GPT Image 2 prompts

Use the uploaded image as the facial identity reference. Preserve the subject’s recognizable facial features, skin tone, facial proportions, hairstyle, and natural expression accurately.

Create a highly realistic full-body formal fashion portrait of the subject standing confidently in an {argument name="setting" default="elegant modern interior"}.

Dress him in a sophisticated {argument name="suit color" default="light-grey"} three-piece plaid suit consisting of a tailored blazer, matching waistcoat, and slim-fit trousers. Pair it with a crisp white dress shirt, a solid black necktie, and a neatly folded white pocket square. Complete the outfit with clean minimalist white leather sneakers.

Pose him standing naturally with both hands inside his trouser pockets, shoulders relaxed, body facing forward, and a calm confident expression while looking toward the camera.

Give him a {argument name="hair style" default="neat short hairstyle"} with natural texture and a clean finish.

Set the scene inside a refined contemporary living room with:

soft grey paneled walls

a large tall indoor plant in a dark planter on the left

elegant neutral curtains near a bright window

a minimalist abstract framed artwork on the wall

a slim black console table with books and a small decorative plant

polished light-grey marble flooring with subtle reflections

Use soft natural window light combined with professional portrait lighting, realistic skin texture, balanced shadows, subtle depth of field, premium menswear editorial photography, sharp suit fabric details, natural proportions, photorealistic quality, full-body composition, vertical 3:4 aspect ratio, ultra-high resolution.

Ramen advertisement pop. Instructed the photo to have a {argument name="atmosphere" default="natural atmosphere, excluding the AI feel"}.

Photo: A realistic beauty content photo of a {argument name="age" default="27-year-old"} female influencer with a short haircut introducing cosmetics to the camera. Vertical, aspect ratio 9:16.

Subject:
Short dark brown hair with clean lines around the ears and nape. Natural makeup with a healthy glow. Skin texture preserved. Wearing a simple ivory top, looking at the camera with a friendly smile.

Action/Product:
Holding a small frosted glass serum bottle next to her cheek with one hand, while supporting it with the other palm. The bottle contains a pale pink liquid and has a white cap. The label faces the camera, with the text clearly visible. The face and product do not overlap; both are sharply focused.

Product Label:
A professionally finished package for a fictional skincare brand named "{argument name="brand name" default="NUVÉA"}." A matte ivory rectangular label is applied straight to the front of the bottle, following the curvature naturally with a subtle paper texture.

The label must accurately print the following text exactly once:
Top: "{argument name="brand name" default="NUVÉA"}"
Center: "{argument name="product name" default="HYDRATING SERUM"}"
Below that: "Niacinamide 5%"
Bottom: "30 mL / 1.0 fl oz"

The brand name uses a slightly larger, sophisticated serif font. The product name and other info use a smaller, legible sans-serif font. Use dark charcoal text with proper spacing and margins. Add a thin dusty pink line. The text must look like actual printing that conforms to the perspective and curve of the bottle. No extra text or certification marks.

Composition/Background:
A portrait shot at eye level from the top of the head to the chest. The entire hair, hands, and product are within the frame. Both the face and product are centered with the same depth of field. The product is large enough for the label to be readable. The background is a bright beige room with a soft-blurred vanity and small houseplant.

Lighting/Finish:
Soft natural light from a window strikes diagonally from the front, creating gentle shadows on the face. Subtle reflections on the glass and natural textures on skin and fabric. Looks like a clean, natural moment captured from a beauty review video on social media.

Constraints:
One person. Fingers and bottle shape must be natural and accurate. Avoid excessive skin filtering, plastic-like skin, or harsh lighting. Text should only appear on the product label. No subtitles, watermarks, or social media UI.

Create a photorealistic retro 1980s synthwave podcast portrait of a smiling {argument name="host description" default="white male host in his 30s with blue eyes, short messy brown mullet hair, light stubble, and a friendly grin"} seated close to camera at a desk, facing forward with relaxed crossed arms. He wears a bold neon windbreaker in {argument name="jacket colors" default="teal, hot pink, purple, and black geometric panels"}, a white T-shirt, and a black digital wristwatch. Place a large vintage silver broadcast microphone on a black boom arm centered in the foreground, with small SHURE branding visible. The background is a dim studio filled with 1980s nostalgia: a bright neon triangle sign behind him reading {argument name="neon sign text" default="Thursday AI"} in pink cursive and cyan tube lights, a stack of exactly 4 cassette tapes on the left labeled “SYNTHWAVE”, “HITS 1984”, “MIXTAPE”, and “GOOD VIBES”, a retro boombox beneath them, and on the right an old CRT television showing a vaporwave sunset with palm trees and a blue grid horizon. Add one framed lightning-bolt poster above the TV. Use saturated magenta and cyan rim lighting, warm skin tones, shallow depth of field, soft film grain, slight dust and scratches, worn rounded photo border, high-detail studio photography, square-ish vertical crop, nostalgic vaporwave color palette, no extra people, no modern devices, no watermark.

{argument name="aspect ratio" default="16:9 widescreen"}, realistic period drama film still. A fictional {argument name="scene" default="Tang Dynasty royal banquet"}, luxurious and warm, where subtle tension is hidden amidst the liveliness.

