Split off from Worm Monster Vore for better visibility. IL equivalent LoRA is still on that model.

LoRA for characters getting swallowed by the monster carnictis, which I believe is from one of the new King Kong movies.

I'll probably revisit this at some point with better training data.

Anm_a1: Not thrilled with the consistency of headfirst vs feetfirst, so I'm calling this an alpha. For tags only I've found CFG=3, steps=30 works well and for natural language + tags mixed CFG=5, steps=40 works alright. Triggers are:

Main: carnictis

Direction: headfirst vore, feetfirst vore

Extra creatures: two carnictis, three, carnictis, four carnictis, five carnictis

Misc: carnictis x-section (for internal shots, not fully functional)