GPT Image 2 prompts

Create a photorealistic luxury wedding portrait of a cute {argument name="cat breed" default="British Shorthair"} cat as the groom, sitting upright on a garden aisle. The cat has plush {argument name="fur color" default="gray tabby"} fur, round golden eyes, a small pink nose, alert ears, and a calm slightly curious expression while looking directly at the camera. Dress the cat in a formal white ruffled collar with a large ivory satin bow tie and one cream rose boutonniere pinned at the chest. Behind and slightly to the right, place a translucent white bridal veil standing upright like an invisible bride, edged with lace; on the veil area add a glowing white abstract Capafy-style logo made from one central diamond/star node connected to 6 round nodes. The setting is an elegant old European garden wedding venue: a stone path covered with pale rose petals, abundant pink, peach, and cream roses on both sides, classical stone statues and urns, an arched stone doorway in the background, tall trees and greenery forming a shaded canopy. Use warm golden-hour backlighting, soft bokeh, realistic fur detail, natural wedding-photography skin/fabric-style realism applied to cat fur and satin, shallow depth of field, cinematic romantic atmosphere, high-end editorial wedding photo look. Vertical 4:5 composition, cat centered slightly left, veil/logo centered right, foreground roses softly blurred, no text, no watermark, no humans.

Goal: Create a photorealistic wide-angle interior design visualization of a small child’s bedroom decorated with IKEA furniture, including annotation callouts naming each new item.

Canvas: Landscape 4:3 image, photographed from the doorway or front-right corner of the room with a slightly elevated eye-level perspective, showing the entire bedroom from left desk area to right bed and closet doors.

Room details: Beige walls, beige carpet, white trim, a large window with white blinds on the left wall, white double closet doors on the right wall, warm natural daylight mixed with soft lamp light. Keep the room tidy, realistic, and practical for a young child.

Layout and furniture: On the far left wall place a white {argument name="desk model" default="LÄRANDE desk"} with a gray rolling desk chair, pencil cups, and a small book or tablet on top. Under the window and along the back wall place a black cube storage unit labeled {argument name="toy storage model" default="KALLAX toy storage"}, filled with exactly 3 gray fabric bins across the bottom row, books, toy vehicles, and small toys. Above the toy storage hang a framed dinosaur print. To the right of the toy storage place a dark wood {argument name="dresser model" default="STORKLINTA dresser"} with exactly 3 large drawers, small plants, a framed photo, and a small lamp on top. On the right side place a white twin {argument name="bed model" default="SLÄKT bed"} with a white headboard, blue quilt, dinosaur-patterned pillow and sheets, positioned lengthwise toward the viewer. Place a black arched floor lamp behind the bed near the dresser. Add a green children’s road-map play rug centered on the carpet between the desk, storage, and bed.

Annotation callouts: Add exactly 7 white rounded-rectangle labels with black text and thin gray borders, each with a white arrow pointing to the relevant item. The labels must read exactly: 1) “LÄRANDE desk” pointing to the desk, 2) “KALLAX toy storage” pointing to the black cube storage, 3) “DRÖNA bins” pointing to the gray fabric bins, 4) “STORKLINTA dresser” pointing to the dresser, 5) “HEKTAR floor lamp” pointing to the black floor lamp, 6) “SLÄKT headboard” pointing to the white headboard, 7) “SLÄKT bed” pointing to the footboard or bed frame.

Visual style: Realistic smartphone interior photo with clean AI home-staging polish, warm cozy lighting, accurate perspective, subtle shadows, readable labels, no people, no clutter, no watermark.

Create a clean futuristic 3D wireframe conversion of a close-up bust portrait of {argument name="subject" default="a young woman"}, preserving the original pose and composition exactly: the subject faces forward with a slightly downward camera angle, short bob-length hair framing the head, one visible ear on the left side, a slender neck, shoulders, and upper chest visible. Render the entire figure as a translucent white mesh made from thin, precise gray polygon and contour lines, showing both outer silhouettes and internal 3D surface structure like a modeling viewport or technical anatomy sketch. Use a pure white background with no scenery. Include exactly 2 main visible elements: 1 wireframe bust portrait and 1 centered opaque square face-cover block. The face-cover block should be a large pale light-gray square with a soft subtle gradient, placed over the entire face from forehead to chin, hiding all facial features while leaving hair, ear, neck, shoulders, and torso mesh visible around it. Emphasize fine crosshatched curved wire lines in the hair, neck, shoulders, and clothing/body surfaces, with delicate semi-transparent linework and high-resolution sharpness. Style: minimalist sci-fi technical drawing, monochrome white and light gray, airy, precise, hollow transparent 3D model aesthetic, no text, no watermark, no color accents.

Create a cinematic 1940s film-noir street scene in high-contrast black and white, widescreen 16:9. In the left foreground, show exactly 1 large heavyset detective or gangster seen from behind, wearing a wide-brim fedora with a feather and studded band, a dark long coat with a pale fur collar and fur cuffs, gloves, and holding exactly 1 rolled document or blueprint in his right hand. He faces across a rain-slick city sidewalk toward a brick storefront on the right with a glowing sign reading {argument name="agency sign text" default="KORSCHE DETECTIVE AGENCY"}. The office entrance has dark shutters, a glass door, and a small paper notice on the door reading {argument name="door notice text" default="NOTICE"}. Include exactly 1 vintage black car parked on the left curb, exactly 3 visible street or exterior lamps casting white pools of light, tall shadowy skyscrapers in the far-left background, wet reflective pavement, misty night air, deep shadows, rim lighting around the character’s hat and fur, and subtle warm yellow highlights only in the storefront and lamp glow. Style: graphic novel realism mixed with classic noir cinematography, dramatic low-key lighting, moody urban detective atmosphere, sharp architectural detail, no modern objects, no extra people, no watermark.

Goal: Create a vertical social-media food review poster that turns a casual restaurant table photo into an Instagram-style barbecue feast cover, with playful hand-drawn Chinese marker annotations and arrows.

Canvas: Square 1:1 image, warm indoor restaurant lighting, close-up overhead three-quarter angle. The background is a dark wooden table with a softly blurred chair and wall behind it. Use a rich brown, orange, and golden color palette with high contrast, glossy food highlights, and shallow depth of field.

