🎬 YouTube Transcript API — Captions to JSON + LLM-ready text

Get any YouTube video's transcript as clean, structured, AI-ready text. Extract transcripts and metadata from YouTube — single videos, or whole channels and playlists — in any available language. Every result ships as JSON and LLM-ready formats (clean Markdown + token-aware RAG chunks), so it drops straight into ChatGPT, Claude, a vector database, or your analytics pipeline.

No API keys. No monthly lock-in. Pay only for what you extract.

---

⚡ Why this one

• YouTube done right. It solves YouTube's JavaScript/PoToken challenge and consent walls (via yt-dlp + deno) — the stuff that breaks naive scrapers. Use residential proxy for scale.
• LLM-ready out of the box. Not just raw captions — cleaned plain text, paragraphed Markdown, and embedding-ready chunks whose size and token_estimate come from a real tokenizer (tiktoken o200k/cl100k), accurate even for CJK/emoji/code.
• Channels & playlists. Paste a channel or playlist URL → it transcribes the videos (bounded by maxVideosPerList).
• Any language. Requests your preferred language and falls back to the closest available variant.

🚀 Build with it

• "Chat with a YouTube channel" — transcribe a creator's catalog, embed the chunks, ship a RAG bot.
• Research & datasets — talks, lectures, podcasts as clean timestamped text in any language.
• Repurposing — turn a video into a blog post, summary, or show notes from the Markdown.
• Analytics — bulk-pull transcripts for sentiment / topic analysis.

📥 Input (highlights)

Field	Type	Default	Description
videoUrls	array	— (required)	YouTube video, channel, or playlist URLs.
language	string	en	Preferred subtitle language; falls back to the closest variant.
maxVideosPerList	integer	20	Cap per channel/playlist (max 200; run ceiling 1,000).
outputFormats	array	all	json, text, markdown, chunks.
chunkTokens / tokenizer	—	400 / o200k	Real, enforced token budget for RAG chunks.
translateTo	string	—	Optionally also return a translated copy. Best-effort, often unavailable — uses YouTube's auto-translation.
proxyConfiguration	object	Datacenter	Enable residential for reliable YouTube at scale.

{
  "videoUrls": ["https://www.youtube.com/watch?v=UF8uR6Z6KLc"],
  "language": "en",
  "outputFormats": ["json", "text", "markdown", "chunks"],
  "proxyConfiguration": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }
}

📤 Output

One dataset item per video: normalized metadata (title, channel, duration, views, upload date), transcript_found, is_auto_generated, selected_language, plus text, timestamped transcript, LLM-ready markdown, and token-aware chunks (with exact token_estimate). Add translateTo for a translated object alongside the original.

💸 Pricing

Pay-per-result — you pay per video successfully transcribed, no monthly subscription. Free Apify tier available to try it.

❓ Notes

Extraction needs the video to expose a caption track (human or auto-generated); no-caption videos return metadata only (enable Best effort). This is the YouTube-focused edition of a broader Video Transcript Scraper that also covers Vimeo, TED, Dailymotion, and 1,000+ caption-bearing sites — same engine.

Keywords: YouTube transcript API, YouTube captions to JSON, YouTube subtitles download, transcript for LLM, RAG, YouTube channel transcript, playlist transcript, video to text, any language.