GPT Image 2Prompts

Goal: Create a vertical cyberpunk anime fighting game character select screen for a fictional game titled {argument name="game logo text" default="SynClub"}, styled like a glossy arcade versus menu. Canvas: Tall 2:3 mobile-poster composition, 675×1200 style, dark futuristic neon city background split diagonally into blue left side and hot pink/red right side, with glowing HUD lines, scanlines, sparks, particles, and high-contrast rim lighting. Layout: At the very top, place angular metallic logo text “SynClub” with electric blue highlights, and beneath it the title “CHARACTER SELECT”. Put “PLAYER 1” in a blue sci-fi tab at the top-left and “PLAYER 2” in a pink sci-fi tab at the top-right. The screen is divided into two large featured character areas: left selected fighter for Player 1 in blue, right selected fighter for Player 2 in red. In the center lower half, place a character portrait selection grid with glowing frames. At the bottom center, place a large luminous “VS”. Along the bottom edge, add a game controller command bar with five options: select, confirm, back, random, option, each with small PlayStation-style button icons. Main characters: Player 1 is {argument name="player one character name" default="REGULUS"}, a dark hooded anime fighter with black cloak, chains, metallic accessories, glowing blue highlights, pale shadowed face, red eyes, one gloved hand lifting the hood, mysterious assassin/rogue mood. Add the Japanese reading “レグルス” under the name. Player 2 is {argument name="player two character name" default="MERU"}, a blonde gothic-lolita anime girl with long twin-tail hair, red and black frilled dress, rose accessories, black gloves, choker, heart gem, playful elegant pose with one hand near her face. Add the Japanese reading “める” under the name. Stats panels: Under each large character name, include a compact rectangular stat box with exactly 5 labeled rows: TYPE, POWER, SPEED, RANGE, TECHNIQUE. For Regulus, the TYPE value is “スピード” and the bars glow blue. For Meru, the TYPE value is “トリッキー” and the bars glow red. Character select grid: Include exactly 14 selectable tiles in the central grid, each with a glowing sci-fi border. Count and arrange them as follows: Row 1 has 3 tiles: hooded dark male portrait, blonde female portrait, dark masked/obscured portrait. Row 2 has 3 tiles: silver-haired maid girl, selected Regulus tile marked with a blue “P1” tag, selected Meru tile marked with a red “P2” tag. Row 3 has 5 tiles: robot/cyborg head, witch girl with hat, blond male portrait, black-haired masked fighter, tan spiky-haired boy. Row 4 has 3 tiles: turquoise aquatic/mermaid-like girl, pale white-haired character, and a black silhouette tile with a question mark labeled “Random”. Use blue selection glow around P1 and red selection glow around P2. Bottom versus panels: Bottom-left blue player panel reads “PLAYER 1 REGULUS レグルス” with an emblem and a weapon box labeled “MAIN WEAPON” showing {argument name="player one weapon" default="黒剣 -ノクターナル-"} beside a sleek dark sword illustration. Bottom-right red player panel reads “PLAYER 2 MERU める” with a rose emblem and a weapon box labeled “MAIN WEAPON” showing {argument name="player two weapon" default="ローズマイク"} beside a rose-themed microphone weapon illustration. Visual style: Ultra-detailed anime key art, glossy UI, esports fighting game interface, neon blue versus neon magenta palette, sharp angular frames, glowing typography, dramatic lighting, dense futuristic city details, polished game-menu screenshot look. Constraints: Keep all text legible and aligned like a real character select UI. Use exactly 2 large featured characters and exactly 14 central selectable tiles. Do not add extra characters outside the grid. No watermark.

Create a highly realistic documentary-style photograph of a clear plastic water bottle lying horizontally on its side in a dry desert, pouring water from its open mouth onto cracked sandy ground. The bottle should be on the right side of the frame, transparent with blue-tinted threads at the neck, ridged sides, visible highlights and refractions, and partially filled with water. A continuous stream of water spills onto the foreground, forming a shallow muddy puddle with ripples, wet sand, small stones, and bright sparkling sunlight reflections. The camera is very low to the ground with a wide-angle close-up perspective and shallow depth of field: sharp focus on the bottle opening, water stream, puddle ripples, pebbles, and gritty desert texture, while the distant background is softly blurred. The environment is an arid desert plain under intense midday sun, with sparse dry shrubs, distant reddish-brown mountains along the horizon, and a clear saturated blue sky. Place the bright sun high in the upper left corner with starburst lens flare and harsh natural shadows. Use photorealistic lighting, crisp detail, realistic water physics, heat-dry atmosphere, no people, no labels or text on the bottle, no watermark.

