About Whisper
Whisper is OpenAI's open-source automatic speech recognition (ASR) model that transcribes audio in 99 languages with remarkable accuracy. It can be downloaded and run locally for free, or accessed via the OpenAI API.
The model handles diverse audio conditions: accented speech, background noise, technical jargon, mixed languages and poor recording quality. It identifies languages automatically, transcribes speech, and translates non-English audio directly into English text.
Multiple model sizes balance speed and accuracy: tiny (39M parameters) runs on any machine, base and small for moderate hardware, medium for good quality, and large-v3 (1.5B parameters) for maximum accuracy. Faster-whisper and whisper.cpp offer optimized implementations.
Whisper is completely free as an open-source download. The OpenAI API charges $0.006 per minute of audio.
This is a basic listing — the summary above comes from the maker and has not yet been reviewed by our editors.
Is this your tool?
Get a full review page with editor scores, pros & cons and pricing — plus a Verified badge and a dofollow backlink.
Upgrade this listing