Original document
GPT Image 2 · en
Create a candid 35mm film-style lifestyle photo of a young woman at a retro American diner during golden hour. She is sitting casually on the hood of a cream-colored vintage station wagon with wood paneling, parked directly outside the diner. The woman has {argument name="hair color" default="long dark brown wavy hair"}, olive-toned skin, thick eyebrows, and a relaxed, slightly smiling expression while looking off to the side rather than at the camera. She wears a {argument name="jacket style" default="navy varsity jacket with white striped cuffs"}, a light ribbed tank top, high-waisted loose blue jeans, white socks with stripes, and black-and-white sneakers. She holds a tall pale milkshake with whipped cream and a red-and-white striped straw near her mouth. On the car hood beside her is exactly 1 red plastic diner basket lined with paper, containing fries or snacks, plus exactly 1 small paper cup. The background shows a busy old-fashioned diner strip with neon signs and warm signage, including visible diner-themed words like “EAT,” “FRIES,” “DINER,” and a Coca-Cola sign; include large windows with seated customers inside, chrome trim, parked cars, string lights overhead, and a sunset sky. Use warm backlighting, soft haze, natural film grain, muted nostalgic colors, slight lens softness, realistic skin texture, and an unposed documentary feel. Frame vertically in a 2:3 portrait composition, with the woman centered on the car hood and the diner storefront filling the upper left background. Avoid a polished fashion-shoot look, avoid modern cars dominating the scene, and do not add extra text beyond plausible diner signage.How to use
- Copy the complete original prompt, including any parameters.
- Replace the subject, text and composition with your own requirements.
- Use a compatible external tool; check its current access conditions and review the result.
Before you start
- Check whether a reference image is required. A compatible model label does not mean this example was generated with that model.
- Keep names, figures and visible text accurate; results can vary between model versions.
