YouTube Transcript Scraper

Get the full transcript of any YouTube video or Short, with timestamps, in the original language or translated into any language YouTube supports. Each result also includes the video's title, channel, views, likes, description, thumbnails and publish date.

You only pay for videos that return a transcript. Videos that are removed, private, have no captions, or fail after all retries are skipped and not charged. They are listed in FAILED_VIDEOS in the run's key-value store with the reason.

What you can use it for

• Feed transcripts to ChatGPT, Claude or your own LLM for summaries, Q&A and RAG
• Turn videos into blog posts, newsletters and social posts
• Research competitors, podcasts, lectures and interviews at scale
• Build subtitle, search or translation datasets

Features

• Any URL format: watch?v=, youtu.be, /shorts/, /embed/
• Language choice: pick a language and get the human-made track if one exists, otherwise the auto-generated one
• Auto-translation: if the language you pick has no track, YouTube translates the video's captions for you (isTranslated: true)
• Spoken language by default: with default, you get the transcript in the language actually spoken in the video
• Timestamped and plain text: every segment has startMs, endMs and startTimeText, plus one transcript_only_text string ready for an LLM
• Retries with fresh residential IPs when YouTube rate-limits or asks for a bot check
• Lean output: the raw player response is off by default, so datasets stay small and fast to download

Input

{
  "startUrls": [
    { "url": "https://www.youtube.com/watch?v=qp0HIF3SfI4" },
    { "url": "https://www.youtube.com/shorts/tPEE9ZwTmy0" }
  ],
  "language": "default",
  "includeTranscript": true,
  "includeFullPlayerResponse": false,
  "maxConcurrency": 10,
  "maxRequestRetries": 10,
  "proxy": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }
}

Field	What it does
startUrls	YouTube video or Shorts URLs
language	default for the spoken language, or a code such as en, es, zh-Hans
includeTranscript	Turn off to get metadata only
includeFullPlayerResponse	Adds YouTube's raw player response. Large, off by default
maxConcurrency	Videos processed at the same time
maxRequestRetries	Attempts per video before it is skipped
proxy	Residential proxies are strongly recommended

Output

{
  "transcript": [
    { "text": "How do you explain when things don't go as we assume?", "startMs": 7490, "endMs": 10690, "startTimeText": "0:07" }
  ],
  "transcript_only_text": "How do you explain when things don't go as we assume? ...",
  "transcriptLanguage": "en",
  "isAutoGenerated": false,
  "isTranslated": false,
  "videoId": "qp0HIF3SfI4",
  "title": { "simpleText": "How Great Leaders Inspire Action | Simon Sinek | TED" },
  "author": "TED",
  "channelId": "UCAuUUnT6oDeKwE6v1NGQxug",
  "lengthSeconds": "1115",
  "viewCount": "21503849",
  "likeCount": "512000",
  "publishDate": "2010-05-04",
  "category": "Education",
  "shortDescription": "Visit http://TED.com to get our entire library of TED Talks...",
  "thumbnail": { "thumbnails": [{ "url": "https://i.ytimg.com/vi/qp0HIF3SfI4/hqdefault.jpg", "width": 168, "height": 94 }] },
  "isShortsEligible": false,
  "isLiveContent": false
}

Also included: canonicalUrl, ownerChannelName, ownerProfileUrl, uploadDate, isFamilySafe, isUnlisted, embed, captions (all available caption tracks) and other flags from YouTube.

FAQ

A video was skipped. Was I charged? No. Only results written to the dataset are charged. Check FAILED_VIDEOS for the reason.

The video has no captions at all. It is skipped. YouTube has no transcript to return.

Can I use it from an AI agent? Yes. It works through the Apify MCP server, and transcript_only_text is ready to pass to a model.

More scrapers by the same developer

• YouTube Super Scraper
• YouTube Comments Super Scraper
• TikTok Videos Super Scraper

Questions or a broken video? Open an issue on the actor's Issues tab and it will be looked at quickly.