GPT Image 2Prompts

Create a photorealistic disposable-camera-style cosplay portrait of a young adult woman wearing a futuristic white cyber maid costume in a modern high-rise apartment kitchen at night. She stands centered from mid-thigh up, facing the camera, holding a round stainless-steel serving tray with both hands at waist level. Her outfit is glossy white with silver hardware: a white ruffled maid headband, long dark wavy hair, white structured bodice with translucent ruffle shoulder straps, a white choker with a large metal ring, silver chain details, metallic wrist cuffs, an upper-arm band, white apron panel, short layered skirt with translucent frills, and black-and-silver garter-like cyber accessories. Use {argument name="costume color" default="white and silver"} as the main outfit palette and {argument name="hair color" default="long dark brown"} hair. The setting is a sleek stainless-steel kitchen with a range hood and reflective metal backsplash on the right, warm under-cabinet lighting, wood lower cabinets, and a large dark window on the left showing a nighttime city skyline with scattered building lights. The image should feel like an amateur flash photo taken on a compact disposable camera: direct flash, slight grain, warm indoor highlights, high contrast, mild lens distortion, candid cosplay event energy, realistic skin and fabric textures. Keep the composition vertical, intimate, and documentary, with the subject occupying most of the frame. Add a large soft rectangular blur over the face to anonymize identity, while leaving hair, costume, and background visible. Avoid text, logos, watermarks, extra people, or exaggerated fantasy anatomy.

Create a nostalgic disposable-camera style night portrait of a young woman cosplaying as a magical girl in a pink-and-white idol costume, photographed outdoors in a dark park or garden with warm fairy lights wrapped around trees in the background. She stands centered in a three-quarter waist-up pose, holding a heart-shaped magical wand with a gold handle and pink heart ornament in both gloved hands near her chest. Her outfit is glossy satin and organza: a pastel pink flared skirt, white iridescent bodice, large pink bow with a red heart jewel at the chest, frilly puff sleeves, lace trim, translucent ruffled overskirt panels, white satin gloves with pink lace cuffs, and a matching headband with a large pale pink bow. She has long dark wavy hair falling over her shoulders, but her face is intentionally covered by a soft square privacy blur. Use harsh direct flash, high ISO grain, slight underexposure, warm orange bokeh lights, deep black shadows, candid amateur cosplay event photography, 4:5 vertical composition, realistic fabric shine, slightly imperfect framing, no text or watermark. Customize the costume palette as {argument name="costume colors" default="pastel pink and white"}, the prop as {argument name="magical prop" default="heart-shaped wand"}, the setting as {argument name="night location" default="dark park with fairy lights"}, the camera look as {argument name="camera style" default="disposable film camera flash"}, and the face treatment as {argument name="face treatment" default="soft square privacy blur"}.

Generate a high-resolution vertical 3D character poster blending Pixar styling with designer-toy sensibility. Subject: {argument name="spiritual keyword" default="SPIRITUAL WORLD KEYWORD"} — rendered as a tall, lean, mildly exaggerated Pixar-style 3D figure. Keep the celebrity's signature face, hair, posture, and wardrobe intact, but translate them into clean, simplified, blocky designer-toy proportions. The figure must stay instantly recognizable, shown full-body, in a pose that conveys their personality, aura, or public persona. Pose: A natural yet expressive stance, with a touch of theatricality where it suits the celebrity. Head tilt, hand gesture, and overall posture should together broadcast who this person is. Background: For every generation, pick one vivid, bold, highly saturated solid color matching the celebrity's temperament and the emotional register of their spiritual world. Acceptable directions: hot red, saturated yellow, blood orange, cobalt blue, vivid green, electric purple, deep cyan, and similar. The backdrop should read flat, clean, graphic, and bold, like a gallery poster or a conceptual exhibition print. No gradients, no muted tones, no dull cinematic gray. The floor can take a related hue, slightly darker or lighter than the wall, with a faint reflective sheen. Keep the environment minimal, infinite, and free of clutter. Lighting: One strong directional source, coming from a side angle or from above. It must cast a sharp, oversized shadow on the wall behind the figure. The lighting needs to feel graphic, dramatic, and deliberate, spotlighting both the character and the symbolic shadow. Core idea, shadow as spiritual projection: The shadow behind the figure must NOT mirror the actual body silhouette. Instead it becomes a symbolic projection of the celebrity's inner world, legacy, psychological core, or most iconic creative or spiritual motif. It should be large, dominant, visually arresting, and blend organically across wall and floor. It should look as if the celebrity's invisible inner life has been projected outward into a visible form. The shadow must visually embody the supplied spiritual-world keyword: {argument name="spiritual keyword" default="SPIRITUAL WORLD KEYWORD"} This symbolic shadow can take an abstract, surreal, poetic, metaphorical, or object-based form, whatever best captures the celebrity's essence. It must function as more than decoration: it is the true conceptual heart of the poster. Typography: Use the spiritual-world keyword, NOT the celebrity's name, as the main poster headline. Render: {argument name="spiritual keyword" default="SPIRITUAL WORLD KEYWORD"} Pick a clean minimalist typeface, sans-serif or serif depending on mood. Place it in a balanced corner slot, usually top-left, with elegant hierarchy and strong poster layout. The celebrity's name may appear smaller in a secondary position if needed, but the keyword must always be the dominant title. Rendering: Pixar-grade 3D rendering

