From what I understand she is pure text to image. I’ve already seen videos of people replicating her in stable diffusion.