
Original document
GPT Image 2 · en
Scene:
Create a square dependent event image in the exact same Halloween night world, preserving the same ember lighting, charcoal depth, restrained fog, and premium suspense from the reference image, but open the scene toward a rooftop party environment. The frame should still feel like the same campaign family, with continuity obvious in light, materials, and threat level, while shifting into a more social nightlife surface.
Subject:
Feature a rooftop setup with candlelit tables, elegant drinks, pumpkins or seasonal props used with restraint, and city-night depth in the background. Preserve exact campaign identity from the reference model while changing the venue emphasis toward elevated nightlife. The scene should feel polished, seductive, and event-promotional, never comedic or bargain-party themed.
Important details:
Keep the same premium Halloween grammar: controlled menace, sharp pools of light, dark surfaces, and one coherent invitation to the night. The rooftop setting should add exclusivity, not spectacle overload. Glass, metal, smoke, and weather cues should all feel real. The result should work as a hero asset for rooftop events, hospitality activations, or nightlife campaign rollout.
Use case:
Halloween Promo series study 2/9.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, preserve exact campaign identity from the reference image, no readable menu signs, no costume crowd, no gore, no giant spider-web cliché, no logo glassware, no fireworks, no cheesy neon, no copyrighted creatures.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.





