Read the caption track of a connected channel's own video to get its transcript.

To transcribe a YouTube video, read the caption track YouTube already holds for it. Do not download the file and run your own transcription.

Call GET /v1/accounts/{accountId}/youtube-captions with videoId (Get a YouTube video transcript), for one of the connected channel's own videos.

Auto-generated (ASR) tracks count. YouTube serves them to the channel owner, which is what the connected account is.

An uploaded track wins over an auto-generated one for the same language.

const { data: transcript } = await zernio.connect.getYoutubeCaptions({
  path: { accountId: '66b2e19d8c3f5a7e9d0b1c2d' },
  query: { videoId: 'dQw4w9WgXcQ', language: 'en' }
});

console.log(transcript.text);

Response (200), trimmed:

{
  "videoId": "dQw4w9WgXcQ",
  "language": "en",
  "trackKind": "asr",
  "source": "cache",
  "fetchedAt": "2026-08-27T21:09:54.000Z",
  "text": "Hey, this is Mickey. I'm the founder of this portfolio of three websites.",
  "cues": [
    { "start": 1.6, "end": 8.88, "text": "Hey, this is Mickey. I'm the founder of" }
  ]
}

format=srt returns the raw SubRip body in srt instead of cues. text is there either way.

Quota. The first read downloads from YouTube and costs 200 quota units. Zernio stores the result, so source reads youtube once and cache afterwards.

Pass refresh=true only when the captions changed on YouTube, since that spends the quota again.

availableTracks lists every track on the video, so you can request another language.

404. Only videos owned by the connected channel are readable; anything else is a 404. A video with no track in the requested language is also a 404, with code: "captions_not_found".

YouTube generates auto-captions only for videos with recognizable speech. It can take a few hours after upload to publish them, so treat that 404 as "not yet" rather than "never".

contentDetails.caption in YouTube's own API reads false on videos that do have a serving auto-generated track. It is not a usable availability signal.