Our primary focus has been on enhancing the quality and photorealism of our actors, which we’ve found is most effectively achieved through video data. Currently, we require approximately 5-8 minutes of recorded video footage to create a Custom Avatar.

We recognize the appeal of generating avatars from images alone, and we’re actively working on a model that can be trained with minimal data, such as a single image, while maintaining the same high level of quality and realism. We appreciate your patience and understanding as we continue to improve our offerings.