Whisper — Open Speech-to-Text
Accurate multilingual transcription you can self-host.
OpenAI's open-weight speech recognition model, strong on accented English and dozens of other languages. Run it locally to caption videos or transcribe interviews with no per-minute fee.
What you get
- Model + code — delivered from the official source, always the latest version.
- Licence: MIT / open source. Safe to use in personal and client work.
- No signup, no email wall, no download manager.
