I’m working on a realtime AI avataar project in which we have to match our AI avataar to a real person, so we need to run Nvidia Omniverse Audio2Face custom blendshapes with live linking transferring the data to unreal metahuman with same blendshape information.