There is a time and place for AI, with gingers like this, it's the time and place. Seriously as long as the devs are clear it's AI then I can accept it. It's mtl, denial of AI use and shit like that I really can't deal with.
- Thread starter 1nsomniac22
- Start date
TheRetroAiden
Well-Known Member
- Apr 4, 2022
- 1,299
- 462
- Jul 16, 2025
- 13
- 17
FatYoda
Well-Known Member
- Sep 12, 2017
- 1,763
- 1,747
yep, i just finished the story. Gotta say taht was 10/10 until the end. I'm guessing you couldn't get the generator to give you penetration scenes or something. Id rather have had still images of him balls deep than an animation that looks like he's just sliding along her clit. Complete loss of immersion when she's talking about how deep you are and all you see is a cock outside a vagina.
- Jul 16, 2025
- 13
- 17
That's not supposed to happen!! Sorry!!! I was using layered sprites that should have given the appearance of penetration, not floating over.
I was trying out some experimental approaches to getting animations into the story without actually needing to render animations.
Oneiros Games
New Member
- Jul 3, 2025
- 5
- 5
- Jul 16, 2025
- 13
- 17
Digging your avatar, thanks!
I have a soft spot in my cold, black heart for redheads...
CutieLvr
Forum Fanatic
- Jun 3, 2020
- 4,867
- 5,792
visually pleasing how ?
In the profile picture of the person above, she looks like a cocaine addict.
Look, I'm just glad the female character is a hot redhead and it's NOT another NTR game. I would rather have AI CG than the same manish-looking Daz3D model with a fucked-ass haircut.
Edit: ah, just saw that you made your own lora for her
- Jul 16, 2025
- 13
- 17
No Jinx LoRA was used; actually built my own novel Gwendolyn Lora.
Here is a quick 'n dirty explainer how I built the LoRA using SD generated images (so, I guess the absolute definition of AI slop, since it was AI images used as input to train a LoRA to produce consistent images).
You must be registered to see the links
Since then, I've worked on what I'm calling my Gen2 LoRAs. I take a first round of a very few images to train up the a LoRA. Then I generate a larger amount of input images using that LoRA with more expressions, poses, clothes, etc to generate the second generation LoRA.
So far, it's been working well for the next project I'm working on. Having a very narrow LoRA to keep 1 character consistent seems a workable solution to consistency.