Transcription modes
Local models, cloud transcription, and how to choose.
Local
Audio is processed entirely on your machine and nothing is sent anywhere. Works with the network off.
| Model | Size | Speed | Accuracy | Best for |
|---|---|---|---|---|
| Tiny | 75 MB | Fastest | Basic | Quick notes and commands |
| Base | 142 MB | Fast | Good | General dictation |
| Small | 466 MB | Moderate | Better | Longer dictation |
| Medium | 1.5 GB | Slower | Great | Detailed transcription |
| Large v3 | 3.1 GB | Slowest | Best | Maximum accuracy |
| Distil Large v3 | ~1.5 GB | Fast | Great | Best speed-to-accuracy ratio |
| Parakeet TDT v3 | ~2.6 GB | Fast | Great | Multilingual, 25 languages auto-detected |
Cloud
Cloud mode is an Enterprise feature. It runs managed Whisper Large-v3-Turbo transcription across 99+ languages with automatic detection, without exposing the underlying provider credential. Audio is uploaded over HTTPS only when cloud mode is selected; local mode remains available.