Short, practical guides to writing prompts for AI image and video models. Each one links to real prompts from the network, published with the result they produced.
A practical structure for Nano Banana prompts: describe the scene in full sentences, lock what must not change, and use JSON when a scene has many parameters.
How to prompt GPT Image in ChatGPT: give ordered instructions, state constraints explicitly, quote the text you want rendered and iterate one change at a time.
The same idea is not prompted the same way in Midjourney, Nano Banana and GPT Image. A side-by-side comparison of prompt style, strengths and when to pick each.
What a JSON prompt is, when it beats plain text, and a reusable structure with subject, setting, composition, lighting and style fields.
How to write AI prompts for product photos: surface, lighting setup, camera angle and negative space, with a template you can adapt to any product.
How to write AI portrait prompts that look intentional: choose a lens, one lighting pattern and a style, and describe the person in a single sentence.
AI video prompts describe time: subject, action, camera movement and pacing, shot by shot. A structure that works for Seedance and other video models.
Plain definitions of the terms used in AI image and video prompts: aspect ratio, negative prompt, reference image, seed, style reference, JSON prompt and more.