Whisper
Also known as: OpenAI Whisper
Whisper is OpenAI’s open, local speech-to-text model. It’s accurate, multilingual, and runs on your own hardware.
- Transcribes audio to text locally.
- 60+ languages.
- Sizes from tiny (fast) to large (accurate).
The local STT default: if you need speech-to-text that stays on your machine, Whisper is the standard choice — and it pairs well with a local LLM for pipelines.