YouTube Transcript Scraper

Download the transcript (subtitles / closed captions) of any YouTube video — in bulk, as clean text, timestamped segments, and SRT — without a YouTube API key, without cookies, without a browser.

You only pay for transcripts that are actually delivered. Videos with no captions, deleted videos, region-locked or restricted videos cost you nothing. The failure shows up in your dataset as a typed error so your pipeline can handle it, and the price of that failure is $0.

$2 per 1,000 transcripts, and the first ~2,500 a month are effectively free: Apify's free plan includes $5 of monthly credit, and Store Actor charges draw on it. Built for apps that hit YouTube's cloud-server block (LOGIN_REQUIRED / "Sign in to confirm you're not a bot"): the Actor rotates player clients and retries through residential IPs so your server doesn't have to.

What you get for each video

Field	Description
text	The full transcript as one clean string — ready for LLMs, embeddings, search indexing
segments	[{ start, dur, text }] — timestamped segments in seconds
srt	The same transcript as a ready-to-save .srt subtitle file
title, channel	Video metadata
language	The language of the returned track
isAutoGenerated	true when the track is YouTube's auto-generated one
availableLanguages	Every caption track the video offers, so you can re-run with a different language

Videos that can't deliver a transcript return { videoId, url, error, errorType } instead — NO_CAPTIONS, NOT_FOUND, RESTRICTED, EMPTY_TRANSCRIPT — and are not charged.

Input

{
    "videoUrls": [
        "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
        "https://youtu.be/dQw4w9WgXcQ",
        "https://www.youtube.com/shorts/abc123def45",
        "dQw4w9WgXcQ"
    ],
    "language": "en",
    "preferManual": true
}

• videoUrls — watch URLs, youtu.be links, Shorts, live, embed URLs, or bare 11-character video IDs. Duplicates are removed (and only counted once).
• language — preferred ISO code (en, es, de, pt, …). If the video doesn't have that language, the best available track is returned and language tells you which one you got.
• preferManual — human-made captions are usually higher quality than auto-generated ones; the actor prefers them by default and falls back to auto-generated.

Why this one

• Failures are free. You pay per delivered transcript, not per attempt.
• Resilient fetching. Automatic retries with rotating sessions when YouTube throttles; typed errors when a video genuinely can't deliver.
• All three formats in one run — plain text for AI pipelines, timestamped segments for search/chapters, SRT for subtitle workflows.
• No setup. No YouTube Data API key, no OAuth, no cookies, no quota project.

Use it from code

curl -s -X POST "https://api.apify.com/v2/acts/beacondata~youtube-transcript-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"videoUrls": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ"]}'

Works everywhere Apify does: REST API, scheduled runs, webhooks, the Apify MCP server (so AI agents can call it as a tool), n8n, Make, Zapier, LangChain.

Honest limitations

• Private, age-restricted, and members-only videos can't be read (RESTRICTED — free).
• A transcript can only exist if the video has captions on YouTube (NO_CAPTIONS — free).
• language selects among the tracks the video actually has; this actor does not machine-translate.

Automate it with n8n

A ready-made workflow fetches a video's transcript, has GPT-4o-mini write an executive brief (TL;DR, timestamped key points, quotes), and emails it to you. In n8n: Workflow → Import from URL:

https://inmanjeffrey89-cell.github.io/data-tools/workflows/t4-youtube-ai-summary-gmail.json

No community nodes (plain HTTP Request) — works on n8n Cloud and self-hosted. More ready-made workflows.

Support

Found a bug or need a field added? Open an issue on the Issues tab — it's read daily and answered.