Download Gauss (Mac) – Download & Review Free

AI image generation is both revolutionary and controversial, but while the tools you need are readily available, they’re not always as accessible as you’d like them to be.

To make things a bit easier for beginners, you can use an app like Gaussian. It is a Swift-based open-source application that allows you to run stable diffusion models locally and generate images with minimal technical knowledge.

Launch and download the model for offline use

There are many online stable diffusion image generators, but they have many limitations. Running the model locally lets you do whatever you like, no internet connection required.

First, you need to download at least one available model, which can be done through the application’s UI. They are both large, so it will take some time to install. Once done, you’ll be able to generate any number of images without relying on the API.

simple but limited

Gauss saves each session as a separate document so you can easily continue working later. Once the project is loaded, go ahead and give various hints to generate the images you need. Figuring out what to input is the hardest part of AI image generation, but the trial-and-error approach works.

Unfortunately, while the app is very intuitive, it also lacks almost all the advanced features supported by the stable proliferation. It does not support negative hints, changing image dimensions, reusing the same seed, and the img2img function.

You can adjust speed and creativity, which should translate to number of sampling steps and CFG scale, and generate one, four or nine images at a time. The same prompt can be run again later, but adjustments are limited if the same seed cannot be used. To save an image, simply drag it to Finder or another application.

Good for beginners, but lacks features

Essentially, Gauss is a great option if you’re just getting started with AI image generation, but it’s too restrictive for anything but occasional use. Nonetheless, the fact that it is written in native code and runs the model natively in an intuitive UI is appreciated.