YouTube Transcript Scraper: captions and subtitles to text

Get the transcript of any YouTube video, Short or live replay, or of the videos in a playlist or channel, as clean text. Paste links and get one row per video with:

• the full transcript as one plain-text block, ready for an LLM, search index or document,
• the same transcript as timestamped lines ([01:05] and that's the idea),
• optional timed segments (start and duration in seconds) for subtitles, clipping or search,
• the video's title, channel, views, duration, description, keywords and thumbnail,
• every caption language the video has, so you know what else you can request.

No YouTube API key, no Google account, no browser. The actor's event fee applies only to transcripts you actually get.

Why this one

• Human captions first. When the creator uploaded captions, you get those. Otherwise you get YouTube's auto-generated captions, joined into clean sentences instead of word fragments. Prefer auto-generated? Flip one switch.
• Any language, any regional variant. Ask for en and you also get en-GB or en-US; ask for es and you get es-419 if that's what the video has. If your language isn't available, you get the video's default captions and the language field tells you which.
• Every link format. watch?v=, youtu.be/, /shorts/, /embed/, /live/, music.youtube.com, or the bare 11-character ID.
• Playlists and channels. Paste a playlist or channel link (@handle or /channel/UC…) and get transcripts of its videos, with hard caps on videos and pages so a big channel can't run away with your budget.
• No charge for failures. Videos without captions, private or removed videos and invalid links are listed in FAILURES and aren't charged a transcript fee.
• Fast. 10 videos in about 18 seconds, 5 in parallel by default.

What you can use it for

• AI and RAG: feed lectures, podcasts, interviews and tutorials into ChatGPT, Claude or a vector database.
• Summaries and notes: turn long videos into articles, newsletters, show notes or study notes.
• Content repurposing: blog posts, social posts and quotes from your own channel's videos.
• Research: analyse what creators, brands, politicians or competitors say, across many videos.
• SEO and accessibility: publish transcripts next to embedded videos.
• Subtitles and clipping: timed segments map every line to the second it's spoken.

Input

Field	Example	Notes
Videos	https://www.youtube.com/watch?v=aircAruvnKk	URLs or IDs, one per line. Duplicates are removed.
Playlists and channels	https://www.youtube.com/playlist?list=UU_x5XG1OV2P6uZZ5FSM9Ttw	Playlist, @handle or /channel/UC… URLs. See below.
Max videos per run	100	Unique videos in total, from both fields. Up to 10,000.
Max playlist/channel pages	10	YouTube listing pages read in total. Up to 100.
Preferred language	en, es, pt-BR	Falls back to the video's default captions when unavailable.
Prefer auto-generated captions	off	On = speech-recognition captions even when human ones exist.
Include timed segments	off	Adds segments: [{ start, duration, text }].
Videos in parallel	5	Up to 20.

Fill in Videos, Playlists and channels, or both. If you paste more than 100 video links, raise Max videos per run.

Playlists and channels

• Playlists: youtube.com/playlist?list=… links. A watch?v=…&list=… link in Videos is just that one video; put it in Playlists and channels to get the whole playlist.
• Channels: youtube.com/@handle (or /@handle/videos) and youtube.com/channel/UC… links, using the channel's Videos tab only (no Shorts, live or podcast tabs). Example: https://www.youtube.com/@GoogleDevelopers. A handle is looked up once per run, with up to 4 tries; looking it up is free and doesn't use up Max playlist/channel pages. If YouTube won't resolve a handle, it shows up in FAILURES and the rest of the run carries on; paste the /channel/UC… link instead. Old /c/ and /user/ links aren't supported.
• Order and duplicates: videos you pasted come first, then each playlist or channel in the order you listed them. A video found twice is fetched once.
• Caps: expansion stops at Max videos per run or Max playlist/channel pages, whichever comes first. A playlist page holds up to about 100 videos and a channel page about 30, so with the defaults you get up to 100 videos. It reads what YouTube lists; it doesn't promise every video of a large channel or a particular order.
• Report: SOURCE_EXPANSION in the key-value store shows, for each playlist or channel, the pages read, videos added, whether a cap stopped it, and any error or notice from YouTube. Playlists or channels that fail are also listed in FAILURES, next to videos without captions.

