GPT Image Prompt Guide
Prompt Structure for Better GPT Image Results
Use a repeatable prompt structure to make GPT Image 2 and ChatGPT image prompts easier to adapt, compare, and reuse.
Start with the target image
Describe the subject, scene, format, and final use before adding style details. A clear target helps the image model understand whether it is creating a product mockup, editorial photo, UI screen, poster, infographic, or edit.
Add constraints after the creative direction
Put camera angle, lighting, composition, aspect ratio, text requirements, and negative constraints after the main idea. This keeps the prompt readable while still giving the model specific boundaries.
Change one variable at a time
When adapting a copied prompt, change the subject, brand style, scene, or output format separately. Comparing small changes makes it easier to learn which prompt details actually improve the result.
Browse structured prompt examples
Open the gallery to compare real image results with the prompt text behind them.
View GPT Image Prompts