WhisperX

Use it when

• produce word-level timestamps and speaker-aware transcripts
• align Whisper transcripts tightly enough for subtitles and transcript editing

What it solves

Not the fit when

• diarization model access and GPU memory may be required

Install

pip install whisperx

Invoke

whisperx audio.wav --model large-v3 --diarize

Alternatives

No reviewed alternatives recorded yet.