Create a vertical 9:16 cinematic 3D animated-style scene in a warm sunlit kitchen about dirty produce and invisible germs. In the foreground on a rustic wooden countertop, show exactly 5 anthropomorphic vegetables/fruits with worried cartoon faces: 1 large red apple in the center with big eyes, raised brown eyebrows, open anxious mouth, water droplets, dirt specks, a stem and one green leaf; 1 red tomato at front left with big eyes and shocked open mouth; 1 pale green lettuce/cabbage head at back left with large eyes, frightened mouth, dirt splashes, and leafy folds; 1 tall orange carrot at back right with leafy green top, dirt on its skin, big eyes and sad frown; 1 small broccoli head partly hidden at far right with worried eyes. Add exactly 7 visible small pests on the produce: 1 cute brown snail with eye stalks sitting on the lettuce, 1 green caterpillar crawling on the apple, 3 black flies with translucent wings perched on the apple/tomato/leaf area, and 2 tiny worm-like bugs near the tomato edge. Scatter clumps of dark soil and crumbs across the table and on the produce. At bottom right foreground, place a large handheld magnifying glass angled over the apple, revealing enlarged colorful cartoon microbes inside the lens, including round spiky germs and rod-shaped bacteria in green, yellow, orange, and red, sharply detailed compared with the rest of the scene. Background: cozy kitchen with shallow depth of field, potted plants on a sunny windowsill, sink and jars softly blurred. Include two framed signs: on the left wall a cream poster with hand-lettered Spanish text {argument name="left sign text" default="FRUTAS Y VERDURAS MÁS LIMPIAS = UNA FAMILIA MÁS SANA"} and a small red heart below; on the right a black chalkboard with white chalk Spanish text {argument name="right sign text" default="LO QUE NO VES TAMBIÉN IMPORTA"}. Visual style: high-quality Pixar-like clay/stop-motion 3D, expressive oversized eyes, tactile vegetable textures, realistic dirt and water droplets, macro photography feeling, warm golden morning light, soft bokeh, crisp foreground focus, wholesome public-health advertisement mood. Use {argument name="main subject" default="dirty fruits and vegetables"} as the scene theme, with {argument name="lighting style" default="warm golden kitchen sunlight"} and {argument name="mood" default="concerned but charming and educational"}. Avoid logos, watermarks, extra text, extra produce characters, or additional pests beyond the counted visible items.

Create a hyper-realistic museum-style product photograph of a single ancient {argument name="artifact type" default="terracotta ceramic vase"} centered on a pure black background. The vessel is tall and ovoid with a narrow foot, rounded shoulders, and a wide flared rim, made of reddish-brown weathered clay with heavy age patina, mineral deposits, chips, scratches, white abrasions, and fine cracks. Add carved ornamental decoration: one continuous geometric band around the neck just below the rim and raised swirling vine-like relief patterns on the body. The front right side has one large jagged vertical rupture exposing the hollow interior, with multiple broken clay shards still attached around the opening; the opening runs from the upper shoulder down toward the lower body and shows a warm copper-orange hammered, dimpled inner surface catching light. Use dramatic studio lighting from the upper left and subtle warm internal/reflected glow inside the broken cavity, emphasizing rough ceramic texture, fractured edges, depth, and archaeological realism. Composition is a straight-on three-quarter product view, full vase visible with slight shadow at the base, high contrast, sharp focus, ultra-detailed, no labels, no people, no extra objects, no watermark.

Goal: Create a wide side-by-side comparison image showing two underwater snorkeling photos of the same scene, as if comparing model outputs.

Canvas: Horizontal 2:1 aspect ratio on an off-white background, with two large rectangular photo panels placed side by side and separated by a narrow vertical gutter. Add small black bold labels above the panels: left label reads {argument name="left label" default="GPT Image 2"}, right label reads {argument name="right label" default="GPT Image 2.5"}.

Layout: Use exactly 2 photo panels. The left panel occupies the left half and the right panel occupies the right half. Keep the labels aligned near the upper-left corner of each panel. No borders other than the off-white background margin and the central gap.

Subject details: Both panels show a realistic underwater wide-angle action-camera photograph in clear tropical turquoise water. The subject is {argument name="subject" default="a young woman snorkeler"} wearing a black wetsuit or rash guard, a large black snorkel mask, and black fins. She is swimming horizontally toward the camera just below the rippling water surface. Sunlight creates bright caustic patterns across the shallow sandy seabed. Small fish are visible in the blue water, and coral or rocky reef formations appear near the edges.

Left panel: The snorkeler is centered and closer to the camera, facing forward with arms tucked back, body streamlined, black mask dominating the face. The water is bright cyan, the sandy bottom is shallow and patterned with sun ripples, and a small reef rock sits in the lower-left foreground. The overall look is clean, slightly artificial, and symmetrical.

Right panel: The snorkeler is angled diagonally from lower center toward upper left, with one arm extended toward the camera in a selfie-like perspective. Her fins are visible behind her near the top-left area, and larger reef rocks appear on the right side. The water has richer color contrast, stronger surface shimmer, more fish in the background, and a more dynamic wide-angle composition.

