Skip to main content
Prompts Hyper-Realistic Golden Hour Portrait Generator

model image_generation template risk: low

Hyper-Realistic Golden Hour Portrait Generator

Creates a hyper-realistic portrait of a specified gender in tailored casual wear, positioned leaning against a weathered brick wall in golden hour light, with natural skin texture,…

PROMPT 

Hyper-realistic portrait of a ${gender:man} in tailored casual wear (dark jeans, quality sweater) ${position:leaning against weathered brick wall} in golden hour light. Maintain original face structure and features. Create natural skin texture with subtle pores and realistic stubble. Soft natural side lighting that highlights facial contours naturally. Street photography style, slight grain, authentic and unposed feel.

INPUTS

gender REQUIRED

Gender of the portrait subject

e.g. man

position REQUIRED

Position and setting of the subject

e.g. leaning against weathered brick wall

EXPECTED OUTPUT

Format
image_prompt

CAVEATS

Dependencies
  • Requires original face structure and features to maintain.
Missing context
  • Reference to source image for 'Maintain original face structure and features' (implies img2img mode).
  • Image generation parameters (e.g., aspect ratio, sampling steps, CFG scale).
  • Negative prompt to avoid common artifacts.
Ambiguities
  • Placeholder syntax like '${gender:man}' is non-standard and may confuse users expecting simple {gender} variables.

QUALITY

OVERALL
0.87
CLARITY
0.88
SPECIFICITY
0.92
REUSABILITY
0.87
COMPLETENESS
0.82

IMPROVEMENT SUGGESTIONS

  • Standardize placeholders to '{gender}' and '{position}' with separate documentation for options.
  • Add explicit aspect ratio, e.g., '--ar 2:3' for portrait orientation.
  • Specify sweater details, e.g., 'navy cashmere sweater' for higher precision.
  • Include strength or denoising level for img2img consistency.

USAGE

Copy the prompt above and paste it into your AI of choice — Claude, ChatGPT, Gemini, or anywhere else you're working. Replace any placeholder sections with your own context, then ask for the output.

MORE FOR MODEL