Whisper large-v3
used in
1
stack
Open-weights automatic speech recognition (STT) model.
OpenAI's open-weights speech-to-text model, widely self-hosted for transcription in voice and meeting-assistant stacks. Open weights let audio be transcribed inside your own environment rather than sent to a third-party API.
Strengths
- Open weights; self-host to keep audio in-environment
- Strong multilingual transcription accuracy
Tradeoffs
- Needs a GPU for low-latency, high-volume transcription
- Diarization (who said what) requires a separate component