# GPT Image Hub prompt

## 심야 자판기 골목 인물 사진
- URL: https://www.gptimagehub.com/ko/library/late-night-vending-machine-alley-portrait
- Generator URL: https://www.gptimagehub.com/ko/generate?promptId=oss-ym-g2-28715
- Markdown URL: https://www.gptimagehub.com/api/agent/markdown?path=%2Fko%2Flibrary%2Flate-night-vending-machine-alley-portrait
- Category: 인물 및 사진
- Model: nanobananapro
- Aspect ratio: 4:5
- Style: photorealistic
- Tags: 자판기, 야간인물사진, 골목길, 스냅사진, 네온조명, 스트리트스타일
- Updated: 2026-10-28T05:36:08.098Z
- Created: 2026-10-28T05:36:08.098Z

늦은 밤 불 밝힌 자판기 앞에서 음료를 고르는 순간을 포착한 사실적인 스냅 포트레이트입니다. 인물, 골목 분위기, 손에 든 소품을 변경해보세요.

### Full prompt
```text
Photorealistic lifestyle portrait of an adult [woman] at night.

- Setting: A dense [Japanese vending-machine alley] at night, featuring multiple glowing drink machines, cool white and blue light panels, colorful product-window glow, glossy side panels, narrow concrete walls, exposed pipes, capsule toy machines, wet pavement reflections, and layered alley depth receding into darkness. Distinctly urban, atmospheric nighttime scene with artificial illumination. Fictional, unreadable product visuals with no real logos or legible text.
- Composition & Camera: Vertical 4:5 lifestyle framing at machine-button height, cropped from mid-thigh or waist to head, with the subject occupying 65–75% of the frame. Strong layered depth: softly blurred machine edge or wet ground reflection in the close foreground, the subject in the middle ground, and distant illuminated machines and pipes in the background. Camera positioned slightly beside the machine display.
- Pose & Action: The subject stands grounded naturally close to the machine, slightly leaning forward to inspect the drink choices. One hand hovers over or presses a selection button, while the other hand holds [coins, a phone] or rests relaxed near her side.
- Gaze: Head remains angled toward the vending machine panel, while only her eyes glance subtly toward the lens, maintaining a candid three-quarter angle.
- Constraints: No posed dramatic twists, no full face turn toward the lens, no wide-angle alley overviews, no real brand logos or readable text.
```