A style LoRA based on the art of hexteknik (@HexTeknik on X), go check out their work. All credit for the style goes to them.
Chunky cartoon furry style: thick clean lineart, flat colors, curvy characters and very expressive faces. And yes, it also learned the artist's famous round chickens. Just write "round white chicken" or "round black chicken" in your prompt and they show up as perfect little balls with googly eyes.
It also knows some of the artist's recurring characters by name: Kana Rowanberry (the fox girl), Malon Collie, Farm Dog and Noki.
There are versions for Krea 2, MOSS, Anima and Anima 2.9B, so pick the one for your base model and read its notes (each version also has its own).
Krea 2
Start the prompt with @hexteknik, 2D digital illustration, and then describe the scene in full sentences. Strength 1.0.
Tested on Neutrino v2 at 8 steps. For Kana's green eyes, write "emerald green irises with black pupils and white sclera, eyes open".
MOSS
Start the prompt with just @hexteknik, and then describe the scene in full sentences. Extra style words like "2D digital illustration" made it weaker in my tests. Strength 1.0 (up to 1.3 if your checkpoint pulls hard toward its own look).
It works on plain MOSS SFT with no turbo (35 steps, CFG 4), but for the best results use a detailer and a turbo LoRA: the standard 8-step Moss Image Turbo, or FurryPlus Turbo for furry characters.
Anima
Start the prompt with masterpiece, best quality, @hexteknik, and then use regular Danbooru tags. Strength 1.0. Use a finetune, and put text, signature, watermark, english text in the negative so no stray letters from the comics sneak in.
Danbooru tags are taken literally: for male characters use adult, mature male, tall instead of "dog boy".
Anima 2.9B: there's a dedicated v1 (Anima 2.9B) version, remapped to the 40-block layout of the 2.9B models. If your loader already remaps older Anima LoRAs (LoRA Manager does), the regular version works too.
Description
Krea 2 version, trained locally with musubi-tuner on Krea 2 Raw (int8) on a 16 GB RTX 5070 Ti.
Dataset: 100 images, captioned in full sentences. Characters are named when known (Kana Rowanberry, Malon Collie, Farm Dog, Noki, plus a few guests like Isabelle or Krystal), and every chicken is described as a "round white chicken" or "round black chicken", so you can call them on purpose.
Training: 3 repeats x 10 epochs overnight, about 28 min per epoch. I compared every epoch against a no-LoRA control. The style shows up from epoch 3, and the chickens turn into perfect little balls from epoch 5 to 7. Epoch 9 was my pick: the most artist-like faces and the roundest chickens.
How to use
Start the prompt with @hexteknik, 2D digital illustration, and then describe the scene in full sentences. Strength 1.0. Tested on Neutrino v2 (Krea 2 Turbo based) at 8 steps.
Things to know
Eyes: the artist often draws dark, half-closed eyes, so a plain "green eyes" can come out black or neon. For Kana, "emerald green irises with black pupils and white sclera, eyes open" works every time.
It tends to undress characters, since a lot of the source art is NSFW. If you want them clothed, put "fully clothed" at the start of the prompt.
The chickens only show up when you ask for them, they don't invade every image. Sadly.









