whisper-transcribe

Transcribe media files by running whisper with your file path and output directory. The Skill converts mp3, wav, m4a, mp4, webm, and other formats to text using OpenAI's Whisper model, supporting multiple output formats including txt, srt, vtt, and json.

Yes. Place a markdown context file in the same directory as your media file to ground terminology. The Skill reads these context files automatically during transcription to correct technical terms, proper nouns, and jargon, reducing manual proofreading.

Whisper transcription supports mp3, wav, m4a, mp4, webm, and additional audio and video formats. Configure output as txt, srt, vtt, json, or generate all formats simultaneously from a single media file.

Pass the --model and --language parameters when running transcription. The Skill satisfies requirements for model selection and language specification, allowing you to choose the appropriate Whisper model size and target language for accuracy.

Yes. Use the configurable workflow to run transcription_with_context.py with your media files and model settings. The Skill automates context-grounding corrections per directory, applying markdown context files systematically across batches.

Install on macOS via Homebrew or on Linux systems. Cross-platform CLI execution enables transcription workflows on multiple environments without additional configuration beyond initial setup.