Prompt library · 195 prompts
195 copy-ready prompts across 7 collections — editorial and fashion, poster and graphic, landscape, botanical still life, cinematic stills, fantasy and sci-fi, animal portraits. Each one shows the image it actually produced, so you can judge it before spending a credit.

Editorial AI prompts for fashion, portraiture, and magazine-style art direction — sketch, illustration, and photographic styles with the generated result beside each prompt.

Graphic design prompts for AI: geometric poster composition, Memphis and Bauhaus-adjacent layouts, colour blocking, and print texture with room left for type.

Landscape prompts for AI image generation covering light, weather, geology, and scale — mountain, coastal, desert, and glacial scenes with the result shown for each prompt.

Studio flower photography prompts for AI: single-stem botanical portraits with controlled backgrounds, petal translucency, and gallery-grade colour separation.

Film-still prompts for AI image generation: lens language, colour grading, and lighting direction that produce frames which read as cinema rather than stock photography.

Fantasy and science fiction prompts for AI concept art: worldbuilding scale, invented architecture, and atmosphere, with the generated result beside every prompt.

Studio-lit animal and pet portrait prompts for AI image generation: coat and feather texture, eye catchlights, and backdrops chosen to separate the subject.
How to use them
The prompts in this library are long on purpose. Each one names the medium, the light and its direction, the materials in frame, the palette, and the things to avoid — because those are the clauses a model can act on. Adjectives like “beautiful” and “high quality” are not instructions, and they are usually what separates a prompt that works from one that does not.
The fastest way to make one of these your own is to change a single clause and regenerate. Swap the subject but keep the lighting. Or keep the subject and change only the grade. Changing everything at once tells you nothing about which part was carrying the image.
FAQ
Yes. Every prompt on this site is free to copy, edit, and reuse, including for commercial work. You only spend credits when you generate an image.
They are written in plain descriptive language rather than model-specific syntax, so they transfer between image models. They were generated and tested in Yiluna, and the result shown beside each prompt is the actual output.
Change one clause at a time. Swapping the subject while keeping the lighting, material, and composition language intact is the reliable way to get a new image in the same style.
A prompt without its output is a guess. Showing the generated image next to the text lets you judge whether the prompt is worth your credits before you run it.