Create a cinematic realistic character design board for a high-budget neo-noir film production set in a rain-soaked futuristic city. Use a dark charcoal and electric cyan color palette with neon reflections in the background. Avoid generic grids or symmetrical layouts; composition should feel like a stylized director’s pitch board. Design a grounded human character with realistic anatomy, subtle imperfections, and strong emotional presence. Include full-body turnarounds, expressive head angles, cinematic portrait, wardrobe breakdown, fabric texture detail, and production notes. Background: blurred cyberpunk city lights, wet glass reflections, moody atmosphere, soft neon glow. Style: semi-realistic cinematic realism, high contrast lighting, shallow depth of field, film grain, emotional intensity.

Create a polished anime-style illustration of {argument name="character name" default="a teenage schoolgirl"} in a cute pastel pink bedroom, standing in the foreground and dramatically holding up a shiny metal spinning top toy on a black handle with its detached circular ring flying above it, connected by a loose black cord. The girl has very long dark navy-blue hair in high twin ponytails with loose strands, a red headband, small hair clips, and a pale blue flower hair accessory; her face is intentionally covered by a soft rectangular blur censor block. She wears {argument name="outfit" default="a cropped red track jacket with sleeve stripes and a blue pleated school skirt"}, and she poses with one arm raised gripping the toy while her other hand is open in surprise, creating an “oh no” action moment. Add exactly 4 yellow sparkle effects around the spinning toy: 1 large diamond sparkle near the window, 2 small sparkles beside it, and 1 small sparkle on the right. The room should be bright, cozy, and feminine: pink diamond-pattern wallpaper with pale heart/cloud motifs, a large sunlit window with pink curtains, a white bed with a plush teddy bear, two pillows colored orange and coral pink, a bedside table with a small pink lamp and round clock, a tall white wardrobe, a standing mirror, and a white desk setup with a computer monitor, keyboard, pink headphones, a small chibi figure, and a framed broken-heart picture on the wall. Use soft afternoon lighting, gentle bloom, warm pink color grading, clean linework, detailed cel shading, slight depth of field, and a square 1:1 composition. Keep the scene playful, chaotic, and cute, with no readable text and no watermark.

Please generate a vertical 2:3 high-definition image with the theme "Architectural Miniature Blueprint". Building Name: {argument name="Building Name" default="Nine-Story Demon Tower"}. The scene is a modern architectural design manuscript, covered with Chinese floor plans, cross-sections, elevations, structural details, dimension annotations, and construction instructions. A highly realistic 3D miniature model of ancient Chinese architecture is "built" out of the blueprint. The main building should occupy 65%–75% of the frame, appearing larger, taller, and more visually striking. Use a low-angle 3/4 top-down perspective to highlight the model's height, volume, and realistic shadows. The building must match the authentic features of the chosen theme and appear to be under construction, about 70%–85% complete. Add miniature construction details: tiny construction workers, scaffolding, wooden frames, ramps, pulleys, wooden beams, stones, tiles, material piles, wheelbarrows, horse-drawn carts, half-finished steps, roof construction areas, etc. The details should revolve around the building itself, forming a simulated miniature construction world. The overall style is a modern architectural firm's concept proposal board + museum-grade miniature model. The blueprint is clean and premium, with colors mainly in off-white, light gray, pale blue, and gray-black line art. All text must be in Chinese, with the main title "[Building Name]" and the subtitle "Schematic Diagram of Architectural Construction Process". Use soft neutral lighting; the image should be clear, modern, restrained, and sophisticated.

Create a cinematic Minecraft-inspired voxel wedding scene that blends realistic newlyweds with a blocky game world. In the center foreground, show a groom in a black suit lifting and carrying a bride in a flowing white lace wedding dress; the couple should be the only mostly realistic photographic figures, with softly anonymized/blurred faces, while the environment and guests are fully voxel/block-style. The bride wears a long white veil, delicate hair accessory, sleeveless lace bodice, and voluminous skirt lifted by motion; the groom faces away slightly, holding her romantically. Place them on a grassy aisle covered with scattered white flower petals, leading toward a square wooden floral wedding arch decorated with dense green leaves, white blossoms, and hanging flower strands. Surround the aisle with exactly two rows of wooden benches, one row on the left and one on the right, filled with blocky Minecraft-style wedding guests in dark formal suits and white shirts, clapping and smiling; include visible lanterns on posts near the seating. Add abundant blocky flowers along both sides of the aisle, mostly white daisies with a few pink blossoms. The background should show cubic trees, layered voxel foliage, wooden blocks, a bright blue sky with rectangular clouds, warm daylight, soft depth of field, and a joyful romantic atmosphere. Use a square-to-landscape 4:3 composition, centered perspective down the aisle, high-detail voxel textures, vibrant greens, warm highlights, realistic fabric detail on the bride and groom, and a seamless blend of real wedding photography with Minecraft game-map aesthetics. No text, no watermark, no UI elements.

