Basically, the mainstream tools for image and video generation are
Stable Diffusion WebUI and
ComfyUI.
If you want to use a website instead, that's fine, but customization options will be limited, and I wouldn't really recommend it.
For AI models, check out
civitai.com. It's probably the largest collection out there. You can also find LoRAs and VAEs there.
If you're aiming to generate photorealistic images or videos, you'll likely succeed. However, if you want to create natural-looking illustrations, it's a bit more complicated. Issues like having six fingers can be fixed using the Badhands LoRA, but unnatural hair details are harder to correct with AI alone. That's why most illustrations posted on Twitter or similar platforms have been manually touched up. If you want high-quality AI illustrations, you'll need to do some manual editing too.
For beginners, I recommend starting with
Stable Diffusion WebUI.
If you don't have a high-spec PC to run SDWebUI, unfortunately, you'll be limited to just playing around, but you can use Huggingface Spaces instead. You can access it
here. At the very least,
animagine xl can generate NSFW-style images.
Good luck!