podcast-transcribe

You can transcribe podcasts by providing an RSS feed, Spotify episode link, or local audio file. The Skill processes the audio using whisper.cpp and outputs a structured Markdown transcript with headers for title, show, date, and source.

Podcast transcription with this Skill requires whisper.cpp, ffmpeg, and Python dependencies including feedparser and requests. You must install these auxiliary tools to process audio files and parse RSS feeds before generating transcripts.

Yes, the Skill automates transcription of latest episodes from RSS feeds. It parses the feed, downloads audio, transcribes with whisper.cpp, and registers each processed episode in sources.jsonl to track what has already been handled.

Yes, the Skill supports language selection and multiple Whisper model sizes. You can choose different models to balance transcription speed and accuracy based on your requirements and hardware capabilities.

Transcripts are output as Markdown files with a structured header including title, show, date, source, language, and model. The clean transcript body is formatted for publication, archival, search indexing, and content reuse.