ONE CHARACTER. A FULL RANGE OF FEELINGS.

A little more
expressive.

The same Zoe you know.
13 emotions + 9 speaking tones.
Every expression is a set of numbers.

NEUTRAL

Neutral

LIVE VECTOR

CONVERSATION PERFORMANCE

Avatar turnResponse tone

What should Zoe say?

Speech speed

Ready to rehearseREST

Text rehearsal uses approximate letter-based mouth shapes. Browser voice timing varies; provider-timed visemes can be connected through the included SDK API. Audio starts only when you click Speak aloud.

TRY A MODEL OUTPUT

Detected emotionAvatar behavior

Label-based mapping. Meaning and context can require a different listener response.

THE COMPLETE SET

Every side of Zoe.

READY FOR YOUR SDK

Mount once.
Change the feeling.

One local renderer. Shared geometry. Interruptible transitions. Use the same parameters in your application.

const avatar = createAvatar(element);

avatar.setEmotion('happy-strong', {
  intensity: 0.85,
  durationMs: 180
});

// Or connect your emotion model:
avatar.setDetectedEmotion('fear', {
  mode: 'listener'
});