Audio Transcription
Kognar can transcribe your voice into text for both the chat input and the Voice Assistant. Two transcription modes are available: local (on-device) and cloud.
Local transcription (on-device)
Your audio is processed entirely on your computer — nothing is sent to the internet. This is the recommended mode for privacy-conscious use or when you don't have a stable connection.
How it works:
- On first use, Kognar downloads a voice model and saves it to your profile
- All subsequent transcriptions run locally using that model
- The model detects the spoken language automatically — it is not locked to the app's interface language
Advantages:
- Full privacy — audio never leaves your device
- Works offline
- Supports longer recordings than cloud mode
Setting up local transcription
- Open Settings > General > Voice
- Set transcription mode to On-device
- Choose a model size (larger = more accurate, larger download)
- Optionally click Download model now to pre-download before first use
Cloud transcription
Audio is sent to Kognar's cloud service for transcription. This requires an active account and internet connection.
Advantages:
- No download required
- Works immediately without setup
Automatic fallback
If local transcription fails and your account has cloud transcription available, Kognar automatically retries using cloud — so you always get a result.
Choosing a transcription mode
Go to Settings > General > Voice to switch between modes at any time. The setting applies to both voice input in the chat and the Voice Assistant.
| Mode | Privacy | Internet required | Recording length |
|---|---|---|---|
| On-device | Full — stays local | No | Longer |
| Cloud | Audio sent to server | Yes | Shorter |