/vt/ - Virtual Youtubers

How do people go about prompting two chuubas into the same image nowadays?
Tried latent couple + composable lora, but it feels like the latter just doesn't do anything (images come out visibly "overcooked" and the loras seem to blend into each other pretty heavily).
Gave regional prompting a shot as well, but the results there make even less sense.
At this point I'm not sure if I'm messing up the settings somewhere, or if this only works on specific versions, or if it's just not compatible with my stupid m1 setup.
Any hint would help.
Pic (mostly) unrelated.