This model is similar to my WCXL model for use on Pony. Its not exactly the same, but its similar enough that I feel comfortable uploading it. Like the pony version this model is meant to be "watercolor" but its not quite watercolor at the same time.

This requires upscaling to get better looking results, as the base outputs will not look very good until they're upscaled.

I'd recommend having it around 0.7 - 1.0 for strength if you have a character as your primary subject.

For landscape images I'd recommend turning it down lower, somewhere around 0.4-0.7, otherwise you might get some "scratchy" results (some of the landscape examples have this.)

Prompt I use:

masterpiece, best quality, amazing quality, very aesthetic, high resolution, ultra-detailed, absurdres, newest, watercolor_\(artwork\),<lora:WCXLN2dot0:1>

You don't need to use watercolor_\(artwork\) but I think it helps slightly. I think. Could just be placebo, feel free to test with or without that tag.

Negative prompt I use:

old, oldest, graphic, text, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra digits, fewer digits, very displeasing, (worst quality, bad quality:1.2), bad anatomy, jpeg artifacts, signature, watermark, username, simple background, conjoined, bad ai-generated,