Create a cinematic photorealistic wide-angle scene of a fortified desert military communications outpost at golden hour, inspired by a modern tactical video game environment. The camera is low to the ground and slightly Dutch-tilted, looking up a dusty dirt road toward a concrete checkpoint compound with a central watchtower/control building, an open dark garage door, stacked concrete barriers, sand-filled blast walls, chain-link fencing, metal catwalks, pipes, antennas, and a large parabolic satellite dish on the roof to the right. Add exactly 2 visible text markings: a concrete barrier in the lower left reading “{argument name="barrier text" default="G.T.I HAWK"}” and a vertical banner on the tower reading “{argument name="banner text" default="G.T.I HAWK"}” with a simple hawk-like emblem. Surround the base with arid desert details: rocky cliffs on the left, several palm trees, scrub grass, dry weeds, scattered crates and utility boxes, worn road tracks, and tan stone walls framing the foreground. Use warm orange sunset light, dramatic partly cloudy sky, long shadows, dusty atmospheric haze, realistic weathered concrete and metal textures, high dynamic range, sharp environmental detail, no people, no vehicles, no UI, no watermark. Overall mood: tense, abandoned, tactical, sun-baked desert base in {argument name="setting" default="Middle Eastern desert"}.

Create a cinematic photorealistic wide shot of a futuristic military command compound in a desert canyon at golden hour. The scene shows a massive concrete-and-steel tactical base stretching diagonally from center-left to upper-right, with angular brutalist architecture, weathered tan concrete, dark metal trim, industrial vents, armored doors, railings, antenna masts, floodlights, and a tall control tower near the upper center. Add exactly 4 prominent text markings on the facility: a black sign reading {argument name="facility name" default="DELTA FORCE"} with smaller text "TACTICAL COMMAND CENTER" beneath it, a large blue wall panel labeled {argument name="sector label" default="A2"} with small "SECTOR" text below, a large orange door number {argument name="door number" default="12"}, and a small triangular warning-style emblem on the control tower. In the foreground, place one large concrete road barrier spanning the lower frame, painted with diagonal black-and-yellow hazard stripes and showing grime, chips, and two metal lifting loops. Behind it, include an empty paved road curving along the base, low security fencing, bollards, stacked crates, and industrial yard details. The background has rocky desert cliffs on the left, a few sparse palm trees, and a warm orange sunset sky with the sun low near the left horizon and soft clouds. Use a slightly tilted dynamic camera angle, 16:9 aspect ratio, low-to-mid viewpoint, high-detail game cinematic realism, dramatic warm rim light, long shadows, crisp textures, realistic dust and weathering, no people, no vehicles, no watermark.

Create a cinematic ultra-realistic wide-angle concept art scene of a futuristic concrete military checkpoint and watchtower in a remote rocky landscape. The main subject is a massive brutalist guard tower labeled {argument name="tower number" default="01"}, centered slightly right, with sloped reinforced concrete supports, weathered beige-gray surfaces, dark metal braces, antennae on the roof, and a glass observation deck wrapping around the top. Surround it with a fortified perimeter wall, a large industrial gate, railings, a raised pedestrian walkway on the left, stacked cargo crates and sandbag barriers near the base, and a sloped vehicle ramp on the right leading into the compound. The foreground is muddy dirt with shallow puddles reflecting the sky, tire marks, gravel, and worn concrete edges. Add a distant rocky mountain on the far right horizon and a deep blue sky with scattered soft clouds. Lighting should be late-afternoon golden sunlight from the left, creating long shadows, realistic highlights on glass, and atmospheric haze. Style: high-end AAA game environment art, photorealistic, detailed hard-surface architecture, no people visible, no vehicles, no extra text except the tower number, 16:9 widescreen composition, low camera angle looking upward for scale, sharp focus, natural color grading.

