Hello AnniBlake, thanks for your comment.
Yes, I do use ComfyUI locally but mostly as a backend server only. I mainly use Krita digital painting software with an AI plugin that connect to that ComfyUI instance so you can generate your image directly into Krita. The main reason is when you need to do some manual editing on your generated image, Krita really shine. You can still do manual editing directly in ComfyUI but I found Krita more flexible IMO. You can download Krita for free here: https://krita.org/en/download/, and download the AI plugin from Acly here: https://github.com/Acly/krita-ai-diffusion.
I use WAI-illustrious-SDXL as base model and I trained my own character lora to keep the character's consistency as close as possible. For my lora training, I'm using AI-Toolkit from Ostris: https://github.com/ostris/ai-toolkit
Hope this answers your questions. If you have other questions, feel free to ask.
Thank you so much for the detailed answer. I'll try playing around with it next week to see if I get good results^^
I haven't had a good experience with AI when it comes to GTS content so far, because sometimes it doesn't understand the scale. But of course, I don't want to rule out the possibility that I might promt bad.
There are some GTS lora available that could help but there is no way to control the character scale unfortunately. Most of the time, I just generate the character with a neutral background, copy that character onto a background I wanted, then rescale the character at the size you need, refine the whole image and that's pretty much it. It may required a lot of manual editing but you can achieve really great result that way.
Here an example of a miniGTS I made a few months ago.
