Transcription
Voquill supports three transcription modes. You can switch between them at any time from the settings page.
Local mode runs transcription entirely on your device using Whisper. Nothing leaves your machine.
- On first use, Voquill downloads a Whisper model (~142 MB for the default
basemodel). - Models are stored in your app data directory under
models/. - GPU acceleration is used automatically when available (Metal on macOS, Vulkan on Windows/Linux).
- You can force CPU-only inference by disabling GPU in settings.
Local mode is ideal when privacy is a priority or when you don’t have a reliable internet connection.
API mode sends your audio directly to Groq’s Whisper API (whisper-large-v3-turbo) for transcription. This requires a Groq API key, which you can add in settings.
- Your API key is encrypted and stored locally.
- Transcription quality is generally higher than the local
basemodel since it uses a larger model. - Requires an internet connection.
Cloud mode routes audio through Voquill’s cloud service, which handles the Groq API call on your behalf. This is the simplest option — no API key needed, just sign in with your Voquill account.
Choosing a Mode
Section titled “Choosing a Mode”| Consideration | Local | API | Cloud |
|---|---|---|---|
| Privacy | Best | Good | Good |
| Accuracy | Good | Best | Best |
| Internet required | No | Yes | Yes |
| API key required | No | Yes | No |
| Account required | No | No | Yes |