If you transcribe at volume (daily meetings, sales calls, episodes), standard Whisper falls short. faster-whisper does the same up to 4x faster with less memory, on your own machine.
Who this is NOT for
It is a Python library, not an app: it exists to be embedded in automated pipelines. If you only transcribe the occasional video, use regular Whisper and keep it simple.
The engine we use in production for all automated transcription. Honest curation note: its last change is over eight months old; we keep it in the Vault because it is a stable, finished piece, not a project that needs weekly news. We re-evaluate it every quarter.
Your first step
Install with pip install faster-whisper, run the README example on one of your audios and compare the time against classic Whisper.
The faster-whisper implementation guide
The project belongs to SYSTRAN and it is free: the link beside asks for nothing. The guide to set it up without getting lost is ours: step by step, common mistakes and a checklist. Leave your email and we will send it.
No spam. Unsubscribe anytime.