Hi guys! I am creating a new visual novel game w/ ai generated image, so please advise!

[shabiphobia]

Hi guys! I am creating a new visual novel game. Our game image is generated by AI(stable diffusion) and this is how the charactor would be like.


Please provide me with advice on it. Thanks!

 

[letsgethappy]

Wellcome and i am not quite sure that i am a good advisor, but what comes in mind:

A little more information couldn't hurt. AI is just the tool to create the images, but what is the game about, what content is planned, what software is used and so on. Something like that.

However, good luck ans success with your project!

 

[desmosome]

Can you keep her consistent between images? That's the first requirement.

 

[Lerd0]

...i'd nut on it.....
...good luck...

 

[Fate stay]

Hi.

 

[S Pied_Piper]

hello!

 

[picobyte]

Here's my tips. Not sure if it will work though.
Keep the seed for consistency, in comfyUI I need to keep it consistent across 2 generations. Take care with eyes and hands.
use simple words that cost one token mostly, try to keep within 75 tokens. use (masterpiece hires, best quality), then prompt.
SD15 is different than sdxl in tokens. sdxl is smarter, but still comma sep fairly simple concepts.
First generate several images. Lora's might help. IP adapter face might work thereafter if you have several images, or you need to train a model.
Focus on the person, keep the background simple. because if you can remove background, than you can use the images on multiple backgrounds.