Create a wide cinematic sci-fi environment concept art scene of a massive concrete-and-metal military communications base on a coastal mountain plateau. The camera is low and slightly tilted, using a 16:9 widescreen composition, looking across an empty industrial landing pad toward angular brutalist buildings. The main structure sits right of center: a fortified bunker-like complex with sloped concrete walls, armored panel seams, a tall central doorway, narrow dark horizontal windows, industrial stairs and ramps, and a large painted number {argument name="building number" default="03"} on the right wall. On the roof, place a tall lattice communications tower crowded with antennas, satellite dishes, vertical masts, cables, and signal equipment. Add a smaller outbuilding to the left with its own antennas, plus a raised bridge or catwalk extending from the right side of the base. In the foreground include exactly 7 discrete industrial objects: 3 rectangular cargo crates on the far left, 2 low concrete barricades with yellow-and-black hazard stripes in the center, and 2 greenish metal equipment cases on the lower right. The setting is a remote coastline with a calm ocean and hazy snow-capped mountains visible on the left horizon. Lighting is golden-hour sunrise or sunset from the far left, creating warm rim light, long shadows, lens haze, and a bright washed-out glow near the horizon. Use a realistic AAA video game / cinematic concept art style, highly detailed hard-surface design, weathered beige-gray concrete, metal grime, subtle panel lines, bolts, vents, railings, and no people or vehicles. Keep the sky deep blue overhead with sparse wispy clouds, high dynamic range, sharp architectural detail, atmospheric depth, and an epic near-future military outpost mood.

Goal: Create a black-and-white manga-style two-panel comic about a cute guilty dog and a broken desktop computer, using expressive screentone shading, clean ink outlines, dramatic reaction marks, and Japanese speech bubbles. Canvas: Vertical comic page, approximately 2:3 aspect ratio, with a thick black outer border and one horizontal divider splitting the page into exactly 2 stacked panels. Panel 1: Show a home office desk scene in grayscale. On the left side of the desk are exactly 4 visible computer-related items: a cracked smoking monitor with a shattered screen, an open desktop PC tower with exposed internal parts, a keyboard in the foreground, and a wired mouse near the center. Add scattered debris, broken fragments, loose cables, small sparks, and dark smoke rising from both the cracked monitor and the damaged tower. In the center foreground sits a small cute puppy, a round-faced mixed-breed dog with floppy ears, large glossy eyes, a white muzzle and chest, darker patches on the ears and body, a collar with a small round tag, and a terrified guilty expression with sweat drops. In the right background, a young woman enters through a doorway, surprised; her face is intentionally obscured by a soft square blur. She has long wavy hair, wears a plain long-sleeve top, slim jeans, and slippers. Add a bookshelf, framed picture, potted plant, chair edge, doorframe, and simple room details in the background. Include exactly 1 speech bubble in this panel, coming from the woman, with vertical Japanese text: {argument name="first speech bubble text" default="あ、あ…"}. Panel 2: Show the same puppy now smiling brightly and posing innocently on the desk, turned slightly backward with one paw reaching behind its body as if hiding or covering the damaged computer tower. The dog should look adorable, playful, and suspiciously pleased, with sparkling effects beside its face. The broken desktop tower is behind the dog on the right, still smoking, with exposed components and a dangling cable. The desk includes exactly 3 visible desk items: keyboard, wired mouse, and a mug holding pencils. Background includes a shelf with books, a framed wall picture, a potted plant, and soft indoor room shading. Include exactly 1 speech bubble in this panel, on the left, with bold vertical Japanese text: {argument name="second speech bubble text" default="何でもない！"}. Subject details: The dog is {argument name="dog breed or type" default="a cute small mixed-breed puppy"}, with {argument name="dog markings" default="white fur with gray patches, floppy ears, and a white blaze down the face"}. The woman is {argument name="woman appearance" default="a young woman with long wavy hair, casual sweater, jeans, slippers, and a blurred face"}. Visual style: Monochrome Japanese manga illustration, soft pencil-like texture, detailed linework, expressive character acting, dramatic smoke and sparks, cute comedic tone, high contrast but not overly dark, realistic home office perspective mixed with exaggerated manga emotions. Constraints: Use exactly 2 panels, exactly 2 speech bubbles total, exactly 1 dog, exactly 1 woman, and exactly 1 broken desktop computer setup. Keep all text in Japanese as specified, avoid extra captions, avoid color, avoid watermark or logo.

