YouTube Transcript Scraper: LLM-ready transcripts for videos, playlists and channels

Extract existing YouTube captions (manual or auto-generated) as clean text, timestamped segments and optional RAG-ready chunks. Give it video links, playlists or channels, mixed in a single run, and get one dataset row per video with title, channel, views and duration included.

1 result = 1 transcript. $1.00 per 1,000 transcripts. Videos without captions are not charged.

What you get

• Clean text of the whole video (HTML entities and whitespace normalized)
• segments: [{start, duration, text}] with timestamps in seconds (optional)
• chunks: text split into pieces of about N characters, each with a start time (optional, for embeddings and RAG)
• Metadata: title, channel, channelId, duration, viewCount, language, isAutoGenerated, isTranslated
• A run summary (ok / no_transcript / error counts) in the key-value store record OUTPUT

Input

{
  "urls": [
    "https://www.youtube.com/watch?v=jNQXAC9IVRw",
    "https://youtu.be/aircAruvnKk",
    "https://www.youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi",
    "https://www.youtube.com/@TED"
  ],
  "languages": ["en"],
  "includeSegments": true,
  "chunkSize": 1000,
  "maxVideosPerChannel": 20,
  "translateTo": "",
  "maxConcurrency": 5
}

Accepted URLs: watch, youtu.be, shorts, embed, live, bare 11-character video IDs, playlists, channels (@handle, /channel/UC..., /c/..., /user/...). Duplicates are removed.

Field	Meaning
languages	Priority list of language codes. Manual captions are preferred over auto-generated ones in the same language. If none match, the first available caption track is used.
translateTo	Optional language code. Uses YouTube's own caption translation and marks the row isTranslated: true. Not every video allows it.
chunkSize	0 = off. Otherwise adds chunks of about this many characters.
maxVideosPerChannel	Cap per channel or playlist URL (default 50).
proxyCountry	Country of the residential proxy used when YouTube blocks direct requests (default US).

Output (one row per video)

{
  "videoId": "jNQXAC9IVRw",
  "url": "https://www.youtube.com/watch?v=jNQXAC9IVRw",
  "status": "ok",
  "title": "Me at the zoo",
  "channel": "jawed",
  "channelId": "UC4QobU6STFB0P71PMvOGN5A",
  "publishedAt": "2005-04-23",
  "duration": 19,
  "viewCount": 444458891,
  "language": "en",
  "isAutoGenerated": false,
  "isTranslated": false,
  "text": "All right, so here we are in front of the elephants...",
  "segments": [{"start": 1.2, "duration": 2.1, "text": "All right, so here we are"}],
  "chunks": [{"start": 1.2, "text": "All right, so here we are in front of the elephants..."}]
}

status is ok, no_transcript (captions are disabled or absent, not charged) or error (video unavailable or repeated failure, not charged; see error).

Run it from the API

curl:

curl -X POST "https://api.apify.com/v2/acts/<USERNAME>~youtube-transcript-cheap/run-sync-get-dataset-items?token=<APIFY_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"urls": ["https://youtu.be/jNQXAC9IVRw"], "languages": ["en"]}'

Python:

from apify_client import ApifyClient

client = ApifyClient("<APIFY_TOKEN>")
run = client.actor("<USERNAME>/youtube-transcript-cheap").call(
    run_input={"urls": ["https://youtu.be/jNQXAC9IVRw"], "chunkSize": 1000}
)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    if item["status"] == "ok":
        print(item["title"], len(item["text"]))

n8n and Make

• n8n: use the Apify node (Run Actor, then Get Dataset Items), or an HTTP Request node with POST https://api.apify.com/v2/acts/<USERNAME>~youtube-transcript-cheap/run-sync-get-dataset-items, header Authorization: Bearer <APIFY_TOKEN> and the JSON input as the body.
• Make: use the Apify module "Run an Actor" and then "Get Dataset Items", or the HTTP "Make a request" module with the same URL and body.
• Each dataset row is one video, so you can feed text or chunks straight into an LLM or vector-store node.

Limitations

• Only existing captions are extracted. There is no AI speech-to-text, so videos without captions return no_transcript.
• publishedAt is best-effort: it is filled when YouTube returns it cheaply and may be null.
• Private, deleted, age-restricted or region-locked videos return error.
• Translation depends on what YouTube allows for each video.
• This actor only reads public data and never logs in.

FAQ

Which languages work? Any language that has captions on YouTube. Set languages to the ones you prefer, in order.

Are auto-generated captions included? Yes, and isAutoGenerated tells you which kind you got. A manual track is preferred when both exist in the same language.

How am I charged? One event per transcript actually extracted. no_transcript and error rows are free. Standard Apify platform usage is included in the price.

Can I limit my spend? Yes. Set a maximum cost per run in the run options; the actor stops when the limit is reached.