CommunityCoding & Developmentgithub.com

dueti/media-transcribe

A Claude Code / Codex skill for local Whisper transcription of video/audio, producing corrected transcripts and Markdown.

What is media-transcribe?

This skill enables AI agent users to locally transcribe video or audio files using OpenAI Whisper, producing a corrected transcript and a Markdown file. It is designed for Claude Code, Cursor, Codex, and similar agent platforms, and works on macOS, Linux, and Windows. The user simply provides a media file path or URL; the skill handles downloading (if remote), transcription via a configurable Whisper model (default: 'base'), and post-processing to fix common transcription errors (e.g., technical terms, punctuation). The output includes a plain text transcript and a Markdown version with speaker labels if available, making it ideal for meeting notes, lecture summaries, or content creation. Configuration is managed through environment variables in a `.env` file, offering flexibility in model selection and output directory.

Works with✓Claude Code✓Codex CLI~Cursor
npx skills add dueti/media-transcribe

Installed? Explore more Coding & Development skills: steipete/bluebubbles, steipete/eightctl, steipete/blucli · View all 6 →

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

What does media-transcribe do?

A Claude Code / Codex skill for local Whisper transcription of video/audio, producing corrected transcripts and Markdown.

Related Skills