Biggest problem these AI game's have, and I feel like potential dev's don't *initially* realize, is that it's practically impossible to make the same character twice. So these games are forced to be very short stories, with not much variance in art, cuz the reality is that they are unable generate art in various poses without the character looking drastically different or deformed.
This one looks a little better, but you can still definitely notices the draw backs and limitations.
thx for the share tho!
Well i think you can somewhat consistently get the same character multiple times, but not by just using prompts.
Instead you need to use one of the "models", essentially data from an ai that has trained exclusively to make a single character.
Alternativly you can go for celebrities, its not as good as a "ai model", but atleast the face will usualy be consistent, depending on what AI you are using.
That been said, consistency would be better if you make a complete CG using strict poses, rather then trying to prompt the AI hundreds of times for individual artworks.
Think in the way of how most japanese rpgbased h-games have hundreds of images but alot of them are just variations that are then copypasted ontop of a base image.