Stable Diffusion
Open-source image model you can run, fine-tune and customise yourself.
Open models (free to self-host); paid API and apps available.

Stable Diffusion is the open-source cornerstone of the image generation category, released by Stability AI. Unlike closed tools, its model weights are open, so anyone can download, run and modify them, which sparked an enormous community of extensions and custom models.
What it does
At its heart, Stable Diffusion turns text prompts into images, but its real strength is customisation. Because the weights are open, developers built fine-tunes, LoRAs, inpainting and ControlNet for precise control over style, pose and composition. You can run it locally for full privacy and control, or use hosted apps and Stability AI’s API when you’d rather skip the setup.
This flexibility makes it a favourite for people who want to tune every detail rather than accept a fixed aesthetic.
Who it’s for
Developers, tinkerers and studios who value control, privacy and cost savings, plus anyone building image features into their own products. The trade-off is complexity: self-hosting needs a capable GPU and some technical comfort. For casual users, hosted front-ends and the paid API deliver the same models with far less friction.
Key features
- Open-weight text-to-image models
- Free to self-host on your own hardware
- Huge ecosystem of fine-tunes and LoRAs
- Advanced control via ControlNet and inpainting
- Paid API and hosted apps from Stability AI
- Active open-source community and interfaces
Pros & cons
👍 Pros
- Free and fully customisable
- Massive community and tooling
- Full control when self-hosted
👎 Cons
- Steeper technical setup
- Local use needs a capable GPU
Stable Diffusion — FAQ
Is Stable Diffusion free?
The models are open and free to download and run on your own hardware. Stability AI also offers paid API access and hosted apps for those who prefer not to self-host.
What hardware do I need to run it?
Running locally typically needs a capable GPU with several gigabytes of VRAM. Lighter setups and cloud services exist for people without powerful graphics cards.
Can I fine-tune Stable Diffusion?
Yes. Its open nature enabled a large ecosystem of fine-tunes, LoRAs and interfaces like ControlNet, letting you customise styles and control composition.
Where can I use Stable Diffusion without installing anything?
Many web apps, plugins and hosted services run Stable Diffusion, and Stability AI provides an API, so you can generate images without local setup.

