Ive got something ive been trying to tackle for a while and im wondering if anyone here has any clue as to how I can make this work. How do I create a commercially usable workflow that can accurately swap faces on ComfyUI? Roop is discontinued, and all other viable methods seem to be using insightface for embeddings, which is not available for commercial use. I dont want to have to train a LoRA on each face I plan to produce images with, what is an alternative?
1 Like