Hook

Their other posts in the index, biggest breakout first.
There are 3 levels of AI and image prompting and most people never get past level 1. But with this simple formula I'm gonna give you, you will reach level 3 every time. So after thousands of images, this is the most reliable prompt structure I've found. Subject, setting, style, lighting, and technical details. You don't always need all 5, but the more specific you are, the less the model decides for you. And the more control you get over the final image. So for level 1, you'll only use the subject and the setting. A classic hourglass on a wooden table. That's enough to get the object, but the model decides almost everything else. For level 2, we add context and specify the lighting. A clear glass hourglass with a dark wooden frame on a rustic wooden table. The sand in the top half is almost gone, with most of it collected in the bottom. Soft morning light comes from a window on the left. A small stack of old books sits beside it. Now we have details about how the hourglass looks, the type of table, the books beside it, and the lighting. You control most of the image. And for level 3, we keep the same prompting, atmosphere, and style, and add technical details. Tiny dust particles float in the light beam. Cinematic style, photography, close-up composition, shot on Fujifilm X-T4 with a 58mm f/1.2 lens, shallow depth of field, Kodak Portra tones, warm earthy color palette, soft film grain. You control the image, not just what to generate, but also how the final image should look. So generating, remember these simple rules and get the results you want.