Visual style: Photorealistic underwater photography, natural sunlight, crisp details, GoPro-style wide-angle lens, high clarity, realistic water refraction, vivid aquamarine and sandy white tones.

Constraints: Include exactly 2 labeled panels and exactly 2 visible text labels. Do not add extra captions, watermarks, UI elements, arrows, or decorative graphics.

Create a wide horizontal ukiyo-e inspired mythic landscape illustration in aged woodblock print style, with visible paper grain, muted ink texture, fine black linework, and a dusky palette of indigo mountains, misty gray-blue valleys, faded salmon pink sky, and soft cream highlights. In the center foreground place one young Greek poet or traveler, {argument name="character name" default="Orpheus"}, with curly dark hair, a laurel wreath, a worried backward glance, a red draped cloak over a pale tunic, and a small lyre held at his side. On the left, emerging from shadowy cliffs and mist, show one translucent ghostly woman in flowing white robes with long dark hair, reaching toward him with one hand, ethereal and partially faded into the fog. The scene contains exactly two flying cranes in the upper sky, one large pale sun near the upper right, one red torii gate on a distant mountaintop, and one winding pale stone path climbing from the lower right foreground through misty mountains toward the gate. Frame the upper left and lower left with dark cherry tree branches covered in pink blossoms, with many loose pink petals drifting diagonally across the whole composition. Add exactly two narrow vertical Japanese paper title plaques: one at the upper left reading {argument name="left plaque text" default="冥より暁へ"}, and one at the right middle reading {argument name="right plaque text" default="音の道"} with a small red seal below the characters. Blend Japanese ukiyo-e composition with a subtle ancient Greek mythological subject, cinematic but flat-print, melancholic and dreamlike, no modern objects, no photorealism, no extra people, no extra birds, no watermark.

Create an ultra-realistic vertical smartphone selfie portrait of a young woman in a cozy bedroom, photographed at close range with natural window light. The subject is {argument name="person" default="a young woman in her early 20s"} with {argument name="hair color" default="dark brown hair"} styled in a messy high bun with loose face-framing strands and flyaway hairs, thick natural eyebrows, brown eyes, long lashes, clear dewy skin with subtle pores and a few small blemishes, and soft neutral pink lips. She looks directly into the camera with a calm, slightly intimate expression while her right hand is raised over her mouth as if blowing a kiss; visible natural nails and a thin gold ring on the ring finger. She wears {argument name="top" default="a ribbed dark brown scoop-neck henley tank top with small buttons and scalloped trim"}. Compose the image as a tight head-and-shoulders portrait, face filling most of the frame, slightly tilted angle, cropped from upper chest to above the bun, with shallow depth of field. Background: softly blurred bedroom interior with warm beige walls, a bedside lamp glowing on the left, a small green plant, a framed wall print reading {argument name="wall print text" default="Good Things Take Time"}, a bed edge, and a white door with a dark handle on the right. Lighting should be soft and flattering, realistic skin texture, no heavy makeup, no airbrushed plastic look. Use natural colors, high detail, candid Pinterest-style realism, as if captured on a modern phone camera, vertical 9:16 aspect ratio.

Create a photorealistic square social-media product shot of a young female streamer in a cozy gaming setup, shot with a slightly wide-angle close perspective from above desk height. She is seated in a gaming chair, leaning toward the camera with an excited open-mouth smile, holding one graded trading card slab prominently in both hands close to the lens. Her look: {argument name="character appearance" default="young woman with wavy black hair with purple streaks, freckles, nose stud, hoop earrings, choker, black cropped top and loose black off-shoulder hoodie"}. The graded slab should have a clear plastic case, a top label reading “NEXUS GRADE,” a large “10” grade box, a small barcode and serial number, and the card art should show a dramatic red-orange robot/monster battle card titled {argument name="featured card name" default="GIGATRON"}. Place exactly 5 visible trading-card/product elements: 1 graded slab in her hands, 1 unopened booster pack on the desk, 1 opened silver foil wrapper on the desk, and 2 loose cards on the desk, one dark pirate-themed card and one colorful cute creature card. The booster pack should read {argument name="booster pack text" default="DOGINAL DOGS LEGENDS RISE OF THE PACK"}. The background is a streamer room with exactly 2 large monitors showing fantasy landscapes, 1 black boom-arm microphone with foam windscreen on the right, 1 glowing ring light partially visible in the upper right, 1 RGB keyboard, 1 gaming chair, and 1 iced pink drink with straw on the desk. Use moody purple and blue RGB lighting mixed with warm screen glow, shallow depth of field but keep the woman’s face and the graded slab sharp, realistic skin texture, authentic cluttered desk, cinematic marketing-still composition. The overall mood should feel like influencer livestream content for a collectible card game product launch. Use {argument name="lighting style" default="neon purple-blue streamer lighting with warm monitor glow"} and a {argument name="camera style" default="high-detail photorealistic wide-angle smartphone-style shot"}. Avoid extra fingers, unreadable main labels, warped slab edges, duplicate microphones, or additional cards beyond the specified 5 product elements.

