Making an ai cg game

[LfkCn]

I've been trying to improve myself in game development for a while now, but I'm moving slower than a baby crawling. My goal is to create a game using StudioNEOV2, but I haven't mastered even the most basic features yet. I'm away from home for work for 14 hours a day, including commuting, so I don't have much time to dedicate to learning game development. I hope that my schedule will change, and once it does, I will have time to learn programs like StudioNEOV2, Daz, and Blender, but for now, I don't. I've gone on for too long. I really want to make a game. My goal isn't to earn money in the short term; in fact, from what I've observed, AI CG games aren't very popular. Still, I want to have my own work. I thought of using artificial intelligence. There are many beautiful-looking games in the AI CG genre. I want to try developing a game using artificial intelligence. Of course, every path has its own challenges, and I'm not saying making AI CG games is very simple, but as I said, I want to try. My questions are as follows: (My questions may seem simple to you, but the only thing I know about artificial intelligence is that I want to translate what I want to write into ChatGPT on f95)

1. What programs are used to make these types of games?
2. Which ones are free? As someone living in a Middle Eastern country, even $20 is a lot for me. I don't think I can afford a paid program.
3. What is the copyright situation? For example, there might be a nice free AI program, but it might be prohibited to use it in an adult content game or in any work where you can earn money. Maybe you can't sell it, but you can receive donations from places like Patreon (like Illusion games).
4. Lastly, what are your thoughts on AI CG games? Are you among those who hate them? I don't think I've played a game like that until today, but the promotional visuals look quite nice.

 

[MissFortune]

14 hours a day? Jesus. I worked that much during the pandemic fairly often and nearly hit $100k that year on supervised internship pay. But it's not something I'd do every day not willingly. Did a number on my health, tbh.

1. What programs are used to make these types of games?
2. Which ones are free? As someone living in a Middle Eastern country, even $20 is a lot for me. I don't think I can afford a paid program.

Kinda depends. You sound like you're not too strong in the hardware department (apologies if assumption is wrong.), so something like SDXL + ComfyUI isn't going to help much if you don't have the hardware to power it. That would also likely apply to many other types of locally run AI/models.

So, this basically forces you into using online services. The issue with online services is that many of the good ones don't allow you to make a large-busted woman with any sort of ease. And the ones that do allow NSFW only allow it until they're profitable enough to ditch the NSFW side of their business.

3. What is the copyright situation? For example, there might be a nice free AI program, but it might be prohibited to use it in an adult content game or in any work where you can earn money. Maybe you can't sell it, but you can receive donations from places like Patreon (like Illusion games).

Most sites allow it, with the exception of Patreon if the models look too realistic. You'll make a decent sum on Steam, and maybe some pocket change on Patreon/Itch/SS/etc.

4. Lastly, what are your thoughts on AI CG games? Are you among those who hate them? I don't think I've played a game like that until today, but the promotional visuals look quite nice.

You've gotta go into this knowing that AI already has a stigma around it - if not multiple. Many see AI art as stolen works being sold commercially, so their line of thought becomes "if they stole it, why should I pay for it?", that's on top of all the shitty Steam cashgrabs using AI to make a quick few grand. So, you'll likely see a lot of pirating. You'll also see a larger amount of people dogging your work for being AI, if only for being AI.

As for my opinion? It isn't worth much, but I think AI art was just the passing fad. They shoveled money into AI art and now the money/VC funds have moved onto more general uses (assistants, etc.). Will you see it? Will you still see work on the systems? Absolutely. But that moment of rapid acceleration is done and progress will be slowing down. The bigger money is elsewhere and the people with knowledge/skills are following it.

That all being said, as long as you can avoid the pitfalls of AI (fifteen fingers, weird eyes, teeth, etc.), then go for it. Just get some thick skin for the initial wave of criticism. It's also worth mentioning that getting a good result for a single image can take hundreds of attempts, even using img2img to keep the same character(s). Since most AI sites run on some coin/etc. system to pay for a generation, you could realistically (and likely will) blow through much of what you pay for before you even get a result you like.

 

[LfkCn]

-

I'm not working 14 hours, but during this time, I'm away from home. I work for 11 hours, take a 1-hour break, and commuting to and from work takes 2 hours. This unfortunately consumes my entire day. However, if everything goes as planned, I should be free in about a year. Other than that, thank you for informing me in detail. After writing this, I did some research and came across a program called Stable Diffusion. It looks quite good, but it requires a good graphics card, which I don’t have. I'm looking into some different alternatives.

 

[FromOtherSpace]

Stable Diffusion (SD) is a model, not the program which run the actual image generation. It's a dataset, think to models as a lot of info about a lot of pictures. You can change models to "reference" different pictures.

A powerful graphic card make generation faster. The hard limit is RAM. If you can load a model, your program will generate images, it may take a lot of time but something will output. My pc has 16G and I can't load XL nor Pony. SD 1.5 (and all 1.5 based model) work fine.

Also, the bigger you want your images, the more RAM and time are needed (You can resize after generation but a small image has fewer details, so it don't scale very well)

Easydiffusion (an open source program to run models) have an option to run on your cpu. It's way slower than any GPU but it's an option, you may want to consider (if you got enough RAM).

 

[Baronluck]

AI might get frustrating to work with if you're using it to generate images for a game. I can imagine it changing the characters' faces and clothes, and you have to try different prompts to get the right look. It'd be faster to learn how to make characters and scenes in Daz. I would consider using AI for story ideas though. AI could generate story outlines, character descriptions, locations, and scene ideas. But anything where it has to reliably generate content would become difficult to work with.

 

[Winterfire]

AI either sucks, or takes a very lengthy time not only to learn but also to make semi-decent results (which are never really good, just look somewhat colorful and good if you look at it for a few seconds). It's not magic, and those AI games that actually look good for more than a few seconds have a lot of work put behind it (I am talking about hours per CG) and deserve praise they often don't get.

Studio is ridicolously easy to use:

You must be registered to see the links

and even though this is KK, the same knowledge can be applied on HS2. If you failed to learn this, you have no chance to learn AI, unless you want to dump random shitty "art" in your potentially good game, get hated by people who have no better way to spend their own time than hating on things, and then eventually just quit because I doubt you want to get shat on after you spent your free time working on something after 14h away from home.

If you want my honest opinion, you should give a real shot to StudioNeo because I refuse to believe it's too hard for you, if the guide I linked you wasn't enough, you can literally find tons of tutorials on youtube or even showcases of people building something on Studio which you can follow along:

You must be registered to see the links

Sure, your first results won't look as good, but that's true for anything, you have to start somewhere, and this is the easiest start. You can only go up from there.

If for some weird reason you really can't wrap your head around StudioNeo, then just drop the art part entirely and do a text based game. They have a market for sure, not that it matters since you don't care about money.
However, even if you did, just look at Lilith's Throne, CoC, and other games of the genre to see that they have a large playerbase and support.

Once you have more free time (or find an artist to work with) you can easily add art to your project.
This project: https://f95zone.to/threads/the-last-sovereign-v0-75-4-sierra-lee.616/ for instance started out with no art at all, and it was pretty popular and well liked. Many years later they started adding art too.