Speech-to-Text via faster-whisper

Engine

Always use faster-whisper as the transcription engine.
faster-whisper is a reimplementation of OpenAI's Whisper using CTranslate2 - up to 4× faster than the original with lower memory usage, while maintaining identical or superior accuracy.

Do not default to openai-whisper, whisper.cpp, or any other engine unless the user explicitly requests it.

When to Apply This Skill

• Transcribing audio files (.mp3, .wav, .ogg, .flac, .m4a, .mp4, etc.)
• Extracting Voice-over (VO) narration from video or podcast files
• Building AI pipelines that ingest spoken content (agents, RAG, subtitle generators)
• Generating timestamped transcripts, SRT/VTT subtitle files, or speaker-diarized output