BYO CHARACTER
Use any VRM character with AStack. Our server sends 52 ARKit blendshapes from Audio2Face — any compatible model animates automatically.
HOW IT WORKS
01
Audio2Face
Server-side AI inference generates 52 ARKit blendshape values from speech audio in real-time
02
WebSocket
Blendshape values stream to the client at 30fps, synchronized with audio playback
03
VRM Render
Client-side Three.js applies blendshapes to any compatible VRM model — your character, your brand
COMPATIBLE CHARACTERS
VRM
Mixamo Characters
Convert Mixamo FBX to VRM via UniVRM Blender plugin
Mixamo + UniVRM
VRM
Your Custom Model
Any VRM model with ARKit morph targets works out of the box
Custom
START BUILDING
Read the integration guide to add custom characters to your AStack application in minutes.
INTEGRATION GUIDE