The AI Stable Diffusion generates images for you using prompts. There are also several image styles available to you.
Prompts for Stable Diffusion
Stable Diffusion works simply: you enter a text and the AI creates an image from it. To make your result as good as possible, there are different prompts and some rules for the optimal commands.
- The prompt should be as short as possible. Stable Diffusion does not understand long, convoluted sentences.
- The AI works best with prompts in English, as it was developed in this language. But prompts in German also work.
- Be as specific as possible with your information. Add adjectives like “red”, “small” or “bright”.
- Also work with negative prompts. This means that if you find that there is an item in all four picture results that you do not want, write it to the negative prompts.
- In addition, Stable Diffusion allows you to add a style statement after your prompt. There you can select, for example, “3D” or “Futuristic”.
Sample prompts for Stable Diffusion
If you don’t know where to start, there is a database on Stable Diffusion where you can view prompts already made and the results. Here we also have some examples in English and German:
- Japanese Zen garden against red sky, setting sun
- rainy city, girl in yellow coat with blue umbrella
- painting of an old man, colourful pop art
- arei dog puppies in water, playing with ball.
- a snowy landscape, illuminated by the northern lights
- a quiet evening in a 1920s jazz club
- a curious owl peeking out from a hollow tree trunk