How to Install Stable Diffusion

To install Stable Diffusion locally will require a bit of Python and GitHub knowledge, but it’s not too difficult and really you can have the whole thing set up and running in less than an hour!

The easiest way to install Stable Diffusion in my opinion is using the AUTOMATIC1111 WebUI tool on GitHub. To do this, first you will need to download the following:

After you have those two installed on your Windows PC or Linux, open up CMD or your Python console in the directory you wish to install Stable Diffusion, and run the following code:

git clone

Wait for it to clone the folder and then navigate to the “models” folder. The path should look something like this:

Here is where you will drop all your .ckpt files (these are models/styles). To start with, you should add the main Stable Diffusion models which are 1.4, 1.5 and 2.0. You can download all of these models from HuggingFace. Navigate to the “Files and versions” tab and you should be able to find the .ckpt files, they’re usually the largest file size. Click on the small downwards arrow next to the file size to download.

Stable Diffusion 1.4 .ckpt –

Stable Diffusion 1.5 .ckpt –

Stable Diffusion 2.0 .ckpt –

After you’ve moved those files into your models folder that’s it! You’ve successfully setup Stable Diffusion and can now locally run it from your machine by opening “webui-user.bat” located in the folder you cloned earlier. Let the CMD console install any necessary prerequisites, once it shows the following:

Running on local URL:
To create a public link, set `share=True` in `launch()`.

You can simply visit “” to open Stable Diffusion in your preferred browser (this is just your localhost address, not your real IP). If it shows a different port number other than 7860 at the end, make sure you replace it to match. If you wish to use any custom models you just need to download their .ckpt files and drop them in the models folder. Some of them may come with required .pt files too, in that case just drop those in the same models folder. You can find lots of models on HuggingFace and Civitai.