Ultra-detailed cinematic cyber-grunge fashion poster of a young man walking confidently toward the camera, centered in frame, full-body composition, intense eyes, soft masculine facial structure, messy wind-swept dark brown hair, subtle stubble, pale skin tone, neutral expression, dramatic editorial lighting. He wears layered black techwear fashion: distressed oversized trench coat with torn edges and flowing fabric motion, heavy draped scarf around neck, oversized black shirt, tactical cargo pants with straps and stitched textures, black leather combat boots, crossbody utility strap, futuristic dystopian styling. Poster aesthetic inspired by avant-garde streetwear campaigns, luxury fashion editorials, neo-Tokyo cyberpunk minimalism, monochrome visual identity, dark moody atmosphere, high contrast black and white palette. Background split vertically into two contrasting halves: Left side: textured off-white concrete wall with grunge paper texture, ink splatters, paint drips, abstract geometric diagrams, thin technical lines, circles, symbols, barcode graphics, minimalist UI-inspired design elements, subtle distressed typography layout. Right side: deep matte black background with smoky haze, cinematic fog, soft atmospheric dust, shadow gradients, minimal texture. Typography integrated naturally into the poster design in English, Korean, and Japanese: Large vertical Japanese kanji typography on right side with rough painted texture, small vertical Japanese quote beside character, Korean phrases placed in lower left and lower right corners, bold English headline text in upper left: “{argument name="headline" default="BEYOND THE SHADOW"}” “{argument name="slogan" default="I WALK ALONE, BUT I CARRY EVERY VERSION OF MYSELF."}” Additional tiny editorial text, coordinates, symbols, futuristic labels, and graphic interface elements scattered subtly across composition. Floor is glossy black with wet reflective texture reflecting the boots and coat movement. Coat fabric flows dynamically as if blown by wind. Composition feels like a premium fashion campaign poster mixed with dystopian anime aesthetics and underground streetwear branding. Ultra-realistic photography, hyper detailed skin texture, cinematic shadows, volumetric smoke, sharp focus, luxury magazine quality, dramatic depth, 8K detail, high-fashion editorial look, realistic fabric physics, perfectly balanced negative space, clean typography hierarchy, sophisticated graphic design composition, poster-ready layout, vertical 4:5 aspect ratio.

Create a cute retro cartoon travel-poster cityscape of {argument name="city name" default="Chengdu"}, in a naive children's picture-book style with ultra-simplified chunky architecture, exaggerated proportions, bold thick black outlines, bright flat colors, halftone comic-paper texture, slight ink misregistration, and a sunny turquoise-blue sky. Use a vertical 3:4 poster composition. At the top, place huge arched yellow bubble letters reading “{argument name="headline text" default="CHENGDU"}” with black shadow, and beneath it red Chinese characters “成都” flanked by two red dots. Make the clear central focus a stylized panda-head building inspired by Chengdu panda landmarks: a large geometric white-and-black panda face structure at the bottom center with a blue glass entrance, stairs, shrubs, and flowers, with a tall slim TV tower rising directly behind it. Include exactly 3 panda motifs: the giant panda-head building, a small cheerful waving panda mascot in the lower left foreground beside a sign, and a red circular passport-stamp graphic in the upper right that says “WELCOME TO CHENGDU” with a panda face. Include exactly 5 puffy cream clouds with peach shading: one large cloud on the upper left, one large cloud on the upper right, one partial cloud entering from the lower left edge, one small cloud near the left-middle skyline, and one small cloud near the right-middle sky. Build a simplified skyline along the bottom with colorful rectangular buildings in teal, yellow, pink, orange, and blue, plus green treetops. Add a traditional Chinese pavilion on the lower left with green tiled roofs, red lanterns, and a small sign with Chinese-style characters; add a yellow vertical building sign on the right reading “太古里”; add a cream arched bridge over bright blue water on the lower right. In the lower left foreground, place a sign reading “I ♥ CHENGDU” next to the waving panda. Keep all forms rounded, toy-like, and geometric, with minimal detail, thick outlines, warm cheerful colors, vintage postcard energy, and no photorealism.

