YouTube Transcript Scraper

What does YouTube Transcript Scraper do?

Give it YouTube URLs or video IDs — it returns the full transcript of each video as clean text plus timestamped segments, ready for LLMs, RAG pipelines, subtitles analysis or content research.

• Batch native — pass one video or five hundred in a single run
• All URL shapes — watch, youtu.be, /shorts/, /embed/, /live/ and raw 11-char IDs
• Language control — preference list (e.g. ["pt", "en"]); manual captions win over auto-generated, and the output tells you which one you got
• You only pay for delivered transcripts — failed videos are never charged

Why this scraper?

Transcript actors are notorious for two things: silently returning nothing and charging for it. This one was built the other way around:

1. Every failure produces an explicit error item in the dataset ("error": "video has no captions at all") — never a silent empty run.
2. Failures are free. You are charged per transcript delivered, not per attempt.
3. YouTube bot-checks are retried automatically across proxy sessions and multiple player clients before giving up.

How do I use it?

1. Click Try for free — the prefilled example works as-is.
2. Paste your video URLs/IDs, set preferred languages.
3. Run and download JSON/CSV — or connect the Apify API / MCP server.

How much will it cost?

$2.50 per 1,000 transcripts ($0.0025 each), plus a $0.005 flat start fee. Failed videos cost nothing. The Apify free plan is enough for roughly 1,900 transcripts per month.

⬇️ Input

{
  "videos": [
    "https://www.youtube.com/watch?v=jNQXAC9IVRw",
    "dQw4w9WgXcQ"
  ],
  "languages": ["en", "pt"]
}

⬆️ Output (real example)

{
  "video_id": "dQw4w9WgXcQ",
  "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
  "channel": "Rick Astley",
  "language": "en",
  "is_generated": false,
  "text": "We're no strangers to love You know the rules and so do I...",
  "word_count": 487,
  "segments": [
    { "start": 18.64, "dur": 3.36, "text": "We're no strangers to love" }
  ],
  "segment_count": 61,
  "duration_seconds": 213,
  "scraped_at": "2026-09-17T22:19:00+00:00"
}

Limitations & honesty notes

• Videos without captions cannot produce transcripts — you get a clear error item instead (not charged). No scraper can transcribe what YouTube doesn't have; actors that promise otherwise are running speech-to-text and charging accordingly.
• YouTube bot-checks datacenter IPs. The Actor retries each video across fresh proxy sessions and three different player clients. If your runs still show Sign in to confirm you're not a bot errors, switch the proxy to RESIDENTIAL — transcripts are tiny (a few MB per 1,000 videos), so the proxy cost stays in cents.
• Playlists/channels are not expanded yet (roadmap) — pass video URLs.
• Age-restricted and private videos may be unavailable.

FAQ

Is it legal to scrape YouTube transcripts?

This Actor collects publicly available caption data only — no personal data, no logged-in areas. Review YouTube's terms and your local rules for your use case.

Which caption track do I get?

Your preferred language's manual track when it exists, then that language's auto-generated (ASR) track, then any manual track. The is_generated and available_languages fields tell you exactly what happened.

Something broke

Open an issue in the Issues tab — it reaches us directly and we ship fixes fast (this Actor is canary-tested daily). Reviews can't notify us.

Your feedback

If this Actor saves you time, a review helps other users find it.