ChatGPT Image Prompt Template

Subject reference (optional but powerful)

Use the uploaded reference image as the identity source for the subject.

Identity lock

Preserve facial features, proportions, age, skin texture, hairstyle, expression exactly.  No beautification. No smoothing. No face reshaping.



Style exclusions (defensive prompting)

No cartoon, anime, illustration, CGI look, waxy skin, plastic texture.



Style directive (positive rules)

Photoreal, high-fidelity photography.

Real materials, natural skin texture, realistic fabric weave, physically plausible lighting.



Camera + framing

Camera: [shot type], [angle], [lens], [distance].

Framing: [full-body/waist-up/close-up], [placement], [headroom], [no cropping].



Pose + action (physics)

Pose: [exact body position].



Action: [what happens], [where in frame], [what moves], [what is frozen].



Physics: realistic motion blur, gravity-consistent debris/liquid behavior.

Wardrobe + grooming



Wardrobe: [specific], [fit], [colors]. No fantasy unless requested.

Environment



Location: [specific]. Minimal clutter. No extra objects unless listed.



Lighting

Key direction, fill behavior, rim highlights, shadow softness.

No glowing edges. No overbloom.



Composition + output

Aspect ratio: [e.g., 1080×1350].



Negative space: [where and why].

Silhouette readable at thumbnail size.



Hard exclusions

No extra fingers, warped hands, duplicate limbs, distorted text, random logos, watermarks.