Goal: Create a black-and-white manga comic page about a dog frantically programming on a desktop computer, then causing a system error. Canvas: Vertical comic page in grayscale, traditional manga ink-and-screentone style, with heavy panel borders and detailed hatching. Use exactly 2 horizontal panels stacked vertically, separated by a thick white gutter. Layout: Panel 1 occupies the top half. Panel 2 occupies the bottom half. Both panels show the same home office desk from a slightly angled side view: a large desktop monitor on the left, keyboard in the foreground, mouse and mug on the desk, a plant and framed wall art in the background. Panel 1: A fluffy medium-sized dog with floppy ears, white muzzle and chest, darker ears and patches, sits at the desk and types furiously with both front paws on the keyboard. The dog’s head area is largely covered by a plain vertical gray rectangle censor block in the center of the panel, obscuring most of the face while leaving ears, paws, body, and tail visible. The monitor displays mixed programming code and gibberish symbols, including C-like code, JavaScript-like lines, Python-like lines, binary digits, and random characters. Add dramatic manga speed lines radiating from the monitor and typing action. Include exactly 7 visible Japanese typing sound effects reading 「カタ」: one near the upper left of the dog/monitor area, one near the top center-right, one near the lower monitor/keyboard area, one near the left keyboard edge, and three clustered around the dog’s lower right body and tail. Add one vertical cloud-shaped thought bubble on the right with Japanese text: 「プログラマーだ！」. Panel 2: The same dog is startled and panicking, leaning back from the keyboard with wide eyes, open mouth, raised paws, a sweat drop on the cheek, and a small shock burst near the head. The monitor now shows a dark system error screen with a warning triangle icon and Japanese text. Use exactly this monitor text: 「システムエラー」, 「問題が発生したため、PCを再起動する必要があります。」, 「エラーコード: 0xC000021A」, 「再起動しています...」, plus a circular loading spinner made of small dots. Add one vertical speech bubble on the right with Japanese text: 「あ、壊れた…」. Visual style: High-detail monochrome manga, expressive animal character acting like a human, dense cross-hatching, screentone shading, strong motion lines, bold hand-lettered Japanese sound effects, slightly comedic tone, clean readable text on the monitor and bubbles. Customizable subject: The dog should be {argument name="animal character" default="a fluffy floppy-eared dog"}. The comic’s first thought bubble should say {argument name="first bubble text" default="プログラマーだ！"}. The second speech bubble should say {argument name="second bubble text" default="あ、壊れた…"}. The monitor error headline should be {argument name="error headline" default="システムエラー"}. The error code should be {argument name="error code" default="0xC000021A"}. Constraints: Keep the image black and white only, preserve the two-panel layout, use exactly 2 speech/thought bubbles total, exactly 7 typing sound effects in the top panel, no extra panels, no color, no watermark.

