For producing images in which a tiny person is partially inserted in a woman's mouth with the implication that she is either sucking on them or about to eat them. I struggled a bit with the feet-first version - headfirst was much easier, I assume because there's less variation in various pairs of legs sticking out of someone's mouth. We'll see about feet-first.
General usage:
g755uckhf, 1girl, tiny <man/woman/person> <poking/holding <teeth/tongue/saliva/open_mouth> ,
Optional terms (mostly self-explanatory):
tiny <man/woman/person>: I'm not sure if really distinguishes and knows what pair of legs to put, though having one of them is pretty necessary for it to render properly.
<poking/holding>: Poking generally renders the giantess' hand such that she's pushing the tiny person deeper into her mouth. Holding renders her hand such that she's grabbing them.
<teeth/tongue/open_mouth>: If none of these are specified, defaults to a sucking motion. teeth generally makes the teeth more prominent, often making it look like she's gripping the tiny person with her teeth. Tongue and tongue_out, put more emphasis on the tongue, though I find sometimes this can interfere with the anatomy of the shrunken person. open_mouth generally opens the mouth.
upper_body: will generally zoom in such that only the giantess' upper half is visible. I find this more reliable than close-up, which often instead produces inset images.
As usual, check the example images for guidance. Adetailer is your friend here unless the camera is zoomed in on the subject's face. Otherwise, no inpainting was used on the examples.