Main layout: Center the composition on a rectangular metal grill tray filled with skewers and roasted items, angled from lower left to upper right. Behind the tray, place a small black handled pot filled with spicy beef stew. At the bottom edge, partially show sauce dishes and side dishes cropped by the frame. Leave dark negative space at the top for large handwritten title text.

Counted food elements: Show exactly 7 prominent food groups: 1 black handled pot of spicy beef stew at the back center, 1 large roasted fish fillet or grilled flatbread-like item across the left side of the tray, 1 cluster of thin grilled skewers extending to the right, 1 group of yellow pineapple chunks near the center front, 1 group of red chili-covered meat skewers at the lower right, 1 set of small dipping sauce bowls along the bottom edge, and 1 partially cropped grilled dish or skewer plate at the bottom right. Show exactly 4 visible small sauce bowls at the bottom: one red chili oil sauce, one bright orange sauce, one green chopped scallion or herb sauce, and one small pale tofu or garnish bowl.

Handwritten graphic overlays: Add colorful doodles in white, yellow, pink, and orange, as if drawn with paint markers. Include exactly 8 annotation clusters: 1 large headline at the top left, 1 pink speech-bubble badge at the top right, 1 arrow label pointing to the beef pot, 1 arrow label pointing to the roasted chips/fish on the left, 1 arrow label pointing to the skewer platter on the right, 1 label near the bottom left for small side dishes, 1 label near the lower right for crispy skewers, and 1 decorative underline beneath the headline. Add small stars, hearts, sparkles, curved arrows, and scribble accents around the text.

Text content: Use a big playful Chinese headline in thick hand lettering: {argument name="headline text" default="今天也要 开心心干饭呀!"}. Put a pink outlined speech bubble at the top right with the badge text {argument name="badge text" default="聚会首选 氛围感拉满~"}. Add a label above the pot reading {argument name="pot label" default="香辣过瘾 牛肉锅~"}. Add a left label reading {argument name="roasted item label" default="烤馍片 焦香酥脆~"}. Add a right label reading {argument name="skewer label" default="烤串拼盘 种类超丰富!"}. Other small labels should be rendered as casual Chinese-style food review callouts, but keep them secondary and less dominant.

Visual style: Photorealistic appetizing food photography combined with hand-drawn editorial doodles, like a trendy Chinese restaurant recommendation post. Food should look freshly grilled with oil sheen, sesame, chili flakes, char marks, and steam-like warmth. The typography must feel hand-lettered, imperfect, energetic, and cute, not like a clean digital font.

Constraints: No people, no brand logos, no watermark, no extra UI frame. Keep the food realistic and the annotations legible, with the headline as the largest text element.

Using REFERENCE_0, edit only the cat’s lower body by adding snug dark gray yoga pants/leggings. Keep the cat’s pose, head, upper torso, fur pattern, tail, collar, floor mat, cardboard, lighting, camera angle, and photo-realistic quality unchanged. The leggings should start around the cat’s waist/belly area with a wide waistband, wrap naturally around the hips and hind legs, include realistic fabric stretch, seams, folds, and shadows, and leave the tail visible outside the pants. Do not alter the rest of the image.

Create a photorealistic architectural interior rendering of a modern dessert cafe named {argument name="cafe name" default="Some ice"}. Show a wide-angle view from the customer seating area toward the service counter in a long, narrow, high-ceiling space. Use a warm minimalist Japanese/Scandinavian style with beige plaster walls, gray concrete floor, textured gray tile at the back wall, light wood ceilings and shelving, and soft indirect cove lighting. The main composition should include exactly 1 large white tiled service counter on the right-center, exactly 1 desktop computer on the counter, exactly 1 suspended wood canopy above the counter with recessed spotlights, exactly 1 gray vertical concrete column on the far right, exactly 1 backlit brand sign above the counter reading {argument name="cafe name" default="Some ice"} in dark letters with a small yellow accent, exactly 4 illuminated menu boards behind the counter, exactly 2 empty wood shelves on the right wall, exactly 3 wall posters on the left wall advertising desserts, exactly 5 round cafe tables visible, and exactly 11 visible chairs/stools: 4 wooden chairs around the two front-left tables, 4 wooden chairs in the middle corridor seating area, 1 tall stool at the counter, 1 pale yellow chair at the front-right table, and 1 dark charcoal chair at the right table. Add exactly 2 small wall-mounted spotlights above the left posters and exactly 2 slim vertical railing posts with a horizontal handrail in front of the counter. The cafe should feel quiet, clean, and newly designed, with soft warm daylight and artificial lighting, realistic shadows, polished materials, and a professional commercial interior visualization look. Avoid people, clutter, food closeups, watermarks, and extra signage.

Create a photorealistic architectural interior rendering of a modern minimalist ice cream shop named {argument name="shop name" default="Some ice"}, viewed from the front seating area looking toward the service counter in warm late-afternoon sunlight. The space is tall and narrow with cream plaster walls, gray concrete flooring, a gray textured back wall, white square-tile counter, warm wood ceiling canopy, hidden cove lighting, and soft shadows from tree leaves cast across the left wall and floor. Place the large shop logo above the counter, dark gray letters with a small yellow accent. The service area contains exactly 1 white tiled front counter, 1 large silver desktop computer on the counter, 1 staff member behind the counter wearing a cream shirt and beige apron, and 4 illuminated menu boards on the back wall labeled Signature, Sweet, Yogurt, and Classic with dessert photos. On the right wall, include exactly 3 wooden retail shelves displaying small branded paper bags or boxes. On the left wall, include exactly 3 framed promotional posters under a slim brass wall light, one featuring a cute mascot and the text {argument name="poster headline" default="Have Someice Day"}. Furnish the customer area with exactly 6 round café tables total: 2 in the left foreground, 2 smaller tables along the left middle wall, and 2 on the right side; include exactly 10 visible chairs total, mostly warm bentwood chairs with chrome legs plus 1 pale yellow chair and 1 black chair on the right. Add a staircase and narrow passage receding into the back left, with wood stair treads, a black metal railing, and a glimpse of brick wall and glass door at the far rear. Use a calm Scandinavian-Japanese café aesthetic, natural materials, realistic global illumination, soft depth, clean composition, no people except the single staff member, no clutter, no watermark, 3:2 horizontal aspect ratio.

