1. Home
  2. Documentation
  3. transcribe_video

transcribe_video

MCPTube transcribe_video tool: transcribe YouTube videos to SRT and plain text from YouTube captions, or with AI on the Pro plan, right in your AI assistant.

Transcribe YouTube video · REST: POST /v1/transcribe

Returns the transcript as SRT and plain text, right in the response. Uses YouTube’s captions when they exist, manual before auto-generated. With no captions, it transcribes the audio with AI, on the Pro plan only (up to 30 hours of audio a month); on other plans, the call is refused and counts toward your quota.

ParameterTypeDescription
urlrequiredstringURL or ID of a YouTube video.
languageoptionalstring (ISO)Caption language, such as "en" or "pt". Without it, the video’s original captions are used.
forceAiTranscriptiondefault: falsebooleanSkips the captions and transcribes the audio with AI. Pro only.
formatdefault: both"srt" | "text" | "both"What to return: SRT, plain text or both.

Example

Ask your assistant
Transcribe this video and sum it up in 3 bullet points: youtu.be/dQw4w9WgXcQ

Your assistant calls transcribe_video, reads the transcript and answers.

Response

Response
{
  "videoId": "dQw4w9WgXcQ",
  "title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
  "author": "Rick Astley",
  "durationSeconds": 213,
  "source": "youtube-captions",
  "language": "en",
  "languageName": "English",
  "cueCount": 61,
  "aiTranscriptionSeconds": null
}

The response also carries the SRT and the plain text, in the srt and text fields, depending on format. On MCP, your assistant gets the same data.

source tells you where it came from: youtube-captions (manual captions), youtube-auto-captions (auto-generated) or ai-transcription (AI transcription). In that case, the response includes aiTranscriptionSeconds and, on Pro, this month’s aiTranscriptionUsedSeconds and aiTranscriptionLimitSeconds.