TubeScribe

Convert YouTube videos into structured transcripts and audio summaries.

Updated Apr 10, 2026
One-click install
npx skills add https://github.com/Glour/ai-office-v2 --skill tubescribe-glour
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: TubeScribe
Source: https://github.com/Glour/ai-office-v2/tree/main/skills/tubescribe
Command: npx skills add https://github.com/Glour/ai-office-v2 --skill tubescribe-glour

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires summarize, pandoc, ffmpeg, yt-dlp, mlx-audio, kokoro, torch, soundfile, numpy, huggingface_hub, spacy, misaki, num2words, and includes scripts (resource) components.

What problem does it solve?

TubeScribe reduces the manual effort of turning YouTube videos into structured transcripts and audio summaries, enabling easy review and documentation of video content.

Core Features & Use Cases

  • Automatically detects speakers and generates a formatted transcript in DOCX/HTML/Markdown.
  • Creates an audio summary (MP3/WAV) for quick on-the-go consumption.
  • Includes clickable timestamps, video metadata, and a concise summary with key quotes.
  • Analyzes YouTube comments to surface viewer sentiment and best comments.
  • Supports a processing queue for batched videos and background processing.

Quick Start

Drop a YouTube URL to start processing; TubeScribe will process in the background and notify you when the results are ready.

Frequently Asked Questions about TubeScribe

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I convert YouTube videos to transcripts with speaker detection?

You can convert YouTube videos to transcripts with speaker detection by providing a URL to automatically extract dialogue, label speakers, and output a formatted DOCX, HTML, or Markdown document with clickable timestamps. It processes multi-speaker interviews and lectures in the background.

Can I generate audio summaries from YouTube videos for offline listening?

You can generate audio summaries from YouTube videos by processing the URL to extract key points and convert them into MP3 or WAV files using on-device text-to-speech via mlx-audio or Kokoro. This creates concise audio files for quick on-the-go consumption.

Does yt-dlp work for extracting YouTube transcripts and video metadata?

yt-dlp works for extracting YouTube transcripts and video metadata by downloading the necessary audio and subtitle streams, which are then processed with speaker labeling and timestamp generation to produce structured DOCX or HTML documents.

What is the best way to analyze YouTube comments for viewer sentiment?

The best way to analyze YouTube comments for viewer sentiment is to process the video URL, which automatically extracts and evaluates comments to surface viewer sentiment and highlight the most relevant discussions alongside the transcript and summary.

How do I batch process multiple YouTube videos into transcripts and summaries?

You can batch process multiple YouTube videos into transcripts and summaries by adding URLs to the processing queue. The system handles batched videos with background processing and notifies you when the transcripts and audio summaries are ready.

Do I need ffmpeg and torch installed to generate YouTube audio summaries?

You need ffmpeg and torch installed to generate YouTube audio summaries because they handle audio processing and machine learning tasks required for on-device text-to-speech generation via mlx-audio or Kokoro to produce MP3 and WAV files.