Create an epic dark fantasy cartographic panorama in a cinematic realistic digital painting style. The scene is a vast medieval coastal world viewed from a very high aerial three-quarter angle, like a game map or fantasy strategy map. In the foreground center is a huge fortified harbor city packed with stone buildings, cathedral towers, walls, docks, small ships, and bridges, surrounded by blue-black ocean and river deltas. A thin bright red route line with small arrowheads is drawn over the landscape, starting at the harbor city, curving across the sea to the right toward a volcanic island, then continuing back through the city and winding northward through forests, rivers, ruined keeps, mountain passes, and finally climbing a monumental vertical ice wall into a frozen northern wasteland. Count exactly 1 continuous red travel path, with multiple visible arrowheads along it. On the right, show a jagged black volcanic island with glowing orange lava, fire, smoke, and ash clouds. In the middle distance, show dark pine forests, scattered castles, villages, roads, rocky mountains, marshy river channels, and mist. Across the upper third, show a colossal icy cliff or glacier wall spanning the entire width of the land, with a frozen plateau beyond it containing snowfields, ancient ruins, spires, distant mountains, and storm clouds. Use cold desaturated blues, grays, charcoal, icy whites, muted browns, and small accents of fiery orange lava and vivid red for the route. Lighting is dramatic and overcast, with heavy clouds, fog banks, sea spray, and high-detail worldbuilding. No labels, no UI, no border, no watermark, no readable text. Aspect ratio 16:9, ultra-wide cinematic composition, highly detailed, sharp, atmospheric, epic scale.

Create a cinematic photorealistic night street scene in rainy {argument name="city" default="Tokyo"}, shot in a wide 16:9 frame with a moody film-still look. Show exactly 1 main subject: a young woman standing on the wet sidewalk outside a brightly lit convenience store, wearing an oversized light beige rain jacket over a long dark skirt, holding exactly 1 transparent plastic umbrella and exactly 1 white takeaway cup near her mouth; her face should be softly obscured by shadow, rain, and shallow focus rather than clearly identifiable. Include exactly 1 yellow taxi on the left side of the street, exactly 1 glowing red traffic light in the upper-left area, and exactly 1 convenience store frontage on the right with warm fluorescent interior lighting, glass windows, shelves, and small Japanese-style notices including a visible “24 OPEN” sign. The asphalt and sidewalk are soaked, with strong reflections of red, amber, and white neon stretching into the distance; raindrops and bokeh speckles appear on the lens, with a dark urban background receding into soft blur. Use {argument name="lighting mood" default="warm convenience-store glow against cold rainy darkness"}, {argument name="camera style" default="cinematic street photography, 35mm lens, shallow depth of field"}, and {argument name="color palette" default="black, wet asphalt gray, amber gold, muted red, and soft beige"}. Make the image realistic, atmospheric, slightly grainy, high dynamic range, no watermark, no extra people in focus, and no readable brand logos.

Goal: Create a vertical Art Nouveau luxury perfume advertisement poster featuring {argument name="character name" default="VoxCat"} as an elegant fantasy anime heroine presenting a perfume bottle as the main foreground visual anchor.

Canvas: Tall portrait poster, 2:3 aspect ratio, highly detailed premium packaging illustration, full-bleed composition with a thin black outer border and ornate metallic gold inner frame.

Layout: Center the character from upper torso to hips, angled slightly right, with her arm extended toward the viewer. The perfume bottle must dominate the lower-left foreground and create strong depth of field, appearing much larger than the torso behind it. Place a large plain beige square over the character's face, centered on the face area, as if the face is intentionally blocked. Put one ornate monogram crest in the upper-left corner and one empty decorative product-label plaque in the lower-right corner.

Subject details: One anime-style female fantasy character with long flowing {argument name="hair color" default="pale blonde"} hair, wind-swept strands, soft luminous skin, and a white-and-gold armored couture dress with gemstone accents and layered translucent fabric. Her visible hand holds the bottle close to the camera, with carefully drawn fingers and glossy nails. Add exactly 6 large white lilies: 3 lilies tucked into her hair near the upper head, 1 lily in the bottom-left corner, and 2 lilies decorating the lower-right label plaque. Add purple flowers and dark green leaves as secondary botanical decoration around the border.

Perfume bottle: One round faceted glass perfume bottle in the foreground, filled with {argument name="perfume color" default="iridescent lavender-gold liquid"}. It has a large crystal sphere stopper, a gold engraved cap, and ornate gold filigree on the bottle face. The central medallion shows a small gold cat silhouette framed by laurel branches and a small blue gemstone at the bottom. Use glass refraction, etched details, bright sparkles, and commercial product-photography highlights.

Decorative packaging style: Deep emerald, midnight purple, and antique gold palette; jewel-tone background with circular halo geometry behind the character; swirling vine lines, gold filigree, thin arcs, stars, gemstones, and botanical motifs. The overall feel should be a collectible limited-edition perfume box mixed with a New Art poster.

Text and marks: Upper-left crest contains a large gold letter {argument name="monogram letter" default="V"} inside a heart-shaped vine frame with a dangling purple gemstone. Lower-right plaque is dark purple velvet with ornate gold trim and mostly empty space for copy; add thin blank gold lines near the bottom and a small handwritten brand signature reading {argument name="brand signature" default="voxCat"} in the lower-right corner.

Constraints: Use exactly 1 character, 1 perfume bottle, 1 upper-left monogram crest, 1 lower-right blank plaque, and 6 large white lilies. Keep the face covered by the plain beige square. No extra readable text besides the monogram letter and brand signature. Make the image polished, intricate, luminous, and high-end, like fantasy character art fused with luxury fragrance packaging.

