Avatar Studio

Starting scene

Voice Line

Audio lip-sync

Voice Yandex marina Speech input Ready AI DeepSeek V4 Flash Lip-sync Loading

Choose GLB Mixamo rig + ARKit + Oculus visemes

Name Body

Mood View

TTS lang TTS voice

Advanced model config

Baseline JSON Retarget JSON Dynamic bones JSON

Saved models

Compatibility checklist

Binary `.glb` full-body character.
Mixamo-compatible skeleton; root object ideally `Armature`.
ARKit 52 face blend shapes.
Oculus 15 viseme shape keys for lip-sync.
Optional: `mouthOpen`, `mouthSmile`, `eyesClosed`, `eyesLookUp`, `eyesLookDown`.