Goal: Create a black-and-white manga page with exactly 2 stacked panels showing a cute small dog discovering a computer keyboard and reacting with excitement. Canvas: Vertical manga page, portrait orientation, two equal horizontal panels separated by a thick black gutter, clean black border around each panel, grayscale screentone shading, hand-inked line art, polished modern manga style. Main character: A small, adorable puppy with floppy ears, large sparkling eyes, a white stripe down the face, darker patches on the head and body, a fluffy raised tail, and a collar with a round ring tag. The dog is seated at a desk in an office chair, facing a desktop computer. Use {argument name="dog breed style" default="cute mixed-breed puppy resembling a Jack Russell terrier"}. Panel 1: The dog sits behind a wooden desk, looking surprised and curious at a large desktop monitor on the left. A keyboard lies in front of the dog. The dog’s mouth is slightly open, eyes wide, with exactly 3 surprise marks above its head. Background: a cozy home office with a window and curtains on the far left, a framed landscape picture on the wall, a small potted plant on a cabinet, and bookshelves on the right. Include one vertical oval speech bubble on the right side containing the Japanese text {argument name="first speech bubble" default="これは何?"}. Mood: confused curiosity. Panel 2: Same dog, same desk and computer setup, now leaning forward happily and pressing the keyboard with one front paw. The dog smiles broadly with tongue visible, eyes shining, tail lifted. Add dynamic radial speed lines in the background and sparkling effects around the keyboard. Include exactly 8 sparkle effects around the dog and keyboard area. Include one vertical oval speech bubble on the left side containing the Japanese text {argument name="second speech bubble" default="ピカピカだ!"}. Mood: delighted discovery. Objects and counts: Exactly 2 manga panels; exactly 1 dog appearing in both panels; exactly 1 desktop monitor; exactly 1 keyboard; exactly 1 office chair; exactly 2 speech bubbles total; exactly 3 surprise marks in panel 1; exactly 8 sparkle effects in panel 2; visible background items include 1 window, 1 framed picture, 1 potted plant, and 2 bookshelves or bookcase areas. Visual style: Monochrome Japanese manga, expressive cute animal character, detailed fur texture, screentone gradients, crisp ink outlines, dramatic but family-friendly expressions, clean readable composition. Keep the speech bubbles vertical and the Japanese text hand-lettered in black. No color, no watermark, no extra panels, no extra characters.

Goal: Create a stylized cartoon version of a video meeting screenshot, like a comic-book Zoom call with exaggerated expressions and blurred private identities. Canvas: Square image with a dark rounded-corner video-call application window on a light background. Add three small gray window control dots at the top left. Inside the window, arrange exactly 5 participant video tiles: 4 tiles in a 2-by-2 grid on top, and 1 wider centered tile along the bottom. Use thin black gutters between tiles. The top-left tile has a bright green active-speaker border. Visual style: High-energy digital comic illustration, bold ink outlines, semi-realistic cartoon portraits, saturated colors, manga speed lines, sticker-like reaction icons, glossy highlights, and dramatic lighting. Faces must be covered by large rectangular blur blocks to preserve anonymity, while hair, poses, clothing, and backgrounds remain visible. Add small blurred name labels in the lower-left corner of most tiles, with one readable label: {argument name="visible participant name" default="Bear Liu"}. Layout and subject details: Include exactly 5 cartoon participants. Tile 1, top left: a blond wavy-haired person in a blue shirt, raising one hand in a rock gesture, orange energetic background, yellow waving-hand icon in the upper left, two white sparkle icons, green speaker outline, blurred name tag at bottom left. Tile 2, top right: a dark-haired person wearing earbuds and a dark jacket, surprised reaction, brown motion-blur hallway-like background, white speech bubble with “?!” on the left, three white stress marks above the head on both sides, lower-left name strip with a muted microphone icon and the readable name {argument name="visible participant name" default="Bear Liu"}. Tile 3, middle left: a seated person in a light gray hoodie in an office chair, hand on chin thinking, pale purple curtain background, thought bubble containing a yellow light bulb, blurred name label bottom left. Tile 4, middle right: a muscular person in a sleeveless black athletic shirt and beige Nike cap, tattooed arm visible, hand on chin, muted green studio background, thought bubble with a black scribble, vertical anxiety lines above the head, blurred name label bottom left. Tile 5, bottom center: a person in a black hoodie and black cap, writing with a white stylus on a tablet, purple burst background, large yellow comic text {argument name="comic slogan" default="LET'S GO!"} at left, two white sparkle icons, desk boxes and a black object in the right background, blurred name tag bottom left. Text and icons count: Show exactly 5 video tiles, exactly 5 face-blur rectangles, exactly 4 visible lower-left name labels, exactly 1 readable participant name label, exactly 1 large comic slogan, exactly 3 thought/speech bubbles/icons with symbols or text: waving hand, “?!”, light bulb, plus one scribble thought bubble. Keep all private names other than the one specified blurred or pixelated. Constraints: Do not show realistic unblurred faces. Do not add extra tiles, extra people, watermarks, captions outside the meeting window, or interface controls beyond the three top-left dots and tile name labels. Keep the overall look like a playful AI-generated cartoon avatar conversion of a video conference screenshot.