Create a live-action cinematic cosplay photograph of an {argument name="character name" default="elegant silver-haired fantasy heroine"} seated alone on an old city bus, framed from the waist up beside the left-side bus windows. The subject is an adult cosplayer with a believable pose, looking quietly out the rainy window rather than at the camera; keep the face mostly hidden by long bangs and soft shadow, with no glamour retouching. Use natural skin texture and realistic fabric texture. The costume should feel elegant and integrated into daily life: exactly 6 visible costume cues, including long wavy {argument name="hair color" default="silver-gray"} hair, a large black ribbon bow behind the head, a white high-collar coat, pale blue and gold embroidered shoulder panels, a large teal-blue chest bow with jeweled ornament, and dangling purple gemstone earrings. The setting is the interior of a worn public bus with dark green-black vinyl seats, metal seat rails in the foreground, tall fogged windows on the left, and a blurred low-saturation city street outside after rain with bare trees, streetlights, and distant traffic. Shoot in a 16:9 film still composition with the subject on the right third, bus windows leading into the left background, shallow depth of field, muted teal-gray color grading, soft overcast winter light, slight glass reflections, subtle motion blur outside, visible film grain, old DV / early-2000s movie still texture, realistic lens softness, and a melancholic quiet mood. Add a small white angular V-like logo in the upper-left corner and a small lowercase serif watermark reading {argument name="watermark text" default="voxcat"} in the lower-right corner. Avoid illustration, anime rendering, plastic skin, studio lighting, exaggerated cosplay posing, and overly clean AI perfection.

Create a poetic character illustration as if it were a collectible still from a French 1980s animated film. Show {argument name="character name" default="Voxcat"}, an elegant fantasy princess-like woman seated alone inside an old city bus, framed in a double-layer bus-window perspective. She has very long {argument name="hair color" default="copper red"} hair flowing over her shoulders and down the seat, a delicate horned crown or ornate headpiece with small dark curved horns, translucent blue veil fabric, gold filigree jewelry, turquoise gemstone ornaments, and a pale blue-and-white ceremonial dress with sheer sleeves and soft draped fabric. Her face should be deliberately obscured by a plain rectangular censor-like block, while the body and costume remain detailed. The bus interior is muted teal-blue with patterned fabric seats, metal window frames, worn panels, and a damp reflective window ledge. Outside the large windows, paint a rainy European boulevard with a bridge railing, bare trees, lampposts, distant classical buildings, wet pavement reflections, atmospheric mist, and a small red traffic light in the far right background. Use hand-painted cel animation language, transparent watercolor edges, visible paper grain, soft ink outlines, subdued blue-gray and beige palette, gentle melancholic lighting, and layered reflections on the glass so the city, character, and bus interior overlap poetically rather than photorealistically. Include exactly 3 visible text/logo elements: a large white monogram logo with the text "VOXCAT" in the upper-left dark bus panel, a tiny square "V VOXCAT" sticker on the seat back to the right of the character, and a handwritten white "voxCAT" signature in the bottom-right corner. Landscape 16:9 composition with rough painted border edges, no extra characters, no modern digital gloss, no photographic realism.

Create a high-end Japanese kaiseki food photograph on a glossy black lacquer plate against a dark black background, dramatic low-key studio lighting, shallow depth of field, macro lens realism. The composition is diagonal and elegant: on the right foreground place exactly 3 thick slices of medium-rare wagyu beef steak, seared brown edges, pink marbled centers, glossy sauce pooling beneath, topped with tiny flecks of gold leaf. On the left midground place exactly 4 translucent white sashimi slices of {argument name="fish type" default="white fish"}, folded and overlapping, with wet reflective highlights, tiny purple edible flowers, and scattered gold leaf. Add exactly 1 large red-orange maple leaf in the front left foreground, exactly 1 curved dry-brushed gold powder arc sweeping around the left and bottom of the plate, exactly 1 upright purple herb flower sprig between the fish and beef, exactly 2 small green wasabi mounds near the back right and back center, exactly 2 pale green rolled garnish rings behind the beef, and exactly 1 small autumn branch in the back with red maple leaves and a green bud. Use {argument name="plate style" default="glossy black lacquer plate"}, {argument name="lighting mood" default="dramatic low-key luxury restaurant lighting"}, and {argument name="background color" default="deep black"}. Make it photorealistic, luxurious, refined, with crisp textures on the food, soft bokeh in the rear decorations, no text, no hands, no utensils, no watermark.

Goal: Create a delicate commercial product visual for a 1/7 scale pre-painted anime figure displayed like a refreshing summer dessert drink.

Canvas: Square 4:5-ish product photo composition, soft high-key lighting, pale lavender, icy blue, and white color palette, shallow depth of field, glossy realistic rendering with a collectible figure advertisement feel.

Main subject: Center-right, show one anime-style female figure named {argument name="character name" default="Emma Crandall"} seated inside a clear cylindrical glass tumbler filled with transparent sparkling water. The figure has very long flowing {argument name="hair color" default="pale lavender"} hair drifting around her in the water, with small dark purple bow accessories. Her face is covered by a smooth opaque pale pink rectangular censor block with no facial features. She wears one frilly lavender summer-lily swimsuit dress with layered ruffles, lace trim, a small dark bow at the chest, a dark bow at the hip, and a translucent scrunchie bracelet. Her pose: one hand near her cheek, the other resting over the glass rim, bare legs dangling underwater with pointed bare feet.

Visible counted elements: exactly 1 female figure, 1 transparent glass tumbler, 1 lemon slice on the right background, 3 prominent ice cubes in the foreground/background, many tiny air bubbles suspended in the water, 2 large blurred pale purple flowers in the background/foreground, and 3 dark purple bow accents on the figure outfit/hair.

Layout: The glass occupies the lower two-thirds of the image, with the figure emerging from the waterline. The top-left area contains elegant product typography and ornament lines, leaving the figure dominant on the right. Background is a dreamy macro tabletop scene with bokeh highlights, icy reflections, blurred flowers, and a lemon wedge.

Text content: Add exactly 4 text lines/blocks in the upper-left. Line 1 large Japanese title: {argument name="Japanese title" default="エマ・クランドール"}. Line 2 smaller cursive English name: {argument name="English name" default="Emma Crandall"}. Line 3 small Japanese subtitle: {argument name="version subtitle" default="サマーリリィ・アフタヌーン Ver."}. Line 4 on a lavender ribbon banner: {argument name="scale label" default="1/7スケール 塗装済み完成品フィギュア"}. Include a thin decorative floral divider between the subtitle and ribbon.

