I'm not sure what the rules on external links are here, so I'll just dump bunch of terms out here
There's a lot of different options out there, some paid, some free. Most of the big paid ones out there are censored (no NSFW), which actually isn't a terrible thing for sharing in public threads on this forum, such as this one. See
here for LC AI rules
Image Generation
Popular paid models
- Midjourney (censored, no NSFW)
- Chat GPT 4o image model (censored, no NSFW)
- Leonardo AI (still censored, but less so, it actually uses some open source models)
Local models
If you have a decent GPU (at least 6GB VRAM, more is better), you can run local (open source) models.
There's a few different base models, but the 2 biggest are:
- Stable Diffusion (which has a few different versions)
- Flux
People have taken these and built finetunes of them, so there are hundreds of models out there (check out CivitAI, it's a site that hosts AI models). Some are more suited to NSFW than others.
The 2 main ways to run models locally are:
- Stable Diffusion WebUI Forge (more straightforward to use)
- ComfyUI (more flexible, node-based)
Local models give you the most flexibility/control/freedom, especially with more advanced things like Loras, ControlNet, IPAdapter, etc.
Video Generation
Popular paid options
- Kling
- Hailuo
- Runway
- Sora (OpenAI)
- Veo3 (Google)
These are all pretty heavily censored
Local options
These are both uncensored. They require even more VRAM (I would say at least 12-16GB, ideally 24GB) to run though (which can be done in ComfyUI)
If you don't have a decent enough GPU but want to use local models (for image or video), there's basically 2 options:
- There's sites that let you use them, usually for a fee/subscription/credit system, but some are free/have trials. One I found mentioned here recently is Perchance, which is free and doesn't even require an account (but seems to be a little limited)
- There's also the option to rent GPU servers from places like Runpod that let you run a ComfyUI instance and you're charged by the time you're running the server, but that's a little more complicated to set up
This is only touching the surface and just some of what I've played around with. There's a bunch more options to look into, and everything changes so rapidly with new models releasing all the time, it's hard to keep up. Youtube is a good place to learn more about all of this