YouTube Transcript Scraper — Any Language, Playlists, SRT

Get the transcript of any public YouTube video. Paste video links, a playlist or a channel. Pick the languages you want, in order. You get one row per video with the text, the timed lines and a clear status.

What it does

• Reads the same caption tracks the YouTube player shows.
• Picks the first language in your list that the video has. Inside one language, it picks the manual track before the auto-generated one.
• Works with any language code YouTube uses, including Arabic (ar), right-to-left text included.
• Can ask YouTube for its own machine translation when your language is missing.
• Expands playlists and channels into their videos, up to a limit you set.
• Splits the transcript by the chapter timestamps in the description. Each chapter gets its first and last spoken line and a word count. No AI is used.
• Gives every video a status, so one video without captions never fails the run.

Input

Field	What it does
urls	Video links or IDs, playlist links, or channel links (@handle, /channel/UC...).
preferredLanguages	Language codes in order, for example ["ar", "en"]. Default ["en"].
allowAutoGenerated	Use YouTube speech-recognition captions when no manual track exists. Default on.
translateTo	Optional. Output language. A native track is used if it exists; otherwise YouTube translation is requested.
allowTranslation	Ask YouTube to translate when your languages are missing. Default on.
fallbackToAnyLanguage	If the wanted language is missing, return the video's own track. Default on.
outputFormat	segments (timed lines plus text), text, srt, or all.
includeChapters	Chapters with first and last line per chapter. Default on.
maxVideos, maxVideosPerList	Caps on videos, and therefore on cost.

Example:

{
  "urls": [
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
    "https://www.youtube.com/playlist?list=PL2jdOr9oJl6L3SZwO8atLeLOJN52BgPpE"
  ],
  "preferredLanguages": ["ar", "en"],
  "outputFormat": "segments",
  "maxVideosPerList": 20
}

Output

One row per video:

Field	Meaning
status	ok, no_captions, blocked or unavailable.
videoId, url, title, channel, channelId, durationSec	Video details.
language	Language of the returned text.
sourceLanguage	Track it came from. Differs from language only when translated.
isAutoGenerated	true for YouTube speech-recognition captions.
isTranslated	true when YouTube machine translation was used.
segments	[{ "start": 18.64, "duration": 3.24, "text": "..." }], times in seconds.
text	The whole transcript as one string.
srt	SubRip subtitle text, when asked for.
wordCount, segmentCount	Size of the transcript.
firstLine, lastLine	First and last spoken line.
chapters	[{ title, startSec, endSec, firstLine, lastLine, wordCount }].
chaptersSource	description when the video has chapter stamps, fixed_blocks for 5-minute blocks.
availableLanguages	Every track the video has. (auto) marks auto-generated ones.
message	Plain explanation when something differs from what you asked.

Sample row (shortened):

{
  "videoId": "4i7c8g5lA5U",
  "title": "العبقري الذي أضاء العالم | الدحيح",
  "status": "ok",
  "language": "ar",
  "isAutoGenerated": false,
  "isTranslated": false,
  "availableLanguages": ["ar", "ar (auto)", "en"],
  "wordCount": 4435,
  "durationSec": 1483,
  "segments": [{ "start": 7.061, "duration": 2.913, "text": "السلام عليكم ورحمة الله وبركاته،" }]
}

Statuses and what you pay

You pay only for rows with status ok. Other rows are free.

Status	Meaning
ok	Transcript delivered.
no_captions	The video has no captions, or none that match your language and auto-caption settings.
unavailable	Private, removed, age-restricted or region-blocked video, or an invalid ID.
blocked	YouTube refused every proxy tier. Run it again later.

Pricing is per delivered transcript. You can set a maximum cost per run; the run stops cleanly when it is reached.

Good to know

• Translation. YouTube sometimes refuses machine-translation requests. When that happens you get the original track, isTranslated is false, and message says so. Turn off fallbackToAnyLanguage if you prefer no_captions in that case.
• Proxies. The run starts without a proxy. If YouTube refuses requests, it moves to datacenter proxy, then residential proxy, and stays there for the rest of the run. Proxy cost is included in the price.
• Chapters. Chapters come from timestamps in the video description. Videos without stamps are split into 5-minute blocks.
• Run summary. The RUN_SUMMARY record lists counts per status, timing, and requests and bytes per video.
• Live streams that are still live have no final captions yet.