Create a vertical retro cartoon travel postcard for {argument name="city name" default="Chengdu"}. Use a bold vintage comic-book style with thick black outlines, bright pop-art colors, visible halftone dot texture, slightly grainy aged print, and playful exaggerated perspective. The sky is saturated cyan-blue with comic dots. At the very top, place huge arched yellow block letters reading {argument name="headline text" default="CHENGDU"}, with black shadow and outline; directly underneath add red Chinese characters {argument name="local city text" default="成都"} flanked by two red dots. In the upper right, add a red circular passport stamp that says “WELCOME TO CHENGDU” with a cute panda face and three wavy postal cancellation lines. The scene should show exactly 5 cream-and-peach puffy clouds: one large cloud on the upper left, one large cloud on the upper right, one partial cloud entering from the left middle edge, one small cloud left of center, and one small cloud right of center. Center the postcard around Chengdu landmarks: a tall TV tower rising from behind a giant faceted panda-head building, with a blue doorway at the panda’s mouth; make the panda head geometric and stone-like with black ears and eye patches. Along the lower skyline, include colorful simplified city buildings and green trees. On the left foreground, include a traditional Chinese pavilion with green tiled roofs, red lanterns, and a sign with Chinese characters. On the right foreground, include a yellow vertical sign with Chinese characters {argument name="landmark sign text" default="太古里"}. Across the bottom, draw a bright blue river with wavy white highlights and an arched cream stone bridge on the right. In the bottom left foreground, place a roadside sign reading “I ♥ CHENGDU” and a cute standing panda mascot waving beside it. Counted visible elements: 1 main headline, 1 Chinese subtitle, 1 red passport stamp, 5 clouds, 1 central TV tower, 1 giant panda-head building, 1 traditional pavilion, 1 vertical landmark sign, 1 river, 1 arched bridge, 1 “I love Chengdu” sign, and 1 waving panda mascot. Keep the composition busy but legible, like a cheerful tourist souvenir postcard, with no photorealism and no extra text beyond the specified lettering.

{argument name="Channel" default="CCTV5"} live broadcast, I am participating in the {argument name="Event" default="Olympic gymnastics competition"}, standing on the balance beam.

Create a highly detailed isometric miniature diorama of {argument name="city square" default="Krakow Main Market Square, Poland"} at warm golden lunch hour, viewed from above at a three-quarter angle like a handcrafted architectural model on a dark square display base. Show ornate European old-town architecture with pastel townhouses along the back and right edges, red tiled roofs, decorative gables, cobblestone paving, and soft shallow-depth-of-field lighting. The main landmarks should be a long Renaissance cloth hall with arcaded sides and an ornate roofline, plus a large brick Gothic basilica with exactly 2 unequal tall towers, dark spires, onion domes, and gilded details. In the open square include exactly 1 central stone monument with a standing statue, exactly 5 horse-drawn carriages, exactly 6 small market stalls or flower kiosks with pale yellow awnings, and exactly 7 white café umbrellas around outdoor seating. Fill the plaza with many tiny colorful pedestrians in summer clothing, clustered naturally around stalls, cafés, the monument, carriages, and church entrances. Use {argument name="visual style" default="ultra-detailed tilt-shift miniature, photorealistic 3D render"}, with warm sunlight, crisp shadows, rich brick and pastel colors, tiny flowers and greenery, and a charming lunch-time tourist atmosphere. Avoid readable signage, modern cars, oversized people, watermarks, or text overlays.