Visual style: Ultra-polished anime figure photography, resin/PVC collectible texture, transparent glass refraction, realistic water caustics, sparkling highlights, soft bloom, pastel elegance, luxury figure promotional key visual.

Constraints: Keep all text legible and placed only in the upper-left. Do not add extra characters, extra labels, logos, watermarks, or packaging. Maintain the censored blank rectangular face exactly as a flat opaque block.

Goal: Create a vertical Chinese filming-direction storyboard infographic for a high-speed train short video, showing how to shoot a passenger in a slightly uncomfortable state from farther away.

Canvas: Tall portrait layout, about 3:4 ratio, white background, black and dark-brown Chinese typography, clean documentary storyboard style.

Text content: Use the visible Chinese copy from {argument name="Chinese text" default="镜头1｜远一点拍出不适状态; 一、机位示意图（俯视图）; 窗户; 人物; 座椅; 包; 过道; 机位; 二、预期画面（示意图）; 拍摄要点; 机位放在过道斜前方; 画面带到座椅、窗户和包; 人物轻捂小腹，表情微微不适; 不要拍得太夸张，像真实乘客状态"}.

Layout: At the top, place a large bold headline centered. Below it, create exactly 2 stacked image panels separated by a thin dotted horizontal divider. Panel 1 is labeled as a camera-position diagram from an overhead perspective. Panel 2 is labeled as the expected shot preview. At the bottom, add a pale cream rounded rectangle notes box with a brown pill-shaped label on the left and a simple yellow lightbulb icon on the right.

Panel 1 subject and composition: Draw an illustrated overhead/three-quarter view inside a high-speed train carriage. A young woman with long dark hair and a deliberately blurred face sits by the window on a dark navy-blue train seat with a white headrest cover. She wears a cream cardigan or sweater, light blue jeans, and white sneakers. Her left hand rests over her lower abdomen to suggest mild discomfort, and her right hand reaches toward a small beige backpack placed on the adjacent seat. Show a large train window to her left with motion-blurred green scenery outside, blue paired seats around her, armrests, and the aisle on the right. Include exactly 6 white rounded callout labels with blue borders and small pointer dots: window, person, seat, bag, aisle, and camera position. Place a black DSLR-style camera on the aisle floor in the lower right foreground. Add exactly 3 directional graphics in this panel: one thick red arrow from the camera toward the seated passenger/bag area, one thick red curved arrow indicating the aisle direction, and one yellow dashed arrow pointing from the camera toward the subject.

Panel 2 subject and composition: Create a realistic cinematic preview frame from the camera angle in the aisle diagonally in front of the passenger. The woman sits in the left-center by the train window, face blurred, one hand lightly holding her abdomen, the other hand near or on the beige backpack on the neighboring seat. Keep the dark blue seats, white headrest covers, beige interior wall, large window with motion-blurred city/green scenery, long aisle receding into the background, and a small monitor screen at the far end of the carriage. The mood should be natural and restrained, not melodramatic.

Bottom notes box: Include exactly 4 bullet points, matching the shooting guidance text. Use a warm cream background, brown label badge, dark text, and a minimal yellow outline lightbulb icon.

Visual style: Semi-realistic storyboard illustration mixed with clean instructional infographic design, soft lighting, slightly textured digital painting, precise callouts, realistic high-speed rail interior, clear teaching-diagram readability. Avoid watermarks, logos, or extra panels.

Create a cinematic anime key visual for a sci-fi fantasy project titled {argument name="title text" default="SHADOW SYNC"}. Wide 16:9 poster composition, ultra-detailed, split-world contrast: the left half is a brilliant utopian future city in daylight with white crystalline skyscrapers, blue sky, sun flare, greenery, waterways, sleek architecture, and golden energy arcs; the right half is a dark cyberpunk megacity at night with storm clouds, neon blue and purple lights, dense towers, holographic signs, flying structures, and rainlike atmosphere. Center foreground: two back-to-back protagonists standing in profile, separated by the divide between worlds. On the left is a blonde young woman with a low ponytail, slim athletic build, wearing a black sleeveless top and dark short outfit, lit by warm daylight; her face is intentionally covered by a soft square beige blur. On the right is a black-haired young man in futuristic black tactical armor, lit by cool neon blue, with a glowing cyan infinity symbol on his shoulder; his face is intentionally covered by a soft square dark blur. Add exactly 2 large flags or banners: one on the bright left side, purple fabric with gold streaks and a glowing blue circular emblem; one on the dark right side, black fabric with a glowing cyan infinity emblem. Add multiple visible infinity symbols across the cyberpunk side, including one large cyan infinity hologram near the center-right, one on the man's shoulder, one on the right flag, one near the top tower display, and one small sign near the lower-right buildings. Include tower text on the dark side reading "REASON", "AGI", "VALUE", "REASON" stacked vertically. Across the lower center, place the main title in large sharp italic cracked lettering, white on the left blending into electric purple on the right, with energy streaks and shards cutting through it. Beneath the title, add small Japanese subtitle text: {argument name="subtitle text" default="影は、未来を選ぶ"}. Visual style: high-end anime illustration, dramatic lighting, extreme detail, glossy sci-fi rendering, dynamic particles, petals and glass shards, purple lightning, blue neon bloom, high contrast, poster-quality composition. No extra logos, no watermark, no additional characters.

Goal: Create an ornate fantasy “Transparency Check” diagnostic card featuring an ethereal moonlit princess character and a Japanese results UI.

Canvas: Square 1:1 image, 1024×1024 style, luminous blue-violet palette, high-detail anime fantasy illustration mixed with elegant magical infographic design. The composition is split vertically: the left 40% is a translucent glass results panel, and the right 60% shows the full-body character standing in a crystal palace.

Main subject: On the right, depict a delicate female fantasy character with extremely long flowing silver-white hair, pale skin, and a graceful posture. Her face is covered by a plain muted pink-beige square censor block. She wears an elaborate layered dress in white, silver, lavender, and deep violet with a dark corset bodice, star-shaped ornaments, gemstone chains, lace, translucent veils, and sparkling crystal embroidery. Her skirt is voluminous and semi-transparent with moon-and-star motifs, flowing fabric ribbons, and glittering highlights. Surround her with hanging crystals, glass bubbles, moonlight, refractions, and a grand icy palace background with columns, arches, chandeliers, and circular celestial architecture.

