Here’s how you can select, import, or fine-tune voices to match your brand, audience, or use case.
Open the Assistant Editor:
At the top right, you’ll see the currently selected voice (like "Emily").
Click the Voice Name:
A menu pops up with available voices.
Select any existing voice to instantly switch your assistant’s sound.
Import or Clone a Voice:
Import: Follow the on-screen instructions or related documentation to bring in a new voice.
Clone: Use the “Clone Voice” feature to create a unique voice.
Tip: Voice changes only apply to the current assistant. Want to use the same custom voice in another? Just repeat the process!
Navigate to Settings in the editor to access detailed voice controls:
Response Speed:
Controls how quickly the AI replies. Default is “1” for real-time. Typically you’ll want to keep it there. Increase for a delay, decrease for more speed.
Sensitivity to Interruption:
Set how much background noise the AI ignores.
For noisy environments (e.g., expos, homes with TVs), reduce to 0.8–0.9.
For quiet settings, leave at the default.
Talking Speed:
Alters how fast the AI speaks. Adjust to match the natural pacing of your chosen voice and language.
Voice Emotion (Temperature):
Sets how expressive the AI sounds.
Lower values (0.9–1.0): More robotic, steady
Higher values (1.15+): More emotional, may sound dramatic
Tip: 1–1.15 is ideal for most voices.
Language Processing:
Define the language the AI understands (e.g., English, Spanish, German). Set this to match your user base.
Speech-Friendly Formatting:
Optimizes how the AI reads out numbers, prices, and phone numbers (English only).
If you notice odd pronunciations, toggle this setting off and on again to reset.
Back Channel Feedback:
Enable or disable natural “ums,” “ahas,” and similar sounds for a more human touch.
Pro Tip: If you notice unusual pitch changes or erratic voice behavior, lower the emotion (temperature) for a smoother, more consistent sound.