Based on 1-3 clear personal photos uploaded by the user, please generate a 2x2 four-grid collage portrait, with a single image ratio of 3:4, and the theme "{argument name="Theme" default="Playful Hanfu Graduation Photo"}". [Character Consistency] Strictly retain the user's real identity features, including face shape, facial feature proportions, eyes, nose, lips, skin tone, sense of age, basic hairstyle, and overall temperament. The four small images must clearly look like the same real girl, not turning into a stranger, not Westernized, not internet-celebrity-like, not overly smoothed, and no AI fake faces. [Overall Style] The entire set of images should be in the new Chinese graduation portrait style popular on Chinese social media in recent years, focusing on a playful, agile, energetic, cute, and endearing girlish graduation vibe. The picture should look like a set of relaxed, fun, and creative Hanfu graduation photos suitable for spreading during graduation season, overall clean and bright, with a slight Chinese sweet aesthetic. [Makeup] The makeup is sweet and clear girlish makeup. The base makeup is light and translucent, and the skin tone is fair and natural. Eye makeup is in light apricot pink, peach pink, and warm brown tones, with fine and soft eyeliner, naturally curled eyelashes, and slight aegyo sal. Blush is peach pink or apple pink, and lip color is cream pink, peach pink, or coral pink. The overall makeup feels young, agile, and energetic. [Hairstyle] The hairstyle naturally extends based on the original hair length in the uploaded photos, but prioritize styles more suitable for a floral graduation cap. - Long hair priority: half updo, double side braids, low double buns, light updo, retaining loose hair and an airy feel. - Medium-long hair priority: half tied hair, double braids, low updo. - Short hair priority: tucked behind the ears, short hair updo feel, French short curls paired with a small floral cap. Do not have a large area of loose hair in all four pictures, and do not let the hair press against the hat and face. [Clothing and Headdress] The clothing is a playful Hanfu graduation portrait style, using soft and bright color schemes such as light green, pinkish white, light yellow, milky white, and pale pink. The clothing is light, with embroidery, pibo (shoulder drape), wide sleeves, and pattern details. It must be paired with a floral bachelor's cap or a floral graduation cap, with fresh and light floral materials, and a small amount of tassels and pendants. [Scene and Props] The scene is a clean and bright Chinese light studio space, with the background mainly in off-white, warm white, and light wood colors, embellished with cherry blossom branches, small wooden tables, flower vessels, books, scrolls, and hanging tags. Props include writing brushes, scrolls, inscription plaques, small wooden tags, and flower branches. For text props, please use Traditional Chinese and brush calligraphy, with clear and readable short sentences, such as: "Happy Graduation", "Bright Future", "Good Luck", "Done Studying". [Lighting and Texture] Use soft and bright lighting, overall clear, clean, and soft, with clear facial features and complete details of clothing and headdresses. The picture has a slight soft-focus feel, but not hard and sharp, not studio-like, and not modern commercial advertising-like. [Four-Grid Fixed Shots] Please strictly follow the 4 fixed shots below to generate the four-grid, each with a 3:4 vertical composition: 1. Top Left: Close-up Main Visual Half-body close-up, the character looks at the camera from the front or a slight 3/4 angle, gently pinching a small wooden inscription plaque in her hand, with a gentle smile or natural sweet smile. Highlight the face, floral cap, makeup, and character recognition. 2. Top Right: Full-body Styling Full-body standing posture, showing the complete clothing and hat. Holding a writing brush in one hand and a vertical inscription scroll in the other, with a natural and bright expression, relaxed and naturally showing teeth while smiling. The body should not be straight and stiff, can slightly step out, slightly turn sideways, and shift the center of gravity to increase agility. 3. Bottom Left: Playful Action 3/4 sideways or half-sideways composition, the character lifts her sleeve, looks back, or slightly tilts her head, and can make richer expressions such as winking, pursing lips to smile, or smiling mischievously. Highlight the "playful feel" and action changes, and do not repeat the expression from the first picture. 4. Bottom Right: Scroll Interaction Sitting or leaning against a table half-body picture, holding a thread-bound book or scroll, with a gentle expression looking at the camera or looking down slightly at the book pages. Hand movements are natural, and the composition includes foreground flowers or desktop props to enhance layering. [Fixed and Unchanged] The following contents are unified across the four pictures: - The same real girl - The same clothing system - The same floral bachelor's cap system - The same type of sweet and energetic makeup - The same set of scene tones - The same type of Chinese prop system [Changes that Must be Reflected] The four pictures must be clearly distinguished: - Close-up / Full-body / Playful action / Prop interaction - Front / 3/4 side / Half side / Sitting - Sweet smile / Natural smile / Playful expression / Toothy smile Do not have all four look like ID photo front faces. [Avoid] Do not have all four looking straight at the camera Do not have all four standing Do not have all hair loose Do not use simplified printed text Do not use modern internet celebrity makeup Do not overly smooth skin Do not have stiff actions Do not have a cluttered background Do not include watermarks and logos

Use this character to create an image for "{argument name="Theme" default="This week's fortune omikuji image"}". IMPORTANT: Make the content completely different every time. The result must be randomly generated and not depend on past context or the content of the reference image. Do not use the reference image as is; change the situation, composition, facial expression, pose, background, props, and magazine design to match the omikuji result. For the overall fortune, use general fortune ranks found in Japanese omikuji culture. Place the overall fortune prominently at the top of the screen. Make the text design fun and information-rich, blending Japanese-style omikuji, magazine fortune-telling pages, variety shows, SNS-oriented images, and internet meme images. The character should speak their reaction or thoughts about the omikuji result in a speech bubble. Change the speech bubble content, facial expression, tension, and reaction every time. Include multiple fortune categories. Add small, cute, fluffy chibi-style cut illustrations to each category. Change the fortune categories, content, page composition, layout, color scheme, font-style expression, and amount of information every time. Include X (formerly Twitter) fortune. Overall atmosphere: - {argument name="Atmosphere" default="Cute, fluffy, slightly weird"} - Makes you want to share on SNS - Feels like the fortune-telling page at the back of a magazine - Heisei to Reiwa internet culture - High information density - Looks good in screenshots IMPORTANT: Significantly change the fortune content, text, chibi character actions, page layout, small jokes, and vibe every time. Do not repeat the same pattern. Negative elements: Avoid exact dates, days of the week, calendar notation, real-world date/time information, specific schedule notation, incorrect date generation, corrupted number expressions, meaningless date text, and unreadable calendar UI.

