109내 사진 무녀 코스프레

Generate this image now. Draw it directly - do not describe it, explain it, or ask questions.
Output size 2480x3508 pixels, A4 portrait, 300dpi, print-ready poster.

[Purpose]
Make exactly ONE finished portrait photograph, ready to post to Instagram without cropping, stitching, a second image, code or another turn. The person in the ONE attached photo becomes the subject of an original shrine-maiden-inspired cosplay editorial. The output is a believable live-action photograph, not anime, illustration, figurine, movie poster, contact sheet or costume catalog. Keep this one archetype and one photographic treatment fixed. Do not ask the user to choose an art style, costume type or location.

[Identity]
The user attaches the photo in the SAME message as this prompt. Preserve the photographed person's apparent age, facial structure, skin tone, hairstyle and any glasses; in particular retain "【사진에서 살릴 특징】" if it is really visible. Maintain identity even when the costume changes. If the stated detail contradicts the photo, the photo wins. Keep the exact number of principal people in the attachment; never invent a companion. This card needs a visible person, so if the attachment has no person, ask for a suitable photo rather than inventing a face. Do not turn an adult into a child or alter body proportions for glamour.

[Scene]
Stage a single candid full-to-three-quarter-length portrait in a quiet, invented Japanese-style courtyard in early autumn. The subject wears a respectfully designed, unbranded white wrap top and deep vermilion pleated hakama-inspired skirt with realistic fabric weight and ties. This is an original costume design with no copied anime uniform, emblem or named character. The person holds a small plain wooden charm at waist height; the charm has no printed writing. One stone path and subdued maple leaves provide depth. There is no large gate or famous shrine landmark to suggest a real place the person visited. The scene looks like a photo shoot, never a documentary claim.

[Pose and expression]
Use the input "【표정】" only to guide the subject's facial expression. If it says "알아서 골라줘", choose a relaxed expression supported by the source face and print none of those instruction words. Keep a natural pose derived from the source photo where possible. Both hands must have coherent fingers; sleeves and fabric folds should follow the arm positions. If the source is a close selfie, retain the face accurately while extending the composition in a plausible way, without inventing distinctive tattoos or jewelry.

[Text]
At 69–77% image height, add exactly ONE short Korean line "【사진에 쓸 한 줄】" in large, clear, slightly irregular GRAPHITE PENCIL handwriting on a quiet dark patch. If the input says "알아서 골라줘", invent a brief original line that fits a playful costume portrait, then render only that line. Never print the instruction phrase. The handwriting is part of the final photo, not a promotional label. Keep the upper 12% and lower 15% free of essential lettering for the site's listing overlays. Do not print the user's identity notes, a fake Japanese inscription, date, location, brand, photographer credit or watermark.

[Photo finish]
Use one live-action editorial look: natural lens perspective, fine facial detail, gentle depth of field, believable skin texture, restrained autumn amber and vermilion color, soft overcast daylight and accurate shadows under the shoes. The white fabric retains shape and folds instead of becoming a blown-out flat patch. Keep the person and one line readable at small list size. Avoid plastic skin, heavy beauty filtering, neon outlines, supernatural effects, weapon props and false signage.

[Avoid]
No specific film, manga, game, character, actor, artist or studio. No multiple shots, split before-and-after, cosplay guide, 3D render, chibi transformation, fantasy battle, altered person count, invented event, extra hands or malformed clothing. Proofread each Korean syllable. Return the single finished image only.