YouTube Transcript Scraper - Subtitles to Text for AI

Turn YouTube videos into text. Paste a list of video URLs and get the full transcript of each one — as plain text, as timestamped segments, and optionally as ready-to-embed chunks for AI and RAG pipelines — together with the video's title, channel, length, views and publish date.

• Bulk: thousands of videos per run, fetched in parallel
• Language choice: ask for en, de, tr in order of preference, with an experimental machine-translation option
• Built for AI: one switch splits the transcript into overlapping chunks, each with its start time
• Honest billing: videos with no captions are reported separately and never charged

Who uses it

• AI and RAG pipelines — feed video knowledge into a vector database
• Content teams — repurpose videos into articles, summaries and newsletters
• Researchers — analyse what is said across hundreds of videos
• SEO — turn talks and tutorials into indexable text

Example output

{
  "videoId": "aircAruvnKk",
  "url": "https://www.youtube.com/watch?v=aircAruvnKk",
  "title": "But what is a neural network? | Deep learning chapter 1",
  "channel": "3Blue1Brown",
  "language": "en",
  "isAutoGenerated": false,
  "isTranslated": false,
  "characterCount": 18430,
  "segmentCount": 286,
  "transcript": "This is a 3. It's sloppily written and rendered at an extremely low resolution…",
  "channelId": "UCYO_jab_esuFRV4b17AJtAw",
  "durationSeconds": 1120,
  "viewCount": 24390646,
  "publishedAt": "2017-10-05T08:11:25-07:00",
  "category": "Education",
  "availableLanguages": ["ar", "de", "en", "es", "fr", "hi", "tr"],
  "segments": [
    { "start": 4.2, "duration": 1.8, "text": "This is a 3." }
  ],
  "chunks": [
    { "start": 4.2, "end": 112.7, "text": "This is a 3. It's sloppily written…" }
  ]
}

Input

{
  "videos": [
    "https://www.youtube.com/watch?v=aircAruvnKk",
    "https://youtu.be/jNQXAC9IVRw",
    "dQw4w9WgXcQ"
  ],
  "languages": ["en"],
  "chunkSize": 1000,
  "chunkOverlap": 200
}

Field	What it does
videos	Watch, youtu.be, Shorts, embed or live URLs, or bare 11-character IDs.
languages	Preferred caption languages in order. Empty means the language spoken in the video.
allowAutoGenerated	Accept YouTube's speech-recognition captions. On by default — most videos have nothing else.
translateIfMissing	Ask YouTube to machine-translate when your language is not available. Experimental; see the notes.
includeTimestamps	Every caption line with its start time and duration.
includeVideoDetails	Channel, length, views, publish date, category, keywords, description.
chunkSize, chunkOverlap	Split the transcript into overlapping blocks for embedding. 0 disables it.

Pricing

Event	Price
Transcript returned	$0.004
Actor start	$0.00005 per run

$4 per 1,000 transcripts, residential proxies included. A video with no captions in your language costs nothing — it is listed under Videos without a transcript with the reason and the languages that do exist. Set Maximum cost per run to cap your spend.

Notes and limits

• Captions must exist. This reads the subtitles YouTube already has; it does not transcribe audio. If a video has no captions at all, no charge is made.
• Auto-generated captions have no punctuation and misheard words are common. isAutoGenerated tells you which kind you got, so you can filter.
• Proxies are on by default. YouTube answers datacenter IPs with a "sign in to confirm you're not a bot" page, so runs use rotating residential proxies. The cost is included in the price above.
• Translation is experimental. It is YouTube's own machine translation, and at the moment YouTube refuses most such requests (in our tests, all of them). Captions that already exist in your language are unaffected. A refused translation is reported with the reason and not charged; isTranslated marks the rows that did get translated. If you need a guaranteed translation, take the original transcript and translate it yourself.
• Private, deleted, members-only and age-restricted videos cannot be read, and are reported with the reason.
• Live streams only have captions once the recording is processed.

FAQ

Can I hand a channel or playlist URL to it? Not yet — it takes individual videos. Pair it with a channel scraper to get the video list first, then feed the IDs here.

Which languages can I ask for? Any language the video has captions in; availableLanguages in each row shows what existed. With translateIfMissing on, YouTube is asked to translate into your first language; see the note on translation above.

Can I use it from my own code or an AI agent? Yes. The API tab has JavaScript, Python and cURL snippets, and the MCP tab exposes it to AI agents.

Something is wrong or missing. Open an issue on the Issues tab with the video URL.