AI
The newest branch of the archive, and the one growing fastest. Since 2023 anyone with a
graphics card has been able to make a passable giantess image in a minute, and since
2025 a passable clip. Most of it is bad, some of it is remarkable, and all of it is
arriving in volumes the hand-made side has never produced.
How it fits together. A base model knows what people and streets look
like; a LoRA teaches it the scale relation, which base models get wrong on their own.
The models below are LoRAs. The tools are what you run them in. The creators are the
accounts posting the results, filed under Artists with
an AI label so you can filter them in or out. Generated work is labelled as such
everywhere on this site; it does not replace the drawn and rendered work and is not
listed next to it as if it did.
Models
Tools
┌─┤
Models
├┐
LoRAs: small add-on weights that teach a base model what a giantess looks like, because none of them know on their own - scale cues, shadows and perspective are exactly what they get wrong. All of these are on CivitAI, which also hosts a generator for running them without a graphics card.
2024-
The most downloaded giantess LoRA on CivitAI. Anime-style, built for Illustrious and later Anima checkpoints, and kept current with new versions through 2026. The one to try first if you draw in that style.
CivitAI
Giantess
City Destruction
Classic Giantess
Macro
2023-
periledascent's long-running size LoRA, for women six to thirty-odd feet tall rather than skyline scale. Teaches the model to shrink the background and keep full-body poses, with separate meso and macro versions. Ported across SDXL, Pony, Z-Image Turbo and Flux.2 Klein as each arrived, which makes it the easiest to run on whatever you have.
CivitAI
Giantess
Amazon
Classic Giantess
Mega
Simply Is
2024
Photorealism on SDXL, still workable with anime. Trigger words split the scale in two: ten to thirty feet, and thirty feet up to mountain size. Built on the Minigiantess categories. Version 1.1 lives on its own page because the author could not update this one.
CivitAI
Giantess
Mega
Macro
Giga
City Destruction
2025
A plain giantess LoRA for Z-Image Turbo, the fast open model of late 2025. No trigger word; put "giantess" in the prompt. Simple, and it works.
CivitAI
Giantess
Classic Giantess
City Destruction
2026
Not a one-word wonder, as its author says: a helper for Z-Image Base that fixes what the base model gets wrong - shadows, perspective and the size relation between her and the street. Size triggers at roughly ten, thirty, a hundred and three hundred metres, and an all-in-one version that replaces the separate ones.
CivitAI
Giantess
Scale Tiers
Mega
Macro
Giga
Perspective
2026
A video LoRA, which is the new frontier: female height growth for the open video models Wan 2.2, LTX and MiniMax H3. The author calls it experimental and says even state-of-the-art models struggle with the concept. They do, and it is still the first thing that animates a growth sequence from a prompt.
CivitAI
Female Growth
Gradual Growth
Clothes Ripping
Outgrowing Spaces
2024
A giant girl in a night city, for Pony checkpoints. Triggers GTS and giantess, recommended at about 0.6 strength. Narrow on purpose, and good at the one scene it was trained on.
CivitAI
Giantess
City Destruction
Macro
Unaware
2023
One of the first, from the Stable Diffusion 1.5 era, and still among the most downloaded: a size-difference concept trained on art of a small boy with a giant girl. The author's own notes warn that it is unstable and picky about checkpoints. Here as much for the history as the results.
CivitAI
Giantess
Shrunken Man
Handheld & Palm Play
Gentle
└┘
┌─┤
Creators working with AI
├┐
From the Artists list, everyone with AI among their media.
AI
UK-based, and among the first to post generated giantess video to DeviantArt - a Kling clip of a giantess seated over a motorway in early 2025. Since then mostly stills, with the model named in the title, and a following in the low thousands that the hand-drawn accounts would envy.
DeviantArt
Giantess
Macro
City Destruction
Unaware
Classic Giantess
AI
Formerly steelknight3000, nineteen years on DeviantArt and now entirely generated: super-strength and growth clips, a lot of it futa, cut together into compilations from Seedance and similar video models. The largest watcher count of the AI accounts here, and a useful picture of where the strength and muscle branches are heading.
DeviantArt
Muscle Growth
Female Growth
Amazon
Transformation
AI
Ten-second clips from Sora 2 and Veo 3.1, nearly all of them boots and bare feet coming down on miniature cities and small creatures, every one of them generated - the bio exists mainly to say so. Narrow, prolific, and the clearest example yet of a branch that video models can do on demand.
DeviantArt
Crushing & Destruction
Foot Crush
Shoes & Boots
Trampling
City Destruction
Unaware
└┘