# New Chinese Style Academic Robe Graduation Portrait | 4 Individually Retouched Photos Please generate **4 different individually retouched graduation portrait photos** at once based on 1-3 clear personal photos uploaded by the user. Theme: "New Chinese Style [Degree Level] Academic Robe Graduation Portrait" Customizable parameters: - [Degree Level]: Bachelor / Master / Doctor - [Major Category]: Literature / Science / Engineering / Agriculture / Medicine / Management / Art / Law / Education / Economics / History / Philosophy / Military Science - [Main Color Scheme]: Default processed according to the common Chinese academic dress system, or fine-tuned according to the actual school version - [Prop Preferences]: Bouquet / Ancient scroll / Ancient books / Calligraphy brush / Inscription scroll / Small wooden sign / Floral hairpin fan - [Background Type]: A. Minimalist retro light brown studio background; B. Dreamy green and gold bokeh background - [Character Temperament]: Sweet / Cool / Gentle / Playful / Scholarly / Energetic ## 1 | Output Format Output 4 different individual photos at once. Each photo is an independent piece: - Vertical composition - 3:4 ratio - Mainly upper body / half body / 3/4 body - Highlight the character, hat, updo, floral hairpin, collar, cuffs, front placket, and scroll props ## 2 | Character Consistency Strictly retain the user's real identity features, including: - Face shape - Facial feature proportions - Eye shape - Nose shape - Lip shape - Skin tone - Sense of age - Basic hairstyle temperament - Overall demeanor All 4 images must clearly look like the same real girl. Do not turn into a stranger, do not Westernize, do not make it look like an internet celebrity, do not over-smooth skin, no fake faces, do not turn the character into an AI model. ## 3 | Core Style Positioning This is a set of New Chinese Style Academic Robe Graduation Portraits. The image must simultaneously have: 1. Clear graduation theme 2. Standard academic cap and robe recognition 3. New Chinese Hanfu details 4. Updo and real flower hairpins 5. Ancient scroll props 6. Gentle, exquisite, girlish, graduation season viral feel The visual logic is: Retain the main body structure of the academic robe, and add New Chinese aesthetics to the collar, cuffs, front placket, hat decorations, and props. ## 4 | Clothing Hard Requirements The clothing must be a New Chinese Style [Degree Level] Academic Robe. ### Must Retain - The main silhouette of a standard Chinese academic robe / graduation gown - Black or dark-colored academic robe main body - Wide sleeve structure of the academic robe - Shoulder and cape-like structure of the academic robe - Academic cap and tassel ### New Chinese Details to be Significantly Enhanced Please significantly enhance the Hanfu design in the following areas: - Collar - Cuffs - Front placket - Local structure on the chest - Local piping - Local embroidery - Frog buttons / Chinese knot buttons - Cloud shoulder-like details - Crossed collar or stand collar layers - Cuff lining layers The clothes should have a more obvious New Chinese exquisite feel, but still be recognizable at a glance as a graduation academic robe. ### Major Color Requirements [Major Category] corresponds to [Major Color]. [Major Color] is only used for: - Collar - Cuffs - Front placket - Local piping - Hat decoration floral materials - Small decorations near the tassel - Scroll binding ribbon Do not change the entire robe into a colorful Hanfu. ## 5 | Hairstyle Hard Requirements The hairstyle must be an Updo. Preferred choices: - Low updo - Low loose bun - Half-tied updo - Braided updo - Low rolled hair - Chinese updo with stray hairs Requirements: - Hair must be clearly gathered up - The updo structure must be visible behind the ears and at the back of the head - Can have a small amount of natural stray hairs to frame the face - If the user's photo has short hair, tuck the hair behind the ears - Do not have long hair draped extensively over the shoulders - Do not have completely loose hair - Do not have everyday loose draped hair - Do not have a wet hair look - Do not have modern internet celebrity big wavy draped hair ## 6 | Academic Cap and Real Flower Hairpins Must retain the standard square academic cap: - Black square cap - Clear cap shape - Clear tassel - Prominent real flower hairpins, but the flowers cannot completely cover the academic cap There must be obvious real flower hairpins at the hat and updo. Hairpin requirements: - More flower volume than ordinary small hair accessories - Must have a real flower texture - Can use pink, white, cream, light apricot small flowers - Have real petal layers - Have stamens, stems, green leaves, small flower branches - Can have a small amount of dangling flower branches or pearl-like small pendants - Floral decorations should be obvious, but do not block facial features and cap shape No plastic flowers, no cheap bead flowers, no ordinary hair clip feel. ## 7 | Prop Requirements The main prop must use an Ancient scroll / scroll. Preferred props: - Ancient scroll - Scroll - Calligraphy brush - Round fan - Thread-bound book - Small wooden sign - "Not studying anymore" inscription scroll Scroll requirements: - Paper is off-white, rice paper feel, slightly yellowed - Has wooden axis or scroll head - Can be gently tied with a pink ribbon - Can have faint patterns or landscape textures Allowed traditional Chinese calligraphy text: - 不讀了 (Not studying anymore) - 畢業快樂 (Happy Graduation) - 前程似錦 (Bright Future) - 金榜題名 (Succeed in the examination) - 書香長存 (Everlasting fragrance of books) The text must have a calligraphy feel, not modern poster fonts. ## 8 | Makeup Requirements The makeup is an exquisite but not overly heavy New Chinese girlish makeup: - Clear base makeup - Retain real skin texture - Soft eye makeup - Natural eyelashes - Bean paste pink, nude pink, soft mist rose pink lip color - Light blush - Gentle, clean, energetic No Korean internet celebrity heavy makeup, no photo studio heavy makeup, no over-smoothing, no plastic skin. ## 9 | Background and Photography Language The background uses a minimalist, clean set that highlights the character, no complex scenes needed, does not emphasize showing legs, no large space environment needed. Choose one uniformly from the following two types of backgrounds: - Type A: Minimalist retro light brown / milk coffee / oil painting feel studio background - Type B: Dreamy green and gold bokeh background, soft focus flower shadows, stronger glowing feel A small amount of the following can be added to the scene: - Bouquet - Flower branches - Ancient books - Round fan - Inscription wooden sign - "Not studying anymore" inscription scroll But overall keep a concise background + highlight the character and upper body, do not clutter the scene. Lighting and Image Quality - The image should have a retouched texture, real photography feel - Lighting is soft, delicate, bright - Can have a slight dreamy soft focus - Character's face is clear, background moderately blurred - The image has a slight creamy feel, airy feel, film-like soft layers - No high-definition hard sharpness, no AI illustration feel - No overexposure, no cheap studio feel ## 10 | Fixed Allocation of Four Photos ### Photo 1 | Main Visual Portrait - Bust or half-body close-up - Looking straight at the camera - Expression is natural and gentle, slight smile - Holding an ancient scroll with both hands or one hand - Hat, tassel, updo, real flower hairpins clearly visible - New Chinese details of collar, cuffs, front placket clearly visible - This is the main visual most suitable for a cover ### Photo 2 | Quiet Profile Mood Shot - Half body or 3/4 body - Head slightly turned to the side - Eyes looking down to the side or at the scroll - Hand gently unrolling the scroll or touching the scroll axis - Expression is quiet, gentle, introverted - Focus on highlighting the side profile, updo, hat decoration, real flower hairpins, collar layers, and cuff embroidery ### Photo 3 | Graduation Atmosphere Enhancement Shot - Half body or 3/4 body - The character holds a vertical inscription scroll with both hands, the scroll crosses diagonally in front of the chest from top left to bottom right, one hand holds the upper wooden axis, the other hand supports the lower wooden axis, the scroll surface is completely in the frame, the traditional calligraphy text "不讀了" is clearly visible. - Expression is more energetic, a genuine toothy smile - The image has a graduation season viral feel - Props are clear, text is clear ### Photo 4 | Playful Interactive Shot - Half-body close-up - Interacting with the scroll, calligraphy brush, round fan, or thread-bound book - Can slightly tilt head, slightly wink, slightly puff cheeks, playful smile - Keep it exquisite and beautiful, do not make it into a meme - Highlight updo, real flower hairpins, academic cap, tassel, and New Chinese placket details ## 11 | Uniform Requirements The 4 images must be uniform: - Same real girl - Same set of New Chinese [Degree Level] academic robe - Same updo styling system - Same real flower hairpin academic cap - Same major color logic - Same studio background - Same photography language - Same graduation portrait theme ## 12 | Key Avoidances No ordinary Hanfu. No ancient costume drama outfits. Do not change the academic robe beyond recognition. No text of any kind on the clothes. No fake flowers. No plastic floral decorations. No cheap photo studio feel. No AI fake faces. No over-smoothing skin.

