Voice
OpenHuman supports voice interaction, allowing you to communicate with the agent when typing is inconvenient.
Voice Settings
Enable Voice
- Go to Settings → Voice
- Enable Voice Interaction
- Select input/output devices
Configuration Options
| Option | Description |
|---|---|
| STT Language | Target language for speech recognition |
| TTS Voice | Voice selection for synthesized speech |
| Wake Word | Trigger word to activate voice (disabled by default) |
| Microphone | Select input device |
Usage
Voice Input
Press and hold the spacebar to speak, release when done. Your voice will be converted to text and sent to the agent.
Voice Output
The agent's responses will be read aloud via TTS. You can choose whether to enable this feature in settings.
Wake Word (Optional)
When enabled, you can say "Hey OpenHuman" to activate voice input without pressing any keys.
Voice Models
OpenHuman provides multiple voice options:
| Voice | Style |
|---|---|
default | Neutral |
friendly | Friendly, warm |
professional | Professional, formal |
Privacy Note
- Voice input is processed locally
- Audio buffers are not stored
- All processing goes through trusted backend services
Troubleshooting
Speech Recognition Inaccuracy
- Check if the microphone is correctly selected
- Confirm network connection is normal
- Use in a quiet environment
TTS No Sound
- Check output device volume
- Confirm TTS feature is enabled
- Try restarting the application