YouTube Transcript Scraper — video transcripts, subtitles & metadata

Extract YouTube video transcripts (captions/subtitles) with timestamps in seconds, plus title, channel and thumbnail metadata for every video. One API call per video — no browser, no proxy, no login.

Why this Actor

• Built for AI/LLM workflows — feed transcripts straight into summarizers, RAG pipelines, knowledge bases, chatbots, translation tools and fine-tuning datasets
• All URL forms — watch?v=, youtu.be/, /shorts/, /embed/, /live/
• Automatic language fallback — if your preferred language isn't available, the first available transcript is fetched (can be disabled)
• Timestamped or plain text — full snippet array (start, duration, text) plus a ready-to-use plain-text version
• Per-video error reporting — a broken video never kills the run; you get structured error fields per item
• Fast & cheap — async concurrency, proxy-free, charged per result

Use cases

• 📄 LLM training / RAG — turn videos into searchable text for AI apps
• 🗞️ Content repurposing — blogs, newsletters, show notes from podcasts/videos
• 🔎 SEO & research — keyword mining from thousands of transcripts
• 🌐 Subtitles & translation — caption extraction for dubbing/subtitling workflows
• 📊 Analytics — transcript + metadata for video performance analysis

Input

Field	Type	Default	Description
startUrls	array	—	YouTube URLs or raw video IDs (required)
languageCodes	array	["en"]	Preferred languages in order
strictLanguages	bool	false	Skip videos without a preferred-language transcript
includeTimestamps	bool	true	Output snippets array + timestamped text
includeMetadata	bool	true	Fetch title, channel, thumbnail via oEmbed
maxConcurrency	int	5	Parallel videos per run (1–20)

Output

One dataset item per video:

{
  "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "videoId": "dQw4w9WgXcQ",
  "title": "Rick Astley - Never Gonna Give You Up",
  "channelName": "Rick Astley",
  "thumbnailUrl": "https://i.ytimg.com/vi/dQw4w9WgXcQ/hqdefault.jpg",
  "languageCode": "en",
  "isGenerated": false,
  "snippetCount": 61,
  "transcriptText": "[00:00] We're no strangers to love ...",
  "snippets": [
    { "start": 0.0, "duration": 3.0, "text": "We're no strangers to love" }
  ]
}

Videos without an extractable transcript return an item with an error field — the run continues.

Pricing

Pay-per-event, charged only when a result is produced:

• $0.0007 per video result (dataset item)
• $0.005 per actor start

No hidden charges. Free tier usage is free.

FAQ

Does it work for videos without captions? No — the video must have manual or auto-generated captions (most videos with speech do).

Is a proxy needed? Not strictly — the actor works proxy-free from residential IPs. For datacenter/large-scale runs, enable the Apify proxy (default) or paste your own. If YouTube shows a "Sign in to confirm you're not a bot" error, add a cookies.txt export from a logged-in browser session (Chrome: use a cookies-export extension; Firefox: EditThisCookie → Export Netscape format).

Can I scrape an entire channel/playlist? Feed the actor the video URLs (e.g. export from a playlist link resolver) or run it on your list. Playlist/channel expansion is on the roadmap.

Support

Open an issue on the actor page or contact the developer. Feature requests welcome: channel/playlist expansion, view/like counts, summary generation.