I've always been a huge fan of that clean cel-shaded look with soft bloom, gentle diffuse lighting, and the subtle cinematic glow you often see in high-quality anime illustrations. I spent a long time searching for a LoRA that could consistently reproduce this specific atmosphere, but I never found one that really matched what I had in mind.

So I decided to train one myself.

The final result honestly exceeded my expectations. Instead of simply changing the colors, this LoRA helps create a much softer, more cinematic cel-shaded aesthetic with smooth lighting transitions while keeping the image clean and detailed.

I've included several reference images below to showcase the exact visual style I was aiming for. After generating your own images, feel free to compare them with those examples and see how closely the LoRA recreates that distinctive cel-shaded bloom look.

I encourage you to experiment with different prompts and models—I think you'll be surprised by how versatile it is.

Recommended LoRA Strength: 1.0

Here are the reference images I had in mind when creating this LoRA. I used them, along with many other samples, to train the model.