OpenZhenZhen AI Prompt Gallery · gpt-image-2.5 · image
Neon Rainy Night Selfie by the Crosswalk
Neon Rainy Night Selfie by the Crosswalk — explore this gpt-image-2.5 creation on OpenZhenZhen, with its original prompt and a reusable generation example.

Prompt
Chinese prompt
创作一张4:5竖幅电影感街头人像。夜晚小雨刚停,一位成年女性站在人行横道旁的安全路缘处,远离行驶车道;她将智能手机自然举向镜头,像随手拍下的一刻,手机位于前景但不要遮挡面部。她直视镜头,带着温暖自然的笑容;深色外套和极简配饰,微乱发丝被微风轻轻吹动。 人物身份严格遵循已绑定的身份参考,保持真实面部特征、脸型和肤色。近距离广角街拍视角,略带自然透视张力,但面部和手部比例真实,不做夸张鱼眼变形。 背景为繁忙霓虹城市街道的柔焦光点、湿润路面反射、远处车辆灯流和高楼招牌色块;背景车流保持在安全距离,不出现近身车辆或可辨识的其他人物。雨后空气微湿,路面明亮反光。 青绿、青蓝霓虹映在深色街景中,与自然暖肤色形成对比;地面路灯反射提供柔和补光。使用柔和对比、提亮阴影、霓虹高光细微光晕和真实胶片颗粒,保留自然毛孔与手持抓拍的不完美感。主体面部清晰,手机和背景可略柔,景深自然。整体浪漫、真诚、充满都市夜生活氛围。 约束:只出现一位成年女性,站在安全的人行区域;保持人物身份、五官、手指和手机结构自然。不要交通危险动作、额外行人、文字、标志、水印、塑料磨皮、动漫风格、过饱和霓虹或虚假的景深边缘。
English prompt
Create a cinematic 4:5 vertical street portrait at night, just after a light rain. An adult woman stands safely at the curb beside a wet crosswalk, away from moving traffic. She naturally raises a smartphone toward the camera as if capturing a spontaneous moment; keep the phone in the foreground without covering her face. She looks into the lens with a warm, genuine smile, wearing a dark coat and minimal accessories. A light breeze moves a few slightly tousled strands of hair. Match the subject's identity strictly to the bound reference, preserving her real facial features, face shape, and skin tone. Use a close handheld wide-angle street-photo perspective with lively but natural depth; keep facial and hand proportions believable and avoid exaggerated fisheye distortion. In the background, show soft bokeh from a busy neon city street, wet-pavement reflections, distant traffic light trails, and colorful sign shapes among tall buildings. Keep traffic at a safe distance and show no recognizable extra people. The post-rain air feels lightly humid and the pavement glows with reflected light. Mix cyan and green-blue neon against the dark street with warm natural skin tones; use streetlight bounce from the pavement as a soft fill. Keep contrast gentle, shadows slightly lifted, neon bloom subtle, and film grain natural. Preserve visible skin texture and the imperfect feel of a handheld candid shot. Keep the face sharp while the phone and background may fall slightly soft with natural depth of field. The mood is romantic, candid, and unmistakably urban nightlife. Constraints: show one adult woman in a safe pedestrian area. Preserve identity and keep facial features, fingers, and phone geometry natural. No dangerous traffic interaction, extra pedestrians, text, logos, watermark, plastic skin smoothing, anime styling, oversaturated neon, or fake depth-of-field edges.
Explore more AI image and video prompts · Try image generation · Try video generation · Create on the AI canvas
Related works
- Bright Outline Paper Airplane Sticker Icon
Bright Outline Paper Airplane Sticker Icon — explore this gpt-image-2.5 creation on OpenZhenZhen, with its original prompt and a reusable generation example.
- Fluffy Sofa Graffiti Portrait in a White Studio
Fluffy Sofa Graffiti Portrait in a White Studio — explore this gpt-image-2.5 creation on OpenZhenZhen, with its original prompt and a reusable generation example.
- Cinematic Visual Sequence Inspired by User Lyrics
Cinematic Visual Sequence Inspired by User Lyrics — explore this gpt-image-2.5 creation on OpenZhenZhen, with its original prompt and a reusable generation example.
- Original chalk drawing of a female sea explorer on a blackboard in an empty classroom
Original chalk drawing of a female sea explorer on a blackboard in an empty classroom — explore this gpt-image-2.5 creation on OpenZhenZhen, with its original prompt and a reusable generation example.
- Original Adventurer Chalk Drawing on an Empty Classroom Blackboard
Original Adventurer Chalk Drawing on an Empty Classroom Blackboard — explore this gpt-image-2.5 creation on OpenZhenZhen, with its original prompt and a reusable generation example.
- Portrait Through a Car Window in Retro Night Breeze
Portrait Through a Car Window in Retro Night Breeze — explore this gpt-image-2.5 creation on OpenZhenZhen, with its original prompt and a reusable generation example.