n8n/Make set-ups) gives up after 300 seconds while the run keeps going, and an automatic retry would start a second paid run.
• Results come back in the order videos finish, not input order. Match rows on input or videoId, not on row number.
• Spreadsheets (CSV/Excel): choose outputFormat: text. With segments every line becomes its own column.
• If a run restarts on Apify's side (e.g. a server migration), videos already done are skipped, not charged again.
• Speed is about 80 videos per minute, so the default 24-hour timeout fits about 100,000 videos. For more, raise the timeout in the run options or split the list. If the timeout is near, the run stops cleanly and tells you how many videos were not processed (not charged).
• Auto-generated captions have no punctuation for some languages; isAutoGenerated tells you which kind you got.

Is it legal?

The actor reads only publicly available captions and does not log in. Transcripts can contain the names or opinions of people speaking, and video content is protected by copyright. Make sure your use (e.g. research, analysis, quotation) complies with copyright law, GDPR and YouTube's terms in your jurisdiction. Don't use the transcripts to train AI models on content you have no rights to. You are responsible for any personal data in the transcripts you collect. If in doubt, ask a lawyer.

Feedback

Found a video that should work but doesn't? Open an issue on the Issues tab with the video ID and we'll look into it.