Goal: Create a wide comparison graphic showing two generated versions of a rhode beauty static ad for Peptide Lip Treatment in Watermelon Slice, with a black presentation background and two square ad mockups side by side.

Canvas: 16:9 horizontal frame, black background, high contrast, clean social-media comparison layout.

Layout: Place two rounded-corner square ad images centered horizontally with generous spacing. The left panel is labeled above in bright orange text: “GPT-Image-2”. The right panel is labeled above in the same bright orange text: “GPT-Image-2.5”. Add one small italic orange caption centered near the bottom reading “prompt in comments”.

Ad panels: Use exactly 2 square beauty advertisement panels.
1. Left panel: soft pink background, rounded corners, elegant editorial beauty-ad composition. At the upper left, large serif headline text reads “A SLICE OF GLOSS.” with smaller text below reading “Peptide Lip Treatment”. A pink rhode Peptide Lip Treatment tube floats diagonally from upper right toward lower center, with vertical white “rhode” branding. Surround it with a glossy pink gel swirl, transparent splash droplets, and watermelon slices. Include a large circular pink sunburst shape behind the product. Watermelon elements count: 3 visible slices or wedges — one large lower watermelon slice, one right-middle slice behind the tube, and one small upper-right wedge. Keep the product large and central.
2. Right panel: more dynamic, brighter, more detailed version of the same rhode ad. Pink background, rounded corners, upper-left serif headline “A SLICE OF GLOSS.” and smaller “Peptide Lip Treatment”. A larger pink rhode tube floats diagonally near the center, surrounded by a dramatic translucent glossy pink splash ribbon, many bubbles and droplets, and watermelon slices. Watermelon elements count: 5 visible pieces — one large foreground watermelon strip at the bottom, one left-middle wedge, one right-middle wedge, one small upper-right wedge, and one partial slice behind the splash near the center. The product should appear sharper, glossier, and more premium than the left panel.

Visual style: polished product advertising, vivid watermelon pinks and reds, glossy liquid textures, studio lighting, high-end cosmetics campaign, crisp typography, saturated color, realistic 3D product rendering mixed with surreal splash photography.

Text and customization: Use the top labels {argument name="left model label" default="GPT-Image-2"} and {argument name="right model label" default="GPT-Image-2.5"}. Use the ad headline {argument name="headline text" default="A SLICE OF GLOSS."}. Use the brand on the tube {argument name="brand name" default="rhode"}. Use the bottom caption {argument name="bottom caption" default="prompt in comments"}.

Constraints: Keep exactly 2 comparison panels, exactly 2 orange model labels, and exactly 1 bottom caption. Do not add extra logos, watermarks, UI chrome, or additional text outside the specified labels and ad copy.

Create an ultra-realistic square travel portrait of a young woman on a high-rise New York City observation deck, photographed through floor-to-ceiling glass with the Manhattan skyline behind her. She is seated or leaning casually in the foreground, shoulders bare in an off-the-shoulder cream ribbed knit sweater, wearing a burgundy baseball cap embroidered with {argument name="cap text" default="BROOKLYN"}, small gold hoop earrings, and a thin gold necklace. She has {argument name="hair style" default="long dark curly hair"}, warm medium-brown skin with natural freckles, glossy lips, and closed eyes, smiling softly as she tilts her face up toward warm sunlight. Behind her, show the Empire State Building centered slightly left, distant skyscrapers and hazy water on the horizon under a clear blue sky. Add subtle white lettering on the glass reading {argument name="observation deck text" default="OBSERVATION DECK"} and {argument name="level text" default="LEVEL 86 - 1,050 FT"}. Use natural golden-hour sunlight from the right, realistic reflections and vertical glass seams, shallow depth of field with the skyline slightly softened, crisp skin texture, cinematic color grading, authentic smartphone/editorial travel photography, no watermark, no extra text.

Goal: Create a horizontal social media comparison graphic showing the same photorealistic action portrait rendered by three AI image models, emphasizing improved realism, water detail, lighting, and skin texture.

Canvas: 16:9 landscape layout, approximately 1200×675 px, with a dark navy-to-electric-blue gradient background that is nearly black on the far left and vivid cobalt blue on the far right.

Layout: Place exactly 3 vertical portrait image panels in a row with even spacing and generous margins. Each panel is a rectangular photo with no border, aligned along the same baseline. Above each panel, add a bold white uppercase label centered over the image.

Text content: Use exactly 3 labels: left label “{argument name="left model label" default="GPT IMAGE 2"}”, center label “{argument name="center model label" default="GPT IMAGE 2.5 FLARE"}” split across two lines, and right label “{argument name="right model label" default="GPT IMAGE 2.5 SUNBURST"}” split across two lines. Use a clean bold sans-serif font, all caps, white, with strong legibility.

Subject in all 3 panels: A close-up outdoor sports portrait of the same athletic adult man with tan wet skin, messy dark wet hair, neon lime wraparound sport sunglasses with orange mirrored lenses, and a pale mint-white sleeveless athletic tank top. He is being splashed with water against a clear bright blue sky. Show droplets, streams, wet fabric, wet hair, highlights on skin, and reflective eyewear. The expression is open-mouthed and breathless, like mid-workout or after splashing water on his face.

