
Documento original
GPT Image 2 · en
Scene:
A square photorealistic image of a fictional rail seat reservation card clipped to a train seatback tray. The card is printed on satin paper with rounded corners, tiny punched holes, and a folded crease. Behind it, the train interior is softly blurred: fabric seat, window light, and a metal tray hinge. The camera is front-facing so the text plane is readable.
Subject:
Create a fictional rail reservation specimen. The exact readable text should include: SPECIMEN, RAIL PASS, TRAIN T26, CAR 05, SEAT 08A, FROM NOVA, TO HARBOR, DATE 23 APR 2026, DEPART 14:20, SAMPLE ONLY. Add a decorative non-scannable barcode-like strip and a simple carriage seat diagram made of rectangles. Do not include real railway operators, real station names, passenger name, ticket number, QR code, payment data, or terms and conditions.
Important details:
The test is transit-document realism in context: crisp ticket typography, small icon alignment, paper thickness, punched holes, clip shadow, mild paper bend, and believable train-seat materials. The reservation must look product-real but clearly fictional and non-operational. Keep the specified route and seat fields readable; all other background marks should be abstract texture. No real transport brand identity should appear.
Use case:
Pseudo-Real Documents single study 8/8, testing readable rail reservation layout, satin card material, transport context, and safe fictional ticket design.
Constraints:
No readable text except the specified rail reservation strings, no watermark, no logos, preserve exact 1:1 aspect ratio, no real rail company, no real station names, no passenger name, no ticket number, no QR code, no scannable barcode, no payment data, no extra schedule text.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.