Left diagnostic panel: Use a transparent frosted-glass rectangular panel with thin ornate silver borders, filigree corners, tiny star accents, and glowing divider lines. All text should look like elegant Japanese fantasy typography, mostly white with pale lavender and cyan glow.

Text content: At the top, large Japanese title text: 「透明度診断」 with small English subtitle “TRANSPARENCY CHECK”. Below it, a small label: 「診断結果」. The main score line reads 「透明度：{argument name="transparency score" default="92"}%」 with the number very large and glowing. Under the score, add a horizontal iridescent progress bar filled nearly to the end, with a small crystal marker on the right. Beneath the bar, show the character name: 「{argument name="character name" default="セシル＝リア＝クラランドール"}」.

Rank card: Include exactly 1 rank section with a circular crescent-moon crystal emblem on the left and the label 「透明度ランク」 above a huge glowing rank value 「{argument name="rank" default="SS"}」 on the right.

Type and attribute section: Include exactly 2 labeled rows with small icons. Row 1 label 「透明タイプ」 and value 「{argument name="transparency type" default="月光水晶型"}」 with a crystal droplet icon. Row 2 label 「属性」 and value 「{argument name="attribute" default="月光水属性"}」 with a crescent moon water icon.

Quote section: Add a centered poetic line in Japanese: 「静かな意志は、澄みきった光となる。」 with small ornamental flourishes above and below.

Auxiliary indicators: At the bottom of the panel, create exactly 4 auxiliary metric rows next to a pale blue-purple radar chart. The radar chart has a pentagonal web with a softly glowing filled polygon. The 4 rows must be: 「うるみ度 95」 with a water droplet icon, 「きらめき 90」 with a sparkle icon, 「儚さ 84」 with a feather icon, and 「神秘性 93」 with a crescent icon.

Visual style: Ultra-detailed, luminous, crystalline, dreamy, transparent materials, moonlight, prismatic reflections, soft bokeh, floating bubbles, tiny stars, silver filigree, high fantasy anime art, elegant UI overlay, cool blue lavender lighting, magical atmosphere.

Constraints: Keep the exact visible structure and counts: 1 title area, 1 score area, 1 progress bar, 1 name line, 1 rank card, 2 type/attribute rows, 1 quote, 1 radar chart, and 4 auxiliary metric rows. Do not add extra characters, logos, watermarks, or modern objects. Preserve the face-covering square over the character’s face.

Goal: Create a warm fantasy manga page about forest children listening to elf and dwarf music, in a hand-painted storybook manga style. Use {argument name="language of visible text" default="Japanese"} for all dialogue and narration, with vertical manga lettering.

Canvas: One vertical comic page, about 3:4 aspect ratio, centered on a pale cream background with a thin black page border. Inside the page, use fine black panel gutters and watercolor-like fantasy illustration with dense foliage, flowers, dappled sunlight, and golden-green tones.

Layout: Use exactly 5 comic panels. Panel 1 is a wide full-width top panel. Panel 2 is a wide full-width middle panel. Panel 3 is the lower-left panel of a split row. Panel 4 is the lower-right panel of the same split row. Panel 5 is a wide full-width final panel at the bottom. Add small musical-note symbols in panels 2 and 4.

Characters: Include exactly 6 recurring characters: 1 tall blonde elf woman with pointed ears, flower hair ornaments, elegant white-and-green embroidered dress, holding a lute; 1 stout bearded dwarf man with pointed ears, round cap, layered travel clothes, and a small hand drum; 1 small tan-haired animal-eared child; 1 blonde elf girl child; 1 dark-haired human boy in a blue tunic and red scarf; and 1 braided brown-haired dwarf girl child with a decorated headband. The elf woman is {argument name="elf musician" default="a gentle blonde elf woman playing a lute"}; the dwarf musician is {argument name="dwarf musician" default="a cheerful red-bearded dwarf playing a hand drum"}. Keep the children expressive, wide-eyed, and delighted.

Panel details and text: Panel 1 shows the dwarf and elf seated under trees performing music while children gather around them; use exactly 2 speech bubbles: one from the elf saying {argument name="elf opening line" default="風がそよぎ 木々がうたう 大地の恵みに ありがとう"}, and one from a child saying "わっっっは！いい歌だ！". Panel 2 shows close-up reactions of the children from left to right, smiling and clasping hands as music floats around; use exactly 4 speech bubbles: "きれいな 音色だね…", "おじさんの リズムが すっごく楽しい！", "ふたりの音が ぴったり 合ってる！", and "もっと 聴きたいな！". Panel 3, lower-left, shows the six characters walking away together along a sunny forest path toward a cozy village house; use exactly 2 speech bubbles: "うん！" and "ねえみんなも 一緒に歌おうよ！". Panel 4, lower-right, shows the elf and dwarf playing side by side, smiling warmly; use exactly 2 speech bubbles: "音楽は ぶつけ合いじゃなく 橋になるんだな" and "ええ 一緒に奏でると 心が通うの". Panel 5 shows the elf and dwarf seated in a flower-filled clearing, surrounded by the four children listening happily; include exactly 2 vertical narration boxes: left box says "小さな村に、 音楽と笑顔が あふれる日々が 続いていきました。" and right box says "エルフの優しい旋律と ドワーフの楽しいリズム、 そしてみんなの歌声が ひとつになってーー". Add a small decorative final caption at bottom left reading "❀おわり❀".

Visual style: Detailed Japanese fantasy manga with watercolor textures, ink outlines, ornate clothing, soft faces, expressive eyes, luminous forest atmosphere, fairytale village background, warm sunlight, and gentle wholesome mood. Avoid modern objects, avoid extra panels, avoid extra characters, and keep all text inside the specified bubbles or narration boxes.

