General description
Style LoRA that turns any prompt into a climax shot.
Sharp, jagged edges everywhere: angular shapes, spiky ink lines and hard diagonals over dry painterly brushwork, back-lit dramatic lighting, fire against deep shadow, and a camera that never stands still.
Most style LoRAs change how the picture is rendered and keep the base model's composition. This one also re-stages the scene: lower angles, stronger foreshortening, the monster leaning over the hero. Same prompt, same seed, a different film.
Monsters are its specialty. Dragons, kaiju and demons come out huge, toothy and scary. Put a tiny figure next to them for scale.
Shines in wide formats. The source art is mostly cinematic 21:9, so 1536×640 and 1344×768 give the strongest results. Open the wide images to see them uncropped.
Available for two bases — pick the version that matches your model:
Anima — trigger @snkbrush
Krea 2 — trigger @snkbrush. (new — natural-language prompts)
This is a style, not a character — name the character you actually want.
Inspired by the art of @sonic1nstinct. Not an official model by the artist; please support the original artist.
──────────────────────────────────────────────────────
About this version — Anima
Trigger: @snkbrush at the very start, followed by a comma, then regular Danbooru tags. No style tags needed.
Prompt template:
@snkbrush, safe, masterpiece, best quality, score_7, <your tags>
Example: @snkbrush, safe, masterpiece, best quality, score_7, no humans, kaiju, giant monster, rising from sea, glowing eyes, roaring, storm, lightning, waves, ships, night
Negative:
worst quality, low quality, score_1, score_2, score_3, jpeg artifacts, logo, watermark, signature, twitter username, bad anatomy, bad hands, missing finger, extra digits, fewer digits, nsfw, nude, explicit, cleavage
LoRA strength: 1.0
Recommended:
DPM++ 2M SDE, simple, 25 steps, CFG 5, 1536×640 (21:9) or 1344×768 (16:9) for cinematic shots, 832×1248 (2:3) for portraits
Hires fix 1.5×, lanczos, 20 steps, denoise 0.4 — this is where the ink scratches and brush texture really come out.
Tips
Ask for the shot: from below, dynamic pose, wide shot, small figure, looking up, light rays, sparks, smoke, rain. The LoRA pushes them further than the base model does.
Monsters: dragon, giant monster, kaiju, demon, glowing eyes, sharp teeth plus 1girl, small figure, standing for scale, or no humans for a creature portrait.
Break the warm palette: it leans to sunset orange and sepia. night, rain, blue theme, ice, neon lights, moonlight give cold blue and teal frames.
It can trade the face for the angle: characters turn their back to the camera. Add looking at viewer or reroll.
Busy battles can break weapons and chains. Simpler action reads better. Signs and neon text come out as gibberish.
Showcase images use hires fix 0.4, with full prompts in the metadata. Before/after pairs use the same prompt and seed without the LoRA.
Epoch 12 of 14 (about 1400 steps), not the final epoch.
Training
Base: Anima Base v1.0. DiT with AdaLN trained, text encoder and LLM adapter frozen.
Rank/alpha 32/32 · 117 images from 50 digital originals: every picture whole, the widest ones also cut into 3:2 tiles · short tag captions, nothing about style, colours or lighting · 1024 res, buckets 512–1536 · AdamW8bit · LR 1e-4, constant, 5% warmup
Trained on Anima Base v1.0 only. Behavior on Anima Turbo and third-party merges is unknown.
──────────────────────────────────────────────────────
About this version — Krea 2
Trigger: @snkbrush. at the very start, then describe the scene in plain sentences. No tags, no style words needed.
Prompt template:
@snkbrush. <who or what>, <what they are doing>, <where>, <light and mood>.
Example: @snkbrush. A small lone figure stands in a ruined city and looks up at a giant monster with glowing eyes rising through the smoke, dramatic light from behind, wide shot.
LoRA strength: 1.0
Recommended:
Krea 2 Turbo, 8 steps, CFG 1, euler / simple, 1536×640 (21:9) or 1344×768 (16:9) for cinematic shots, 832×1248 (2:3) for portraits
Hires pass: lanczos up about 1.5×, same prompt and seed, 8 steps, denoise 0.25 — sharper lines, more grain and brushwork, while the composition and faces stay. At 0.35 it starts inventing details.
All showcase images use the hires pass (832×1248 to 1216×1824, 1536×640 to 2304×960) and keep their prompts and settings. Before/after pairs use the same prompt and seed without the LoRA and the trigger.
Tips
Monsters and scale: a small figure against something huge is where it shines.
Play silly subjects straight: a giant shiba over a burning castle or a hamster knight against a snake get the full epic treatment.
Epic scenes change the most. A calm portrait of a famous character changes much less than a battle.
Characters usually keep their canon look, but may turn their back to the camera. Reroll for the face.
Action scenes can add small extra figures or a second weapon. Reroll.
Signs and neon text come out as made-up letters.
The same seed and settings keep the shot, but small details change between runs, so copying a showcase image gives a close match, not an exact one.
Two files, pick either: fp32 is the original, bf16 is half the size. In our test they were no further apart than two runs of the same file.
Epoch 3 of 8 (702 steps), not the final epoch — later epochs add scarves and props nobody asked for, and a bare prompt keeps drawing the same character.
Training
Base: Krea 2 Raw. LoRA on the DiT.
Rank/alpha 32/32 · the same 117 images, each with a long and a short natural-language caption · 1024 res, buckets · batch 1 · AdamW8bit · LR 1e-4, constant · flow shift 2.5 · Musubi Tuner
Tested on Krea 2 Turbo (fp8) in ComfyUI only.