Panel details: Left panel: frontal portrait, more static and less refined, water streaming straight down from his mouth and chin, sunglasses large and centered, shoulders visible. Center panel: more energetic and dynamic, head turned slightly to viewer’s right, dense suspended water droplets around face, stronger contrast, sharper texture, more dramatic wet hair and skin. Right panel: cleaner and more controlled realism, head turned slightly upward and to viewer’s left, smooth natural lighting, water splashing diagonally across the chest and face, refined reflections and skin detail.

Visual style: Photorealistic editorial comparison, high-resolution sports photography, bright daylight, crisp water interaction, realistic wet skin texture, high shutter speed droplets, saturated blue sky, vivid neon eyewear.

Constraints: Include exactly 3 labeled image panels and no additional logos, captions, icons, arrows, or watermarks. Keep the background outside the panels as a simple gradient. The three panels should clearly depict the same core prompt but with subtly different realism and composition.

Goal: Create a 3:4 vertical cinematic visual poster of a futuristic design studio on the top floor of a Tokyo high-rise during blue hour after heavy rain, featuring a plush whale character designing a mechanical butterfly.

Canvas: Vertical 3:4 composition, ultra-detailed cinematic realism, shallow depth of field, glossy reflections, warm desk lights contrasted with cool rainy blue city light. The camera looks slightly downward across a cluttered designer’s desk toward floor-to-ceiling windows.

Main subject: A large cute plush {argument name="character name" default="blue whale"} mascot sits in a black office chair at the desk, wearing a dark gray hoodie with white drawstrings. It has soft fuzzy blue fur, a white belly and muzzle, small black glossy eyes, side fins, and a small rounded fin on top. The character holds a stylus and draws on a large pen-display tablet.

Scene and environment: Behind the character, huge rain-covered windows show a vivid Tokyo night skyline with wet glass, bokeh neon reflections, skyscrapers, and Tokyo Tower glowing orange in the distance. Realistic raindrops streak down the glass. Add indoor plants, shelves, desk lamps, and metallic studio equipment. On the left wall, place a minimalist poster with a black circular logo and Chinese text reading 「更快・更锐利・更智能」.

Desk contents: Include exactly 9 distinct desk items: 1 large pen-display tablet showing butterfly concept sketches, 1 black keyboard, 1 transparent glass mug of water, 3 Polaroid photos of the whale character, 1 open notebook, 1 black pen lying in front of the notebook, and 1 black cylindrical pen cup. The three Polaroids should show: the whale wearing sunglasses, the whale wearing a red cap, and the whale drinking from a cup.

Butterfly design: Show exactly 2 prominent mechanical butterflies: one larger glowing orange-and-glass butterfly floating in front of the tablet, and one smaller transparent glass-wing butterfly shown in a floating UI card. The larger butterfly should have a metallic orange body, delicate transparent wings with mechanical vein structures, warm LED highlights, and sparkling reflections. The tablet screen should also show multiple drawn stages of the butterfly design, including pencil line sketches and colored mechanical renderings.

Interface overlay: Add exactly 2 floating translucent UI bubbles on the right side. The top bubble contains a small round icon of the whale and Chinese text: 「只把蝴蝶翅膀改成透明玻璃，其他部分保持不变。」 The lower bubble contains the smaller glass-wing butterfly preview. Use rounded rectangles, subtle blue-gray glassmorphism, and soft glow.

Notebook text: On the open notebook in the lower right foreground, write large handwritten Chinese text: 「想象，不止于提示词。」 and below it write “GPT-Image 2.5”.

Visual style: Hyperrealistic 3D cinematic advertising poster, immaculate material rendering, wet glass, polished black tabletop, metal edges, water droplets, soft plush fur, transparent glass wings, orange neon glints, blue ambient city light, volumetric lamp glow, high contrast but clean, almost no noise.

Constraints: Preserve the exact counted items and text placements, keep all text legible, do not add extra characters, avoid cartoon flatness, avoid watermark or logo overlays beyond the described wall poster and notebook text.

Goal: Create a cinematic photorealistic architectural cutaway scene of a glass apartment building viewed through a handheld magnifying glass, with annotated floor-by-floor labels.

Canvas: Landscape 4:3 image, ultra-detailed evening cityscape at sunset after rain. A waterfront modern glass building stands beside a reflective river or lake, with distant skyline, warm orange sun near the horizon, blue storm clouds, and a faint rainbow on the left.

Main composition: In the extreme foreground, a human hand holds a large round magnifying glass at the right edge. The lens dominates the center of the frame and contains sharp focus, visible rim distortion, and many raindrops on the glass. Inside the magnified circle, the front building is crisp and enlarged, revealing a transparent nine-floor cross-section with people, furniture, and warm interior lights. Outside the lens, the same building and city are slightly softer and less magnified, emphasizing the optical effect. Water reflections, wet rocks, plants, and a few small fish near the clear shallows appear along the bottom.

Building details: A tall modern steel-and-glass rectangular building on piloti columns at the waterline, with warm yellow interior lighting and transparent walls. Show exactly 9 stacked visible levels, each as a distinct room or activity scene with tiny realistic people.