Create a simple full-body vector illustration on a clean white background of a friendly young woman standing upright, slightly tall with long legs, suitable as an illustration for student internship or presentation materials. She has {argument name="hair color" default="black"} shoulder-length bob hair with a side part, a minimal faceless or softly obscured face, and a calm professional pose. Dress her in an {argument name="jacket color" default="bright orange"} cardigan-style jacket with exactly 3 visible black buttons over a white shirt, beige high-waisted wide-leg trousers, white socks, and orange flats. Her left arm is extended holding an open gray laptop angled outward; her right hand rests casually in her trouser pocket. Add exactly 1 floating object to the right of her upper body: a teal closed book or notebook with a thick black outline, white page edges, and a centered white heart icon on the cover. Use a cute modern Japanese-style flat illustration aesthetic with thick rounded black outlines, soft cream shading, simple shapes, no text, no shadows beyond subtle flat shading, and plenty of white negative space.

Create a highly detailed cinematic steampunk fantasy portrait of a cute small {argument name="dog breed" default="Shih Tzu"} standing full body in the center of the frame, looking directly at the viewer with large glossy dark eyes and a calm heroic expression. The dog has fluffy cream and tan fur, long floppy ears, a rounded muzzle, and a plume-like tail curling behind its right side. Dress it as an ornate Victorian airship pilot: exactly 6 main outfit/accessory elements are visible — 1 brown leather aviator cap with brass goggles on top, 1 fitted brown leather jacket with cream shearling collar and cuffs, 1 layered waistcoat and shirt with tiny brass buttons, 1 diagonal strap and belt system with pouches, 1 pair of tailored dark trousers, and 1 pair of worn brown lace-up boots. Add numerous tiny brass gears, compass-like medallions, chains, buckles, pins, and clockwork ornaments on the jacket and belts, but keep the dog’s face unobstructed. Set the scene in a wide {argument name="setting" default="golden steampunk city and airship dock"} with wet reflective cobblestone pavement, ornate brass machinery, pipes, towers, glowing round windows, railings, steam vents, and warm golden lights. Place one large blurred zeppelin airship in the upper right background, with industrial towers and round mechanical structures fading into mist on the left and center background. Use dramatic warm backlighting, soft fog, floating golden dust sparks, shallow depth of field, realistic fur texture, polished brass reflections, rich brown and gold color grading, ultra-detailed 3D/photoreal fantasy styling, symmetrical centered composition, low eye-level camera angle, 4:3 landscape aspect ratio. No text, no watermark, no extra animals, no cropped feet.

Goal: Create a vertical one-page sci-fi shoujo manga page in a gritty watercolor-and-ink style, featuring {argument name="character name" default="Zero"} and a pale gothic heroine in a ruined post-apocalyptic city as a mechanical dragon attacks.

Canvas: Tall portrait manga page, about 2:3 aspect ratio, with a thin black outer border, off-white paper tone, diagonal white panel gutters, rough ink texture, muted gray, lavender, beige, and dusty brown colors.

Layout: Use exactly 6 comic panels. Panel 1 is a large full-width establishing panel across the top half: collapsed urban ruins, rubble, smoke, broken skyscrapers, multiple armored soldiers, a white-haired cybernetic male soldier standing with a rifle, and a pale long-haired girl kneeling in the debris. Panel 2 is a lower-left close-up portrait of the girl with very long platinum hair, black gothic dress, cross necklace, and a flat beige censorship rectangle covering the center of her face. Panel 3 is a narrow upper-right action panel showing a heavy armored/mechanical leg or weapon crashing through debris with speed lines. Panel 4 is a small middle-right panel with a partial character close-up and another beige censorship rectangle covering the face. Panel 5 is a large bottom-left monster reveal: a huge biomechanical dragon or kaiju head roaring with jagged teeth, red mechanical eyes, plated armor, debris flying, and tiny human figures in the foreground. Panel 6 is a bottom-right action panel showing the white-haired armored male protecting or pulling the pale girl back while soldiers and mechanical debris surround them.

Characters: The male lead is a tall, slim, white-haired anime cyborg soldier in black tactical armor with silver mechanical limbs, holding a rifle, calm but alert. The heroine is a delicate pale teenage girl with very long flowing platinum-blonde hair, black frilled gothic outfit, lace details, cross pendant, fragile shoujo expression, and kneeling or being escorted through rubble. Include several background armored troops and ruined military figures, but keep them secondary.

Text content: Include vertical Japanese manga speech bubbles and sound effects. Use exactly 5 speech bubbles: 2 in the top panel, 1 in the girl close-up, 1 in the small middle-right panel, and 1 in the bottom-right panel. Suggested bubble text defaults: {argument name="top left speech" default="旦那様の目が本物か、見つかるといいんだけど……"}, {argument name="top right speech" default="危険区域だからな、足場に確保してもらいたいものだ"}, {argument name="center speech" default="この気配まさか……？"}, {argument name="middle right speech" default="……何か、来るぞ。"}, {argument name="bottom right speech" default="ファンネリア！下がっていろ！"}. Add bold hand-drawn Japanese sound effects: a large red-black roar beside the monster and a dynamic impact sound across the bottom-right panel.

Visual style: Romantic shoujo manga character design mixed with dark military sci-fi, scratchy watercolor wash, dense pen hatching, dramatic diagonal composition, chaotic rubble detail, soft luminous hair, high-detail ruins, cinematic action, gritty texture rather than clean vector art.

Constraints: Keep the page black-and-white/muted color manga-like, avoid modern UI elements, avoid photorealism, no watermark, no English lettering, preserve the exact 6-panel layout and the two beige face-obscuring rectangles.

Goal: Create a high-end dark luxury advertising demonstration graphic for an AI image editor, showing a cosmetic serum ad being decomposed into editable layers.

Canvas: Wide 16:9 banner, deep navy-black background with gold light streaks, cinematic glossy lighting, premium beauty-brand mood, ultra-detailed photorealistic product rendering.

Layout: On the left, show one complete vertical finished poster inside a thin gold border. In the center, place a large glowing gold arrow pointing from the finished poster toward the editable composition. Across the top center, add the title “{argument name="software title" default="Reve 2.0"}” in large elegant serif gold-white lettering, with the subtitle “AI画像をレイヤー編集” beneath it in white. On the center-right, show the enlarged source model image: a woman in a dark satin evening dress photographed from behind with a shiny shoulder, hair rim-lit by warm light, city night bokeh behind her, and a large soft rectangular face-obscuring block over her face. On the far right, display separated editable layer elements floating with blue dashed selection boxes and small square resize handles.

