Stable Diffusion is a deep learning, text-to-image tool released in 2022. It is primarily used to generate detailed images conditioned on text descriptions.
As a free and open source tool, it can be run directly on your own computer or using a cloud service. Te Wāhi has its own Stable Diffusion web tool you are free to use (opens in a new tab) and can be run on any device with an Internet connection.
Navigating and downloading from the site is simple, and does not require an account.
Take care when using Stable Diffusion tools directly on your computer (rather than using a cloud tool). Stable Diffusion requires a very powerful computer in order to run locally.
This is the homepage of Generate (opens in a new tab), the Stable Diffusion web instance maintained by Te Wāhi. Click "Start generating" when you are ready to begin.
For whatever you want to generate, it is recommended that you add the text “mdjrny-v4 style highly detailed…”, and include as many details and descriptors as you can.
e.g. “mdjrny-v4 style highly detailed, ancient guardian creature of the forest and flora”
You can cycle through new generations by clicking "Generate" again if you aren't happy with a result. Please, don't spam this button too quickly, otherwise the site will timeout for you and you will be temporarily banned.
From here you can continue working on this image in a tool such as Photoshop, or simply use it as inspiration for another work.