A simple workflow I use when generating images locally. It includes a first generation pass, a rescale with second pass to refine details and a Face Detailer for good measure.

n.2 Versions include a refined negative prompt as well as two new detailers: hands & feets.

n.1 Versions have some clarification about the aspect_ratio, width and height fields on the Configuration node and changed the save node to use comfy save node as it embeds the workflow on the saved image by default.

It should work for models other than ANIMA, but that would require some tweaking of nodes and parameters.

V3 Leverages on ComfyUI subgraphs to keep everything neat.

V2 Doesn't

Both work the same way, generate the same quality and have the same speed.