As I've been posting more, I've felt motivated to start to tidy up and share some of my go-to ComfyUI workflows. This is my "basic" for most offsite generations: a two step workflow that combines an illustration/anime model that provides good prompt adherence and interesting compositions with a realistic model that transforms the image to an aesthetic that I usually prefer.

I make no pretense that any of this is particularly groundbreaking, original, or advanced, and I think many others have better techniques than me. However, perhaps some users will find it useful.

It uses minimal custom node dependencies (impact pack only, for some nice quality of life nodes).