Create a gritty flash-photography candid photo in a dim high school hallway, vertical 2:3 aspect ratio, with a teenage student in a dark Korean/Japanese-inspired school uniform posing full-body beside a long wall of pale blue-gray metal lockers covered with stickers, labels, tape marks, doodles, and scratches. The main subject is centered slightly right, leaning one hand against the lockers with one leg crossed in front, wearing {argument name="school uniform style" default="black blazer with gold buttons, white button-up shirt, loose black necktie with a crest, dark pleated mini skirt"}, black thigh-high socks, chunky black-and-white platform sneakers, and carrying a glossy black shoulder bag with silver chain straps. She has {argument name="hair color and style" default="long messy black hair with small white hair clips"}; her face is intentionally anonymized by a square gray blur. The corridor recedes deeply into the background with fluorescent ceiling lights, reflective gray floor tiles, scattered papers on the floor, and groups of other students in uniforms standing and talking on the left side, their faces also soft or blurred. On the right edge include a red fire extinguisher and a Korean emergency sign with a downward arrow. Use a raw 2000s digital-camera look: harsh direct flash, cool desaturated colors, strong shadows, slight grain, imperfect composition, realistic school clutter, no glamour retouching, documentary snapshot mood. Add a small subtle watermark-like text reading {argument name="watermark text" default="PromptHero"} in the bottom-left corner.

