GPT Image 2.0 Prompt - Low-Angle Candid Night Street Action Portrait
- Model
- GPT Image 2.0
- Aspect ratio
- auto
- Quality
- 1K
Prompt
[Location]: subway exit / vending machine / bus stop / convenience store
[Outfit]: jacket + fitted top + mini skirt/shorts + knee-high socks + sneakers
[Action]: crouching, sitting low, bending down, fixing hair/shoes, caught mid-motion
Create a 9:16 ultra-realistic night portrait of an adult East Asian woman in her early 20s. Use a close 28–35mm perspective with a slightly low camera angle, letting one knee or leg enter the foreground for stronger depth.
The body should stay in an unfinished everyday action rather than a perfect model pose — low center of gravity, natural hip/waist twist, loose shoulders, hair slightly messy from movement, then suddenly looking directly into the camera.
Use harsh on-camera direct flash as the main light. Keep the subject much brighter than the dark environment, with realistic highlights on skin, hair and fabric, subtle high-ISO grain, imperfect nighttime exposure, and natural background lights.
Prioritize real skin texture, visible pores, natural flyaway hair, believable clothing folds and anatomy. Avoid plastic skin, beauty filters, studio lighting, CGI look, stiff posing, extreme wide-angle distortion, or overly cinematic grading.
Use this image prompt
- 01
Start with the prompt
Copy the prompt, or open it with the generate button.
- 02
Make your changes
Adjust the subject, setting, or details. Add reference media if needed.
- 03
Check before generating
Review the model and output settings. Results can vary between generations.







































