Candid Winter Street Smartphone Travel Portrait Prompt

Here's the thing.

@saniaspeaks_ (𝗦𝗮𝗻𝗶𝗮) opens with one line, Created with GPT image 2, then drops the entire prompt as a structured JSON object right in the tweet body. No need to dig into images or links, the deliverable is the prompt and the prompt is the post. Two reference frames come with it, both of the same woman in the same scene.

The setup is a candid full-body outdoor smartphone portrait on a quiet urban street in soft winter afternoon sunlight. One young woman, long straight dark-brown hair flowing past her shoulders, gentle cheerful expression, eyes softly closed, head tilted slightly to one side, playful candid pose. Standing comfortably with arms crossed while lifting one leg slightly behind her. Wearing an oversized pastel lavender-blue button-up shirt over a delicate white lace-trimmed top, a short flowing white tiered skirt, a small cream-colored backpack hanging over one shoulder, clean white sneakers.

The background is what makes the scene read as candid street rather than staged portrait. Old textured cream-colored wall with geometric square relief details on the right, bare climbing vines dried against the wall, small dark wall sign, parked Range Rover with a Korean license plate (37허 6318) half-visible on the left, narrow urban street receding into the distance, bare winter trees and old residential buildings beyond. Soft winter afternoon sunlight raking from the left, casting warm light onto her hair with subtle glow, gentle natural shadows, faint realistic lens flare on the upper-left of the frame. Skin texture has subtle imperfections, fabric texture is realistic, the whole frame has the soft pastel color grading and subtle vintage film grain of an authentic mobile-camera photo. Aspect ratio 4:5 vertical.

The two frames read as the same moment, two beats of the same pose. Frame 1 has her right arm extending toward the camera in a soft reaching gesture, eyes closed mid-smile, head tilted left, her long hair half-covering her left cheek. Frame 2 has her arms tucked down together in front, eyes open and looking straight at the lens, the smile smaller and more direct, the camera a half-step closer to her face. Same outfit, same wall, same warm afternoon light. Just the same scene shot twice with a small change of body language.

The prompt, as posted. The author ships it as a structured JSON object with explicit blocks for prompt text, subject, outfit, background, lighting, photography, mood, quality, aspect ratio and a separate negative prompt. That structure is the point of the share, it lets you swap individual blocks without rewriting the whole thing.

{
  "prompt": "A highly photorealistic full-body outdoor smartphone photograph of a young woman standing on a quiet urban street during soft winter afternoon sunlight. She has long, straight dark-brown hair flowing naturally over her shoulders, with realistic individual hair strands. She has a gentle, cheerful expression with her eyes softly closed and her head tilted slightly to one side, creating a playful and candid pose. She is standing comfortably with her arms crossed while lifting one leg slightly behind her.",
  "subject": {
    "gender": "young woman",
    "hair": "long, straight dark-brown hair flowing naturally over her shoulders",
    "expression": "gentle, cheerful, relaxed smile with eyes softly closed",
    "pose": "head slightly tilted to one side, arms comfortably crossed, one leg lifted slightly behind her in a playful candid pose"
  },
  "outfit": {
    "outerwear": "oversized pastel lavender-blue button-up shirt",
    "top": "delicate white lace-trimmed top",
    "bottom": "short flowing white tiered skirt",
    "bag": "small cream-colored backpack hanging over one shoulder",
    "shoes": "clean white sneakers"
  },
  "background": {
    "location": "quiet urban street",
    "wall": "old textured cream-colored wall with geometric square details",
    "details": [
      "bare climbing vines",
      "small dark sign",
      "parked car",
      "subtle urban street elements"
    ]
  },
  "lighting": {
    "time": "soft winter afternoon",
    "style": "natural warm sunlight",
    "effects": [
      "subtle warm glow around her hair",
      "gentle natural shadows",
      "faint realistic lens flare"
    ]
  },
  "photography": {
    "camera_style": "natural smartphone photography",
    "aesthetic": "candid travel-fashion photography",
    "skin_texture": "realistic natural skin texture with subtle imperfections",
    "focus": "shallow depth of field",
    "color_grading": "soft pastel color grading",
    "film_effect": "subtle vintage film grain",
    "realism": "natural body proportions, realistic fabric texture, realistic lighting, authentic mobile camera appearance"
  },
  "mood": "dreamy, peaceful, cheerful, soft and candid",
  "quality": "highly photorealistic, natural-looking, realistic smartphone photo",
  "aspect_ratio": "4:5",
  "negative_prompt": "cartoon, illustration, CGI, 3D render, overly smooth skin, plastic skin, artificial face, distorted body, extra fingers, extra limbs, blurry face, unrealistic proportions, oversaturated colors, studio lighting, heavy makeup, exaggerated lens flare"
}

A small note on the original. The author is 𝗦𝗮𝗻𝗶𝗮, posting as saniaspeaks_, an X account publishing GPT Image 2 prompt shares in the same candid travel-portrait lane as Aqsa's kitchen prompt and a dozen others in this library. The post is in English, the prompt ships in English as structured JSON, and the model is ChatGPT Image 2. The two reference frames are framed for the same scene shot twice with slightly different poses, the prompt itself is generic enough to be re-shot in any quiet winter street with any cream wall and any parked car as a backdrop.

At the time of writing, the post sits at 86 likes, 11 retweets, 29 bookmarks and 3,160 views, with 29 replies under it. Original thread here, x.com/saniaspeaks_/status/2096576034900471875