YouTube Transcript Scraper

Paste a list of YouTube video or Shorts links. Get back the full transcript of every one — as clean text, timestamped lines and timed segments — in the video's original language or the language you choose.

No YouTube API key, no Google account, no quota.

What it's for

• AI and LLM workflows — feed transcripts to ChatGPT, Claude or your RAG pipeline to summarise, answer questions or write show notes.
• Content research — find what top creators in your niche actually say, and how often they mention a product or keyword.
• Repurposing — turn videos into blog posts, newsletters, social threads and quotes.
• Subtitles and accessibility — timed segments with start time and duration for every caption line.
• Monitoring — check what was said about your brand in reviews, interviews and news coverage.

What you get for each video

Transcript	Full text, word count, character count; optional [01:05] timestamped lines
Segments	Every caption line with start time and duration in seconds
Language	Language code and name, human-written or auto-generated, whether it is the video's original language
All languages	Every caption track the video offers, so you know what else is available
Video	Title, channel, channel ID, length, exact view count

Real results from a test run of 42 videos:

Channel	Language	Words	Length
Veritasium	English (original)	8,346	47 min
MrBeast	English (original)	17,462	58 min
BBC News	English (auto-generated)	6,289	32 min
officialpsy	Korean (auto-generated)	194	4 min

Picks the right language for you

Big channels add a dozen translated caption tracks to every video. This actor knows which one is the original and returns that by default — not whichever language happens to come first in the list. Want another language? Set languages to your preferences in order, for example ["es", "en"]; en also matches en-US and en-GB. Human-written captions always beat auto-generated ones in the same language.

Input

{
  "videoUrls": [
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
    "https://youtu.be/IwZVXmQdX1E",
    "https://www.youtube.com/shorts/-62C9R_41Sc"
  ],
  "languages": ["en"],
  "includeSegments": true,
  "includeTimestampedText": false
}

Field	Default	Notes
videoUrls	—	Required. Watch pages, youtu.be links, Shorts, live streams, embeds or bare video IDs. Duplicates are removed.
languages	original	Preferred language codes, in order.
fallbackToAnyLanguage	true	When none of your languages exist, return the original transcript instead of skipping the video.
includeAutoGenerated	true	Accept YouTube's automatic captions when no human-written ones exist.
includeSegments	true	Add timed segments.
includeTimestampedText	false	Add [mm:ss] text lines.
maxConcurrency	5	Videos fetched in parallel.

Speed

Measured: 42 videos from 8 channels in under a minute, every one with a transcript, including hour-long videos with 17,000+ words.

You only pay for transcripts you get

Videos without captions, private, deleted or mistyped links are never written to the dataset, so they are never charged. They are listed with the reason in RUN_SUMMARY.

Limitations

• Only languages the video has. YouTube does not let logged-out visitors use its machine translation, so a video with English captions only cannot be returned in Spanish. availableLanguages shows exactly what exists.
• Videos with captions turned off, age-restricted and members-only videos have no transcript available and are reported as failed.
• Live streams in progress have no transcript until YouTube processes them.

Legal

Reads only what YouTube shows any visitor without logging in. You are responsible for how you use the data, including the creators' copyright in what they say.