Automatic1111’s Stable Diffusion Web UI is a highly popular web user interface for accessing Stable Diffusion’s AI models without the need for coding. As the leading browser-based interface using the Gradio library, it comes with a variety of features to improve the user experience and produce excellent results. The UI stands out for its extensive range of extensions and constant innovation, which sets it apart from other user interfaces like InvokeAI. I still advocate for InvokeAI, not only because it’s easy to install on a Mac M1 or M2, but because the user interface and outpainting features are amazing.
Looking for more information on Stable Diffusion? Check out these other articles I’ve written about Stable Diffusion:
This article explores the most valuable feature of the Automatic1111 Stable Diffusion Web UI: it’s extensions.
Extensions for Automatic1111’s Stable Diffusion Web UI
In my view, the extensions are what set this web UI apart and make it great. While there may be other UIs available that are superior, such as InvokeAI, they do not keep up with the same level of innovation as the Automatic1111 UI.
Extensions are an additional feature of the Automatic1111 Stable Diffusion Web UI that provides a more convenient way to use user scripts.
They can be easily installed using the web UI. And all extensions are located in their own folder within the extensions folder of the web UI.
For power users they can also be installed manually. To install an extension manually, one can use git to clone it into the extensions folder. For example, to install an extension from https://github.com/AUTOMATIC1111/stable-diffusion-webui-aesthetic-gradients, one can use the following command: git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui-aesthetic-gradients extensions/aesthetic-gradients.
Alternatively, one can copy and paste a directory into the extensions folder. And for those interested in developing extensions, there is a guide available for developing extensions.
Top 5 Automatic1111 Stable Diffusion Web UI Extensions:
1. ControlNet
https://github.com/Mikubill/sd-webui-controlnet
ControlNet is an absolute game changer for my personal workflow, and it is the coolest thing in Stable Diffusion to date in my opinion.
The ControlNet extension allows you to use different models like Canny, Depth, HED, OpenPose, etc. in order to make your output match any input you can imagine. There is a great explanation for how this extension works on Reddit.
2. Dreambooth
https://github.com/d8ahazard/sd_dreambooth_extension
Dreambooth is probably one of the most well known extensions. Dreambooth allows you to create custom models and checkpoint files using your own datasets to be used with Stable Diffusion. Want to train your face, an object in your house, or a style into a model so you can make images with it? This is the extension for you.
You can also train LORAs using this extension as well. If you’re curious, I’ve written before about Customizing Stable Diffusion models with your own data.
3. Deforum (Animations)
https://github.com/deforum-art/deforum-for-automatic1111-webui
Deforum is useful for creating animations with Stable Diffusion. That’s right: video, GIFs, and stuff like that. Kind of amazing that this kind of thing is even possible. They are not to the level of a blockbuster movie or animation studio. But with some creativity you can really create some awesome stuff!
4. Dynamic Prompts
https://github.com/adieyal/sd-dynamic-prompts
Dynamic Prompts is an extension that implements an expressive template language for random or combinatorial prompt generation along with features to support deep wildcard directory structures.
I use this extension all the time in order to help create prompts and ideate. If you feel like you have prompt writers block, this may be the extension for you! I’ve also found this very useful for storing common terms and organizing information in one spot to help build prompts.
There are tons of cool features as well, like attention grabber, that will automatically add emphasis to random parts of your prompts.
5. CLIP Interrogator
https://github.com/pharmapsychotic/clip-interrogator
I use the CLIP Interrogator to help with building prompts. It runs pretty fast and generally provides good output.
The CLIP Interrogator is a prompt engineering tool that combines OpenAI’s CLIP and Salesforce’s BLIP to optimize text prompts to match a given image. Use the resulting prompts with text-to-image models like Stable Diffusion on DreamStudio to create cool art!
Want to figure out what a good prompt might be to create new images like an existing one? This extension might be useful for you!
Other extensions
These extensions are my top 5 favorites. There are so many others! See the full list here: https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Extensions
Automatic1111’s Stable Diffusion Web UI is a powerful tool for anyone looking to use AI models without coding. Its range of extensions, constant innovation, and user-friendly interface make it a standout choice in the field. I still strongly advocate for InvokeAI, but the extensions above make it well worth having a copy of this UI to use along side it on your machine.