Excited to try this out! I've been wanting to make my own game as well. What AI do you use to generate these pics and what prompts do you use? Any AI I try doesn't get the size difference correctly.
Viewing post in Into the Portal: Giantess world journey comments
I'm using Krita which is a digital art software and I'm using it with the AI plugin from Acly. This plugin work with Comfy UI as a backend and you can generate an image directly into Krita that way, you can easily edit the generated image and obtain great result that way. You can check the github repo for the plugin here:
https://github.com/Acly/krita-ai-diffusion
It takes a lot of setup to get it work but it’s worth it. I only use two models (IllustriousXL and Qwen Edit 2511) and a few custom loras but a gts lora is not needed to achieve that kind of result with Krita using the layer. For prompting, I don't use anything special. As example, the close-up of the giant girl looking at the MC on her hand palm, I generate only the close-up of the girl first looking at her hand using illustrious, I select the hand area, then add the MC kneeling on her palm using Qwen Edit model and voila you get that kind of result.
Hope it help.