YouTube Channel Scraper — AI-Ready Transcripts & Stats

The fastest YouTube channel scraper on the store. Full transcripts, likes, comments, views, descriptions and thumbnails — in ~15 seconds per channel. No API key. No browser. No headless Chrome.

---

How it compares to the alternatives

	This actor	Premium transcript scrapers	Standard channel scrapers	Budget scrapers
⏱️ Speed (7 videos)	~15 seconds	~16–30 seconds	60–120 seconds	2–5 minutes
💰 Price per 1,000 videos	$5.00	$8.00–$15.00	$2.40	$0.50
📝 Full transcript	✅ 17k–45k chars	✅	⚠️ often broken	❌ URL only
👍 Like count	✅	✅	✅	❌
💬 Comment count	✅	❌	✅	❌
📅 Smart date filter	✅ stops early	❌ scans all	❌ scans all	❌ scans all
🔑 YouTube API key	✅ None needed	✅ None	✅ None	✅ None
🩳 Shorts filter	✅ Built-in	❌	❌	❌
🌐 Multi-language transcript	✅	✅	❌	❌

The only actor in this price range that gives you transcripts AND comment counts AND smart date filtering in the same run.

The key advantage: Smart date filtering. Most scrapers scan the entire channel history every run. This actor stops the moment it reaches your cutoff date — so scraping a 500-video channel for the last 7 days costs the same as scraping a 10-video channel.

---

What you get per video

{
  "youtube_id": "jqoFP9QapXI",
  "channel_id": "UC2ojq-nuP8ceeHqiroeKhBA",
  "title": "32 Tricks to Level Up Claude Code in 16 Mins",
  "video_url": "https://www.youtube.com/watch?v=jqoFP9QapXI",
  "thumbnail_url": "https://i.ytimg.com/vi_webp/jqoFP9QapXI/maxresdefault.webp",
  "description": "Full video description with links and timestamps...",
  "transcript": "Hey what's up everyone welcome back. So today we're going to cover 32 tricks...",
  "upload_date": "2026-04-27",
  "view_count": 42625,
  "like_count": 1873,
  "comment_count": 137,
  "length_seconds": 975,
  "type": "video"
}

Every field is live, real data — not estimates, not placeholders.

---

Input

Required

Field	Description	Example
Channel URLs	One or more YouTube channel URLs	https://www.youtube.com/@mkbhd
Oldest Post Date	Only return videos on or after this date (YYYY-MM-DD)	2025-01-01

Optional

Field	Default	Description
Video Type	long	long = full videos only (no Shorts) · short = Shorts only · all = everything
Include Transcript	true	Set false to skip transcripts — even faster and cheaper
Transcript Language	en	Language code (en, fr, de, es, pt…). Falls back to English auto-captions
Sort By	newest	newest · oldest · popular (by view count)
Max Videos	0 (unlimited)	Hard cap per channel — useful for testing

Supported URL formats

All of these work:

https://www.youtube.com/@mkbhd
https://www.youtube.com/channel/UCBcRF18a7Qf58cCRy5xuWwQ
https://www.youtube.com/c/MarquesBrownlee
https://www.youtube.com/user/marquesbrownlee

---

Real benchmark

Channel: @nateherk · Filter: last 9 days · Videos found: 7 · Transcripts: all 7

Pass 1 (channel listing):   2.5 seconds   — no proxy
Pass 2 (video metadata):    4.5 seconds   — parallel, no proxy
Transcripts (all 7):        0.6 seconds   — parallel CDN download
─────────────────────────────────────────
Total:                      ~12 seconds
Cost:                       $0.03

Transcript sizes ranged from 17,000 to 44,780 characters (full spoken content, not summaries).

---

Use cases

Content intelligence Track what competitors publish, analyze transcripts for keywords, topics, and talking points. Run daily on a set of channels and feed the results into your AI pipeline.

AI / LLM training data High-quality spoken-word transcripts at scale. Each transcript is 10,000–50,000 chars of clean, real human speech.

SEO & keyword research Extract descriptions, titles, and transcripts to find the exact language creators use in your niche.

Newsletter & digest automation Pull the week's latest videos from 10–20 channels in a single run. Schedule it and pipe the results anywhere.

Market research & trend detection Track engagement velocity — which videos are getting likes and comments fast — to spot emerging trends before they peak.

Podcast / content monitoring Get alerted when a creator publishes on a specific topic. Filter by date + transcript keywords downstream.

---

Running multiple channels

{
  "channelUrls": [
    "https://www.youtube.com/@mkbhd",
    "https://www.youtube.com/@LinusTechTips",
    "https://www.youtube.com/@veritasium"
  ],
  "oldestPostDate": "2025-04-01",
  "videoType": "long",
  "includeTranscript": true,
  "transcriptLanguage": "en"
}

Each channel is processed in sequence. All results land in one dataset, sorted newest-first per channel.

---

Transcript quality

Transcripts are fetched directly from YouTube's caption CDN — the same source YouTube uses for its own subtitle display. Three-layer fallback ensures maximum coverage:

1. Innertube caption URLs (fastest — direct CDN, no proxy)
2. youtube-transcript-api (fallback for edge cases)
3. yt-dlp subtitle extraction (last resort)

Result: transcripts succeed on every video that has captions enabled, in 0.5–1 second per video.

---

FAQ

Does this need a YouTube API key? No. The actor uses YouTube's internal Innertube API — no quota limits, no key setup, no 403 errors after 10,000 daily requests.

How does date filtering work exactly? Videos on YouTube channels are listed newest-first. The actor fetches in batches of 10 and stops the moment any video in a batch is older than your oldestPostDate. For a channel with 500 videos, a 7-day filter typically fetches 10–20 videos and stops — you never pay for the other 480.

What about private or age-restricted videos? They're skipped silently — they don't show up in results and don't cost you anything.

Can I skip transcripts to go faster? Yes — set includeTranscript: false. Pass 2 alone (metadata + all stats) completes in under 8 seconds.

What counts as "long" vs "short"? Shorts are videos under 60 seconds. long filters them out. all keeps everything including Shorts and live stream replays.

Does it work on huge channels (1,000+ videos)? Yes. A tight date filter (7–30 days) on any size channel will only fetch the videos in that window. maxVideos gives you a hard cap if you need one.

What's the output format? JSON by default. Apify automatically converts it to CSV, Excel, XML or HTML from the dataset view.