Which AI model works best for this prompt?

It’s optimized for Sora 2 but works with any text-to-image model supporting detailed scene prompts.

Can I modify the prompt variables?

Absolutely—replace `{subject}` and `{location}` with any person and setting you like.

What kind of output will I get?

A realistic candid photograph featuring a person looking down at their phone in the specified environment.

Candid Subway Photo Prompt | Create Realistic AI Images | sora2hub

Candid Subway Photo Prompt — Sora 2

Generate ultra-realistic candid photos of people on their phones in subway settings with this flexible prompt—perfect for urban photography enthusiasts.

提示词 · EN

一位 argsubject: 美丽的女性 正低头看着手机，身处 arglocation: 地铁上，以抓拍照片的风格呈现。

关于这个提示词

Generate ultra-realistic candid photos of people on their phones in subway settings with this flexible prompt—perfect for urban photography enthusiasts. 它更适合作为 GPT Image 2 的概念艺术起点：先保留画面结构、主体关系和风格约束，再替换成你的品牌、人物或场景。

使用时优先替换 subject、location 等占位符，并保持镜头、构图和材质描述的顺序不变。这样更容易获得稳定的画面结果，也方便继续做多版本对比。

如何使用这个提示词

Copy the prompt from this page.
Open Sora 2 or your preferred AI image generator.
Paste the prompt and customize the {subject} and {location} placeholders (e.g., "businessman", "train station").
Generate your image and download the candid-style photo.