Settings → Transcription Models
This screen manages which Whisper model sizes are downloaded — separate from the Whisper.cpp engine itself, which you install in Dependencies.
The model table
| Tiny | 75 MB — Fastest |
|---|---|
| Base | 142 MB — Fast |
| Small | 466 MB — Medium |
| Medium | 1.4 GB — Slow |
| Large (V3) | 2.9 GB — Slowest, best quality |
| Silero-VAD | 2.5 MB — powers the "Voice Activity Detection" option in General settings, not a transcription model itself |
Larger models are more accurate but slower to run. Each row shows a live download-progress bar while installing, then "Installed" or "Not Installed," with a Download or Remove link on the right (Remove uses the same click-to-arm, click-again-to-confirm pattern used throughout Settings).
Manually-added models
If you've manually copied model files into the app's models folder — shown at the bottom of this screen — click the Refresh button in the header to detect them without restarting the app. The models folder path is displayed directly on screen; model files should be named ggml-*.bin.
FAQ
Which model should I start with?
Base or Small are a good balance of speed and accuracy for most content. Reach for Large if accuracy matters more than speed, or Tiny for quick drafts.
Do I need to install every model?
No — install only the sizes you actually plan to use. You can add more later.
Downloading is disabled — why?
Whisper.cpp itself needs to be installed first, in Settings → Dependencies.