Left finished poster details: The poster advertises “{argument name="product name" default="Noxellia 3X"}” with the English sublabel “NIGHT REPAIR SERUM.” It contains exactly 1 female model, 1 dark blue serum bottle with gold dropper, 1 pale gold flower accent, 1 nighttime city skyline background, and Japanese ad copy in white and gold. Include the large central Japanese headline “翌朝、艶印象が変わる。” in ornate serif type, small top copy “一晩で、肌の修復力を最大化する,” mid copy “高濃度ナイトリペアセラム,” and bottom supporting text “通常の3倍濃度の独自成分を、1滴に凝縮。” The woman’s face in the poster is also covered by a soft rectangular blur/censor block.

Editable layer callouts: Show exactly 7 separate blue dashed selection rectangles with glowing cyan corner handles: 1 around the large central headline “翌朝、艶印象が変わる。”, 1 around the smaller central supporting sentence “通常の3倍濃度の独自成分を、1滴に凝縮。”, 1 around the top-right Japanese phrase “一晩で、肌の修復力を最大化する”, 1 around the right-side phrase “高濃度ナイトリペアセラム”, 1 around the isolated product bottle plus flower, 1 around a cropped city skyline strip, and 1 around a thin gold curved line near the bottom right.

Bottom feature row: Add exactly 3 gold icon-and-text feature blocks along the bottom of the central/right area: 1 droplet icon labeled “独自3倍濃度” with small explanatory Japanese copy, 1 crescent moon and stars icon labeled “夜の集中ケア” with small explanatory Japanese copy, and 1 bottle icon labeled “上質な所有感” with small explanatory Japanese copy.

Visual style: Luxury skincare advertising, black and midnight blue palette, champagne gold typography, sparkling lens flares, elegant serif Japanese type, glowing cyan UI handles, clean composition, realistic reflections on the serum bottle, glossy skin highlights, dramatic rim light, premium cosmetic campaign aesthetic.

Constraints: Keep all Japanese text legible and positioned as described, preserve the dark gold luxury look, use no watermarks or app chrome, make the layer boxes look like non-destructive editing overlays, and include exactly the counted elements and selection boxes.

Goal: Create a vertical Japanese manga-style educational infographic about planning a comic before making it with {argument name="tool name" default="GPT Image 2"}. The main headline should read 「GPT Image 2で 漫画を作る前に決めること」 with 「決めること」 emphasized in yellow, and the subtitle should read 「いきなり作らず、まず“誰に何を伝えるか”を決めよう！」.

Canvas: Portrait 2:3 infographic, clean white background, bright blue accent color, thick rounded comic-panel borders, cheerful instructional tone, polished digital anime illustration. Use large readable Japanese typography with blue, black, yellow, and red emphasis. Add small sparkle and motion-line decorations near the title.

Layout: Arrange exactly 6 numbered manga panels in a 2-column grid, numbered with blue circles 1 through 6 in the upper-left corner of each panel, followed by one wide bottom summary box.

Panel 1: A young male student in a blue hoodie sits at a wooden desk with a laptop that has a cat sticker, a notebook, and a pen. He looks confused with a scribble above his head and a question mark. Add a speech bubble: 「漫画を作りたいけど…何から？」. His face is covered by a soft beige rectangular blur.

Panel 2: The same boy worries at a cluttered desk covered with papers. Around him float exactly 5 planning sheets or sticky notes labeled: 「キャラ」, 「背景」, 「色」, 「セリフ」, and 「効果・演出 コマ割り…」. Add a speech bubble: 「先に絵柄を決めるの？」.

Panel 3: A young female teacher/advisor in a yellow cardigan holds a checklist and raises one finger, standing against a warm yellow radial burst background. Her face is covered by a soft beige rectangular blur. Add a speech bubble: 「まず中身を決めよう！」. The checklist title reads 「チェックリスト」 and shows three checked boxes.

Panel 4: A clean list card titled 「この5つで整理！」. Show exactly 5 colored rows with matching icons and labels: 1 「誰に？」 with a people icon and note 「ターゲット・読者」; 2 「何を？」 with a pencil icon and note 「伝えたいテーマ・内容」; 3 「どう感じて？」 with a heart icon and note 「読後の気持ち・目的」; 4 「どんな雰囲気？」 with sparkles icon and note 「トーン・世界観・絵柄」; 5 「どこまで？」 with a flag icon and note 「コマ数・尺・伝える範囲」.

Panel 5: The boy is now smiling and writing, with sketches and sticky notes on the wall. Add a speech bubble: 「テーマが決まると楽！」. Include a paper titled 「設計メモ」 with exactly 5 checklist items corresponding to the same five questions: 「誰に？」, 「何を？」, 「どう感じて？」, 「どんな雰囲気？」, 「どこまで？」.

Panel 6: The boy in the blue hoodie and the female advisor in the yellow cardigan stand together happily. The boy's face is covered by a soft beige rectangular blur. The advisor holds a storyboard sheet labeled 「1ページ漫画の設計図（6コマ）」 containing exactly 6 small storyboard thumbnails numbered 1 to 6. On the right, place a large yellow starburst speech shape reading 「最初はプロンプトより設計！」 with 「設計！」 in large red letters.

Bottom summary box: A wide rounded rectangle with a blue ribbon label 「まとめ」. Divide it into exactly 3 columns with blue check icons. Column 1 text: 「誰に向けるか決める」 with a people icon. Column 2 text: 「伝えたいことを1つに絞る」 with a pencil icon. Column 3 text: 「作る前に設計する」 with a lightbulb icon.

Visual style: Friendly Japanese educational manga infographic, crisp line art, soft shadows, expressive anime characters, simple school/desk props, blue and yellow highlight palette, high legibility, balanced spacing, no photorealism.

Constraints: Use exactly 6 main panels, exactly 5 planning questions in the list, exactly 3 bottom summary points, and keep all visible text in Japanese as specified. Do not add extra panels, extra checklist rows, watermarks, logos, or unrelated text.