Output

{
  "videoId": "jNQXAC9IVRw",
  "url": "https://www.youtube.com/watch?v=jNQXAC9IVRw",
  "title": "Me at the zoo",
  "channelName": "jawed",
  "channelId": "UC4QobU6STFB0P71PMvOGN5A",
  "channelUrl": "https://www.youtube.com/channel/UC4QobU6STFB0P71PMvOGN5A",
  "durationSeconds": 19,
  "viewCount": 445773513,
  "isLive": false,
  "description": "...",
  "keywords": ["me at the zoo", "jawed karim", "first youtube video"],
  "thumbnail": "https://i.ytimg.com/vi/jNQXAC9IVRw/hqdefault.jpg?...",
  "language": "en",
  "languageName": "English",
  "isAutoGenerated": false,
  "availableLanguages": [
    { "code": "en", "name": "English", "isAutoGenerated": false },
    { "code": "de", "name": "German", "isAutoGenerated": false }
  ],
  "transcript": "All right, so here we are, in front of the elephants the cool thing about these guys is that they have really... really really long trunks and that's cool ...",
  "transcriptWithTimestamps": "[00:01] All right, so here we are, in front of the elephants\n[00:05] the cool thing about these guys is that they have really...\n...",
  "wordCount": 39,
  "scrapedAt": "2026-10-22T10:08:42.164Z"
}

With Include timed segments on, each row also has segments: [{ "start": 1.2, "duration": 4.08, "text": "All right, so here we are, in front of the elephants" }, ...].

Use it from code

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("santiagomv/youtube-transcripts").call(run_input={
    "videos": ["https://www.youtube.com/watch?v=aircAruvnKk"],
    "language": "en",
})
for video in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(video["title"], video["wordCount"])
    print(video["transcript"][:500])

For a playlist or channel, pass "sources": ["https://www.youtube.com/@GoogleDevelopers"] and set "maxVideos".

More examples (Python, Node.js, cURL, n8n) are in the code examples repository, including a playlist and channel tutorial.

Which YouTube actor do I need?

• Captions as text: this actor. It only works for videos that have captions available; videos without them go to FAILURES.
• Just the video list (URLs, titles, views, durations, approximate upload dates): YouTube Channel Videos is cheaper per video. It reads the channel's Videos tab only, and views and dates are rounded or approximate as YouTube lists them.
• Comments and replies: YouTube Comments.

All three accept channel and playlist links directly.

Pricing

$4 per 1,000 transcripts ($0.004 each), charged as an actor event for each transcript returned. Listing the videos of a playlist or channel and looking up @handles aren't charged as events, and neither are videos without captions, private, removed or age-restricted videos, failed playlists or channels and invalid links. Your run's maximum cost limit is always respected.

FAQ

Does every video have a transcript? No. Most videos in common languages have auto-generated captions, but some creators turn captions off, and music videos often have none. Those videos show up in FAILURES and aren't charged a transcript fee.

Can it transcribe audio itself? No. It reads the captions YouTube already has. It doesn't download audio or run speech recognition.

Can it translate? Not in this version. Ask for a language the video offers (see availableLanguages), or translate the text afterwards.

Do I need a YouTube API key or proxies? No. Proxies are included in the price.

Channels and playlists? Yes, with limits: playlist links and channel links (@handle or /channel/UC…, Videos tab only), capped by Max videos per run and Max playlist/channel pages. 100 transcripts cost $0.40; set a run cost limit for big channels. /c/ and /user/ links aren't supported; open an issue if you need them.

Is this allowed? Read this before using it. This actor collects publicly available captions without logging in. Automated access may be against YouTube's Terms of Service, and YouTube may block or rate-limit the IP addresses or accounts involved. Transcripts are the copyrighted work of their creators. You are responsible for how you use this actor and the data, including copyright, personal data and YouTube's terms.

This actor is not affiliated with, endorsed by or sponsored by YouTube or Google LLC. YouTube is a trademark of Google LLC.