Goal: Create a polished fictional SaaS landing page hero for a data infrastructure and analytics platform named {argument name="brand name" default="NOVALITH"}, in a premium Japanese B2B technology style. Canvas: Wide desktop web landing page, 16:9 aspect ratio, clean white background with pale cyan gradients, subtle network-line geometry, faint dot matrices, and glossy floor reflections. Use a futuristic corporate palette of deep navy, teal, cyan, and white. Header: Top-left logo with a large angular gradient cyan-to-navy “N” icon followed by spaced uppercase wordmark “NOVALITH”. Across the top center place exactly five Japanese navigation items: 「プロダクト」, 「ソリューション」, 「ドキュメント」, 「リソース」, 「会社情報」. Top-right has 「ログイン」 and a rounded teal CTA button labeled {argument name="contact button text" default="お問い合わせ"} with a right arrow. Hero layout: Left side contains a large bold navy Japanese headline in two lines: {argument name="headline text" default="データを、未来の意思決定に変える。"}. Under it place a smaller Japanese subheadline: {argument name="subheadline text" default="NOVALITHは、開発・運用・データの複雑さを解消し、信頼できるデータ基盤で、事業の成長を加速させます。"} Feature row: Beneath the subheadline show exactly three compact feature blocks, each with a thin teal line icon, bold Japanese title, and two-line description. The three blocks are: 1) stacked-layers icon, title 「統合された基盤」, description 「データとシステムを一元的に統合」; 2) shield-check icon, title 「堅牢なガバナンス」, description 「セキュリティと品質を標準で確保」; 3) rising chart arrow icon, title 「高速な意思決定」, description 「信頼できるデータで意思決定を加速」. Hero buttons: Below the feature row place exactly two rounded buttons: a filled teal primary button labeled {argument name="primary cta text" default="プロダクトを見る"} with a right arrow, and a white outlined secondary button labeled 「資料をダウンロード」 with a download icon. Right side visual: Create a large tilted floating dashboard panel with rounded corners, soft shadow, and a glossy white UI. Behind it place a monumental 3D translucent teal/navy “N” structure and angular cyan glass panels, extending off the right edge. The dashboard should show a sidebar, top utility icons, analytics cards, line charts, a pipeline diagram, an alert table, and a donut chart. Dashboard details: Include exactly six sidebar menu items: 「ダッシュボード」 highlighted in teal, 「データカタログ」, 「パイプライン」, 「モニタリング」, 「ガバナンス」, 「アラート」. Include exactly four top statistic cards with labels and values: 「データパイプライン」 128, 「データ品質スコア」 98.6%, 「アクティブユーザー」 1,256, 「処理データ量」 2.45 PB. Include exactly three nodes in the pipeline diagram connected by lines, an alert table with exactly four rows, and a donut chart labeled 96% with a three-item legend. Visual style: Crisp vector-like UI mixed with photorealistic 3D glass, airy spacing, premium startup landing page aesthetics, high readability, soft shadows, subtle cyan glow, accurate Japanese typography, no people, no clutter, no watermark.

Goal: Create a polished mobile onboarding/top screen for a fictional B2B data platform named {argument name="brand name" default="NOVALITH"}, in a premium Japanese SaaS visual style. Canvas: Vertical smartphone screenshot, 9:16 aspect ratio, white background, iOS status bar at the top showing 9:41 with signal, Wi-Fi, and battery icons. Use generous margins and a clean app-advertising composition. Layout: Place the NOVALITH logo near the top left, with a geometric turquoise-blue “N” mark and wide-spaced navy wordmark. Below it, set a large bold Japanese headline in dark navy: {argument name="headline text" default="データを、未来の意思決定に変える。"}. Under the headline, add a smaller two-line Japanese description: {argument name="body text" default="開発・運用・データの複雑さを解消し、信頼できるデータ基盤で、事業の成長を加速させます。"}. On the lower left, show exactly 3 feature blurbs with thin cyan outline icons: 1) stacked layers icon, title “統合された基盤”, description “データとシステムを一元的に統合”; 2) shield check icon, title “堅牢なガバナンス”, description “セキュリティと品質を標準で確保”; 3) rising bar chart icon, title “高速な意思決定”, description “信頼できるデータで意思決定を加速”. Hero visual: On the right half, create a large glossy 3D smartphone card tilted slightly backward, overlapping a translucent teal-blue architectural block and a giant faceted 3D letter N behind it. The phone screen shows a Japanese dashboard UI with exactly 4 dashboard sections: 1) top metric cards, with “データパイプライン 128 +2.4%” and “データ品質スコア 98.6% +2.4%”; 2) pipeline status row with three connected icons labeled “データソース 24”, “処理中 128”, and “出力先 36”, plus small status dots; 3) line chart titled “データ量の推移” with dates from 5/06 to 5/12; 4) donut chart titled “データソース健全性” showing “96% 健全” and a legend with “健全 46”, “警告 3”, “エラー 1”. Keep the dashboard text small but believable and neatly aligned. Bottom navigation and actions: Center exactly 3 small carousel dots above the buttons; the first dot is teal and active, the other two are pale gray. Add exactly 2 rounded buttons at the bottom: a filled teal gradient primary button labeled {argument name="primary button label" default="はじめる"} with a right arrow icon, and a white secondary button with teal outline labeled {argument name="secondary button label" default="あとで始める"}. Visual style: Futuristic enterprise data analytics, crisp vector UI mixed with soft 3D glassmorphism. Palette is white, deep navy, cyan, teal, and aqua gradients. Add subtle polygon network lines, tiny nodes, faint dotted patterns, and light geometric overlays across the background. Use soft shadows, clean Japanese typography, high contrast, and a premium App Store marketing feel. Avoid clutter, avoid extra buttons, avoid extra feature items, and do not add watermarks.