
Documento original
GPT Image 2 · en
Scene:
A square large-format city spectacle where architecture bends, folds, loops, or stacks while remaining physically convincing enough to feel like a photographed event. Use broad daylight or clear night atmosphere, dense urban material detail, and a sense of immense scale achieved through believable streets, facades, traffic patterns, and tiny witnesses. Build a downtown renovation zone where scaffolding, cranes, and office towers rise into a vaulted urban canopy over the street.
Subject:
The subject is the city itself as a time-distorted machine, with roads, bridges, towers, or plazas curling into impossible but coherent formations. Human figures should remain small, secondary, and practical, acting only as scale markers for the urban transformation. Construction workers and orange cones stay tiny at ground level, reinforcing the megastructure above.
Important details:
The signature rule is to make the impossible read as engineered reality rather than fantasy painting. Preserve concrete, glass, asphalt, railings, shadows, and structural rhythm with enough discipline that the folded geometry feels like a captured phenomenon. Avoid cartoon simplification, magical particles, or decorative sci-fi ornament. The vault should feel assembled from practical urban systems such as scaffold tubes, netting, temporary lights, and tower cores.
Use case:
Time-Folded Urban Spectacle single study 8/9.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no fantasy castle shapes, no random levitating debris without structure, no superhero cues, no branded billboards, no unreadable poster text, no broken anatomy, no low-detail background, no accidental text fragments.Cómo usar
- Copia el prompt original completo, incluidos los parámetros.
- Adapta el tema, el texto y la composición a tus necesidades.
- Usa una herramienta externa compatible; comprueba el acceso y el resultado.
Antes de empezar
- Comprueba si necesita una imagen de referencia. La compatibilidad no demuestra qué modelo generó el ejemplo.
- Revisa nombres, cifras y textos; el resultado puede variar entre versiones del modelo.





