AuViMi stands for audio-visual mirror. The idea is to have CLIP generate its interpretation of what your webcam sees, combined with the words thare are spoken.
Why do you think that https://github.com/Zasder3/CLIP-Style-Transfer is a good alternative to AuViMi