Another try with Comfy. Getting better at combining words. Will eventually refactor my current workflow, but for now it's doing beautiful. Thinking about adding more complexity, though. This general-rated otter, though, should be perfect for now.
Tool: ComfyUI Used Custom Nodes: + WAS Node Suite - adds a lot of small, but useful nodes, like Constant Number and other similar stuff. + Plasma Noise for ComfyUI - adds ability to start generation from plasma noise - allows to have better shadows and highlights with no additional cost. To see the workflow (model, prompts, seeds, etc.), download the image and drag it into ComfyUI. It will open the workflow.
Other Useful Nodes: + BlenderNeko's Cutoff - gives more precision about details, allows to limit sets + Davemane42's Custom Node - gives easier control over image segregation, combining workflow of 3-8 nodes into one. + Advanced CLIP Text Encode - gives finer control over CLIP encoder, allowing to simulate results closer to A1111 tool. + Ultimate SD Upscaler - node that makes HD images within a single node instead of 10 different ones. + ImagesGrid - allows to simulate X/Y plot script of A1111. + EllangoK's Post-Processing Nodes - gives some funny filters to increase the quality of finished image. + Vextra Nodes - gives a bit of funny filters, if you are into it. + ControlNet Preprocessors - adds preprocessing, that helps a lot when you try to make sharpening operations