# GPT Image Hub prompt

## 도쿄 거리에 앉아 있는 애니메이션 소녀
- URL: https://www.gptimagehub.com/ko/library/anime-girl-sitting-in-tokyo-street
- Generator URL: https://www.gptimagehub.com/ko/generate?promptId=oss-ym-g2-21955
- Markdown URL: https://www.gptimagehub.com/api/agent/markdown?path=%2Fko%2Flibrary%2Fanime-girl-sitting-in-tokyo-street
- Category: 애니메이션 / 만화
- Model: gpt-image-2
- Aspect ratio: 3:4
- Style: anime
- Tags: 애니메이션, 합성, 거리 사진, 도쿄, 믹스드 미디어, 캐릭터 디자인
- Updated: 2026-10-27T02:30:23.098Z
- Created: 2026-10-27T02:30:23.098Z

2D 애니메이션 캐릭터를 실사 도쿄 거리 배경에 합성하고 호기심 어린 행인들의 반응을 담아냅니다. 도시 위치, 간판 텍스트, 보행자 수를 변경해 보세요.

### Full prompt
```text
Using the reference photo as the character base, composite the 2D anime character into a realistic city street scene.

Subject: A full-body front-facing anime girl sitting on the sunlit brick pavement facing the camera, with one knee raised and the other leg folded outward. Preserve her blue angelic streetwear design, adding oversized blue-and-white platform boots, long trailing ribbon straps, a thin white halo above her head, and a small white plush accessory held near her raised knee.

Environment: A photorealistic Japanese urban shopping district resembling [Shibuya, Tokyo], captured from a low street-level perspective. Feature tall glass buildings, vibrant storefront signage, billboards, and a prominent store sign reading "[TSUTAYA]".

Interaction & Crowd: Surround her with [10] photorealistic pedestrians in a loose circle, some holding smartphones to take photos or staring in curiosity. Pedestrians have realistic clothing and shadows, with faces slightly blurred or obscured.

Style & Lighting: Crisp 2D anime character with clean line art and cel shading seamlessly composited onto a photorealistic background. Match realistic natural daylight, ground contact shadows, and depth of field so the character feels physically grounded in the real world.

Constraints: Vertical 3:4 framing with the character centered and positioned slightly low. Do not convert the anime character into a photorealistic human. No watermarks, text overlays, or UI elements.
```