dk-transcribe-summarize

To transcribe audio and generate a PDF summary, run the provided Python script using uv. The Skill transcribes the MP3, generates 100-word and 400-word summaries, and outputs PDF, HTML, and Markdown documents automatically.

Yes, you can transcribe YouTube videos using yt-dlp and summarize them locally. The Skill supports local mlx-whisper models for transcription and local LLM fallback for summarization, ensuring offline capability.

Yes, ffmpeg and Node.js are required dependencies for the audio transcription workflow. You also need a Brave browser, LM Studio, and an OPENROUTER_API_KEY environment variable to execute the transcription and summarization processes.

Yes, this transcription tool supports both cloud models and local LLM fallback. It uses the OpenRouter API for cloud processing and mlx-whisper for local transcription, offering flexibility in content summarization.

The best way to convert a lecture recording into markdown and HTML documents is using this automated transcription Skill. It processes the audio, generates concise summaries, and outputs formatted HTML and Markdown files for reference.

Python dependency resolution fails if uv is not installed or configured correctly. The transcription script relies on uv to automatically resolve dependencies like mlx-whisper, fpdf2, and mistune required for document generation.