Floor labels and annotations: Place thin white leader lines from the rooms to clean white text on the right side inside or beside the magnifying glass. Include exactly these 9 labeled floors, top to bottom:
1. “9F · Rooftop Garden” with subtext “Tea under string lights · Evening sky”; show people seated among plants and string lights on an open roof terrace.
2. “8F · Lounge” with subtext “Friends talking · Warm ambient lights”; show a cozy social lounge with sofas and people chatting.
3. “7F · Workshop Lab” with subtext “Careful inspection · Tools neatly arranged”; show a maker workshop with shelves, tools, and a person working at a bench.
4. “6F · Child's Room” with subtext “Story time · Toys on the floor”; show a warm kids’ room with an adult and child, toys scattered.
5. “5F · Mini Gym” with subtext “Treadmill run · Evening workout”; show exercise machines, one person running, another working out.
6. “4F · Reading Room” with subtext “Quiet notes · Bookshelves and desk lamp”; show bookshelves, desk, lamp, and a seated reader.
7. “3F · Design Office” with subtext “Team discussion · UI review in progress”; show a small office meeting with screens and a blue presentation board.
8. “2F · Artist Studio” with subtext “Painter at work · Large canvas by the window”; show an artist painting on a large canvas, art supplies and easel.
9. “1F · Family Kitchen” with subtext “Cooking dinner · Steam and warm light”; show family members at a dining table/kitchen with glowing lamps and steam.

Visual style: Photorealistic, highly cinematic, rich depth of field, warm interior glow contrasted with cool blue evening outdoors, realistic rain droplets, reflections, refraction through magnifying lens, sharp micro-details, realistic people at miniature scale, premium architectural visualization quality.

Text constraints: The annotation text must be legible, white, modern sans-serif, aligned vertically on the right, with thin white pointer lines. Keep all labels in English. Do not add extra floors, extra labels, logos, watermarks, or unrelated text.

Using the provided reference sketch as the composition guide, turn it into a highly detailed cinematic dark-fantasy illustration of {argument name="scene subject" default="a warrior confronting a fire-breathing dragon"}. Keep the same basic layout: one warrior in the lower foreground facing one huge winged dragon above. Replace the simple line drawing with realistic epic concept art: the warrior wears battle-worn armor, a torn red cape, and holds a raised sword; the dragon is massive, horned, spiked, with wide leathery wings, clawed feet, glowing eyes, and a torrent of fire blasting from its mouth toward the battlefield. Add a dramatic environment not present in the sketch: burning ruined castle/city, collapsing towers, banners, smoke-filled storm clouds, flying embers, lava-like orange firelight, and jagged black rocks in the foreground. Style should be {argument name="visual style" default="ultra-detailed cinematic fantasy game key art"}, with intense contrast, dynamic perspective from behind the warrior, warm orange flames against dark gray smoke, no UI elements, no sketch lines, no text, no watermark.

Goal: Create a vertical two-panel design that transforms a cozy realistic cat photo into a minimalist geometric poster illustration.

Canvas: Portrait 3:4 composition, 900 x 1200 style, split horizontally into exactly 2 panels. The top panel occupies about 52% of the height and is a warm realistic photograph; the bottom panel occupies about 48% and is an off-white gallery-poster illustration area.

Top panel: Show one fluffy white cat sitting at a wooden desk in a warm home office. The cat faces forward behind one open silver laptop positioned on the left foreground, with the keyboard barely visible at the bottom edge. The cat has exactly 2 square white paper notes taped over its eyes, each note drawn with one simple black cartoon eye and short eyelashes, giving a funny wide-eyed look. Background is softly blurred with exactly 1 black office chair centered behind the cat, 1 leafy plant on the left, and 1 wooden bookshelf on the right with indistinct books and decor. Lighting is golden, soft, shallow depth of field, cozy indoor ambience, realistic fur detail.

Bottom panel: On a clean off-white background, place a small centered vector poster illustration inspired by the top photo. The illustration contains exactly 1 white cartoon cat, exactly 1 black laptop silhouette on the left, exactly 2 square cartoon-eye notes over the cat's eyes, and the cat's paws on the keyboard. Use crisp orange outlines and flat graphic shapes. Add the large vertical headline text {argument name="headline text" default="LOOK"} in tall condensed orange letters at the upper left of the illustration. Add a small stacked caption on the right reading {argument name="side caption" default="GOOD THINGS AHEAD"}, with a thin black vertical line beside it and a short underline beneath it.

Geometric elements in the bottom poster: Include exactly 8 abstract color-block accents: 1 large orange rectangle behind the cat, 1 yellow circle behind the top right of the cat, 1 cyan-blue square overlapping the yellow circle, 1 tiny magenta square above, 1 small cyan square at the lower left, 1 small yellow horizontal rectangle beside it, 1 long yellow horizontal bar along the bottom, and 1 tiny blue rectangle at the lower right end of that bar. Add a minimal black plus-sign registration mark near the headline.

