Deveolpment Update 3

Hey all,

I want do give you a quick update on what's currently happening:

The last update I had planned to the recently released LoRAs was pushed live yesterday.

CIvitAI was having quite a few issues last week, so getting the models released was a bit annoying, but I got there. The updated models are:

- POV Reverse Cowgirl Anal + Creampie

- BDSM - Tied to bed

- BDSM - Pillory

- Futa TakerPOV Blowjob

- Panties to side

I'm really happy with how all of these turned out now. The only one that will probably get another update is the Panties to side one. For being such a simple concept, it still struggles a bit to do what it's supposed to.

Currently I'm working on getting the LoRA that won the last poll done (Bent over table). I already did a bunch of testing yesterday, and it appears that this is a bit of a tricky one to get right. I will do more experimentation with that, and maybe I'll be able to release it later today already. Let's see.

I will put up a new poll for you to vote on in the meantime. I haven't gotten any new suggestions for topics, so I will reuse some from previous polls or add my own in the mix.

There's also an idea floating in my head about trying to do a style LoRA, which I haven't attempted yet. Don't worry, it will have nothing to do with Anime;)

The idea is to try and make a LoRA that gives the image the whole porn aesthetic. I want it to be able to reproduce two different styles of that in fact, professional and amateur porn. The professional one would try to imitate the bright, flashy and sometimes very colorful environments. The Amateur one would produce lower quality images on purpose, having bad lighting, often in dark rooms while using a camera flash and having more, let's say, average environments.

I'm not sure if that is something people would like or need at all, but for me it's something I'd like to use a lot. When I'm prompting for pornographic images I don't find myself entering things like '8k hd' or 'intricate' and such most of the time. I want a more realistic and relatable look to my outputs, not the high-end, super refined look you'd find in most preview images on CivitAI for example.

I realize you can achieve all of that with proper prompting, but I think having a LoRA for it would make it easier and more reliable. But let's see, I need to dig into how to make proper style LoRAs first anyways.

Also, if you have any feedback for me about my models or this page, please let me know.

And that's it for this week's update, thanks for reading!

-KinkAI