Visual style: Top panel should be realistic warm photography; bottom panel should be clean Bauhaus-inspired geometric poster art with orange, yellow, cyan, magenta, black, and white. Keep generous negative space in the bottom panel, sharp vector edges, playful editorial design, no extra animals, no extra text, no watermark.

Goal: Create a vertical museum-style fashion poster inspired by Dunhuang flying apsaras, combining a realistic mythic dancer with oversized Chinese typography.

Canvas: Tall vertical 9:16 poster, warm ivory paper background, clean negative space, no border or watermark.

Layout: Place four huge dark teal Chinese characters in bold blocky poster type: top row has exactly 2 characters, {argument name="top headline text" default="敦煌"}, spanning nearly the full width; bottom row has exactly 2 characters, {argument name="bottom headline text" default="飞天"}, oversized and partially behind/in front of the figure for a layered editorial look. Add one small slogan block at the lower left with exactly 2 short lines of dark teal Chinese text: {argument name="slogan text" default="穿越千年的\n飞天之美。"}.

Subject details: A single elegant {argument name="main subject" default="Dunhuang flying apsara dancer"} floats diagonally from lower left toward upper right, torso and face visible on the right side, one arm extended horizontally to the left. She has porcelain skin, serene upward gaze, red lips, delicate facial features, and ornate black hair in high double buns with gold floral hair ornaments. She wears antique gold earrings, bracelets, armlets, and a decorative necklace, plus a cropped teal-and-rust silk top and layered translucent flowing skirts.

Fabric and motion: Use exactly 3 major flowing silk ribbon groups: one teal ribbon sweeping behind the extended arm, one rust-red ribbon crossing from left to center, and one long rust-red ribbon curling down toward the bottom edge. The translucent skirt spreads across the lower left with painterly beige, teal, and terracotta textures, suggesting wind and flight.

Visual style: High-end contemporary Chinese cultural poster, cinematic realistic 3D fashion photography mixed with Dunhuang mural color palette, soft natural light, refined shadows, matte paper texture, teal typography, terracotta and jade fabrics, elegant composition, dramatic scale contrast between figure and type.

Constraints: Keep the typography legible and dominant, preserve exactly four large headline characters and one two-line small slogan, avoid extra text, avoid modern accessories, avoid cropped face, avoid cluttered background.

Using the provided reference image, clean up and regenerate the same composition as a higher-quality version. Remove the gritty/noisy texture, compression artifacts, speckling, and rough grain while preserving the original layout, Japanese text, manga illustration style, colors, lighting, characters, objects, borders, and all readable typography exactly as in the reference. Keep the two-part page structure unchanged: the top Japanese text panel and the bottom tavern illustration. Do not redesign, crop, add elements, translate text, or change the scene; only make it smoother, sharper, cleaner, and more polished like a restored high-resolution print. Apply {argument name="cleanup strength" default="strong grain and artifact removal with preserved ink detail"}.

Create a cinematic photorealistic historical portrait of {argument name="character name" default="Elon Musk"} reimagined as a wealthy ancient Mesopotamian or Egyptian-era noble merchant standing in a sunlit riverfront city. Show him from the waist up in the foreground, centered slightly left, looking confidently off to the upper right with a calm serious expression. He wears an off-white rough linen tunic with frayed sleeves, a dark brown patterned draped shawl over one shoulder, layered bead-and-gold necklaces, and multiple engraved gold bracelets with blue stones on both wrists. His hands hold one clay tablet covered in cuneiform-like writing close to his chest. The environment is an ancient desert port city at golden hour: a carved stone pillar with hieroglyphic-style symbols on the far left, palm trees and mud-brick buildings in the background, a broad river or canal to the right with small wooden sailboats, workers carrying and stacking stone blocks, and a stepped ziggurat or pyramid-like construction site in the hazy distance with scaffolding. Use warm amber sunlight, dusty atmosphere, soft background blur, subtle rim light around the figure, realistic skin texture, shallow depth of field, epic film still composition, 35mm anamorphic look, high detail, natural colors, no modern objects, no text overlay, no watermark.

Goal: Create a realistic vertical smartphone screenshot of a ChatGPT conversation showing an AI-generated Chinese history infographic about {argument name="topic" default="安史之乱"}.

Canvas: 576 × 1024 portrait phone screen, clean white iOS-style interface, high-resolution screenshot look with soft shadows and rounded UI elements.

Top status and navigation: Show the iPhone status bar with exactly 5 visible status items: time “8:47”, a small user/person icon, black Dynamic Island with one tiny orange dot, cellular/Wi‑Fi indicators, and a battery badge reading “16”. Below it, show exactly 3 floating navigation controls: a left circular hamburger/menu button with a small blue notification dot, a right rounded pill containing an edit/compose icon, and a three-dot menu icon.

Chat layout: On the right side near the top, place one light-blue user message bubble containing exactly this Chinese text: “生成图片：全面讲一讲\n安史之乱”. Beneath it, show one generated image card with rounded corners. Add exactly 2 small action icons under the card on the left: a share/upload icon and an ellipsis icon. Add one white circular share button overlay at the lower-right corner of the generated image card.

Generated image card content: The card is a dense, horizontal Chinese educational infographic poster about the An Lushan Rebellion. Use an antique parchment and ink-wash historical style with dark red section headers, beige paper background, weathered borders, and a dramatic top banner painting of a burning Tang dynasty city, smoke, soldiers, banners, and a dark horse silhouette. Large brush-calligraphy title at upper left: “安史之乱”. Include a smaller subtitle about the major turning point of the Tang dynasty from prosperity to decline. Keep all infographic text in Chinese and make it appear mostly legible.

Infographic sections: Use exactly 9 visible content sections in the poster: 1) “什么是安史之乱?” explanatory text block, 2) “主要人物” row with exactly 5 illustrated portrait figures labeled An Lushan, Shi Siming, Tang Xuanzong, Tang Suzong, and Yang Guifei, 3) “叛乱的原因” list with exactly 4 numbered causes, 4) “叛乱的过程” vertical timeline with exactly 6 dated milestones, 5) a central “叛乱形势图” map of Tang China with red and black arrows showing rebel and imperial movements, 6) “重要战役与事件” column with exactly 4 thumbnail battle/event images, 7) “叛乱的影响” block with exactly 4 icon-and-text impact items, 8) “历史意义” block with exactly 3 icon-and-text meaning items, and 9) “后世评价” block with exactly 3 quote cards. Add a bottom slogan line in Chinese resembling “以史为鉴，方知兴替。” with a small grayscale battlefield silhouette along the bottom edge.

Bottom input bar: At the bottom of the phone screenshot, show the ChatGPT message composer as one rounded white pill with exactly 4 elements: a plus icon, gray placeholder text “询问 ChatGPT”, a microphone icon, and a blue circular voice-wave button.

Visual style: Modern ChatGPT mobile UI combined with a richly detailed Chinese history infographic. The screenshot should feel like an authentic mobile app capture, with white margins, subtle drop shadows, rounded corners, and no external watermark.

Constraints: Do not add extra chat messages. Keep the infographic fully contained inside the rounded image card. Preserve the visible Chinese text where specified. Use accurate Tang dynasty historical visual motifs and avoid fantasy elements.

Create a realistic vertical smartphone photo of a candid grocery-store shopping moment in a Korean supermarket. A young woman with {argument name="hair color" default="dark brown shoulder-length hair"} is crouching low beside the cereal and granola aisle, wearing a {argument name="hat color" default="red"} baseball cap pulled low over her face, a slightly loose white ribbed long-sleeve henley top, light blue wide-leg jeans, white sneakers with beige soles, and a slim black shoulder bag. She is turned in side profile looking to the right, with her left arm extended toward a black plastic shopping basket on the floor; the basket contains visible groceries and has the white “emart” logo on the side. The scene contains exactly one main shopper, one black basket, two background shoppers, one shopping cart in the background, one prominent overhead aisle sign with the number 6 and Korean category text, one visible price sign reading “2,980,” and long shelves packed with cereal, granola, snack, and packaged food bags with Korean labels. Use a natural consumer-phone snapshot style, slightly imperfect framing, realistic indoor fluorescent lighting, mild motion softness, neutral colors, glossy gray supermarket floor reflections, deep aisle perspective, and authentic everyday details. Make it look like a casual buyer-show or store social media photo rather than a studio fashion shoot. No glamour lighting, no exaggerated posing, no artificial-looking skin, no extra people, no watermark.

Create a clean, photorealistic product-design workspace scene showing an AI-assisted transformation from sketch to finished object. Use a wide 16:9 canvas with soft natural daylight, pastel mint-green wall, realistic shadows from window light, and a modern maker-studio aesthetic. On the left, place a sheet of rough off-white textured drawing paper standing upright on a blue gridded cutting mat, held at the top with one small strip of orange tape; the paper shows a detailed graphite construction sketch of a {argument name="product type" default="mushroom-shaped ceramic table lamp"}, including vertical and horizontal guide lines, a domed shade, a short neck, rounded bulbous base, hatch shading, and a ground shadow. In front of the sketch are exactly 2 drawing tools: one black pencil angled diagonally and one white rectangular eraser. Across the center, add a translucent futuristic UI visualization: exactly 4 semi-transparent glass panels receding from left to right, each showing the lamp evolving from rough sketch to cleaner render to simplified 3D form, with pale cyan outlines and a subtle blue glow. Add exactly 3 interface overlays: one bottom-left rounded rectangle with three horizontal slider controls, one upper-right rectangular crop/selection frame around a dome preview, and one bright blue glowing cursor arrow pointing at the lamp evolution. On the right, place the final physical {argument name="product color and material" default="glossy warm-white ceramic"} lamp on a rectangular speckled stone platform: a smooth hemispherical dome shade, small brass stem, softly glowing warm light under the shade, and a round bulbous base matching the sketch. Surround the scene with studio props: leafy green plants partially visible at the upper left and upper right, a white cup holding paintbrushes at the far right, a small round tray with a red-orange block in the lower right, and stacked sample boards or notebooks in the lower left. The blue cutting mat should cover the tabletop with precise grid lines, reinforcing a design-prototyping environment. No visible text, no logos, no watermark; emphasize the concept of turning a hand sketch into a finished product through transparent AI editing layers.