Create a cute retro cartoon cityscape of {argument name="city name" default="Tokyo"}, in a naive children's picture-book style with ultra-simplified architecture, chunky geometric shapes, exaggerated proportions, and bold thick black outlines. Use a vertical poster composition with a bright cyan-blue halftone sky and a worn beige paper border. Make {argument name="landmark" default="Tokyo Tower"} the clear central focus: a tall red-and-white lattice tower with a smiling kawaii face on its observation deck, rosy cheeks, simple black eyes, and a small open mouth. At the upper left, add large tilted comic lettering reading {argument name="headline text" default="TOKYO"} in red-orange with black outline, and beneath it place the Japanese text 「東京」 in chunky yellow letters with black outline. Include exactly 2 large cream-colored rounded clouds, one on the left and one on the right. Surround the tower with a simplified skyline of colorful rectangular buildings in blues, pinks, yellows, and oranges. Include exactly 5 visible signboards: a red vertical sign reading 「カラオケ」, a pink vertical sign reading 「原宿」, a yellow horizontal sign reading "SHIBUYA", a red vertical sign reading 「秋葉原」, and one billboard showing a pink-haired cute character portrait. Add one additional billboard with a cute white cat face next to the SHIBUYA sign. In the foreground, show a row of rounded green trees, a blue river or canal with small white ripples, and a cream-and-green cartoon train entering from the lower right on curved tracks, also with a smiling kawaii face and rosy cheeks. Apply a vintage comic-book print texture with visible halftone dots, slightly rough ink edges, saturated colors, playful proportions, and no photorealism.

Create a cute retro cartoon postcard cityscape of {argument name="city name" default="Tokyo"}, with the city’s most iconic landmark as the clear central focus: a kawaii Tokyo Tower in the middle, red and white lattice structure, simplified chunky geometry, thick black outlines, a smiling face with rosy cheeks on the observation deck, and exaggerated children’s picture-book proportions. Use a vertical 2:3 postcard composition with a turquoise-blue sky, cream paper border, visible vintage print texture, halftone dots, slight ink misregistration, and bold pop-art colors. At the top left, add large tilted red block letters spelling {argument name="headline text" default="TOKYO"} with black outline, and below it large yellow Japanese characters {argument name="local city text" default="東京"}. Surround the tower with an ultra-simplified skyline of colorful rectangular buildings in blues, pinks, oranges, and yellows. Include exactly 5 visible sign or billboard text elements in the city: a red vertical sign reading 「カラオケ」, a pink vertical sign reading 「原宿」, a yellow horizontal sign reading “SHIBUYA”, a blue billboard with a cute cat face, and a red vertical sign reading 「秋葉原」. Add exactly 2 cream-colored rounded cartoon clouds, one on the left and one on the right. In the foreground, create a band of bright green rounded treetops, a blue river or canal with small white water strokes, and a cute green-and-cream commuter train at the bottom right with a smiling face, black tracks, and rounded windows. Keep all architecture extremely simplified with minimal detail, chunky geometric shapes, naive children’s illustration style, bold thick black outlines, cheerful kawaii expressions, retro comic halftone shading, saturated pop palette, and no photorealism.

Please generate a high-quality home appliance commercial poster, with the main subject being "{argument name="主体" default="咖啡壶"}". With the subject as the visual core, automatically adapt its corresponding lifestyle scene, auxiliary items, decorative elements, keywords, selling point badges, feature icon descriptions, and overall color style. The layout is: large main title at the top + handwritten subtitle + keywords; the main product in the middle placed in an exquisite lifestyle scene that fits its usage logic; circular selling point badges and 4-5 feature icon descriptions added on the left and right; product name tag displayed at the bottom. The overall style requires a warm and healing vibe, soft warm lighting, realistic textures, modern clean typography, commercial advertising quality, low-saturation coordinated color scheme, 4K ultra-clear, suitable for brand promotion, e-commerce posters, and social media visual displays. All elements are automatically generated around the subject and serve the subject, without overshadowing it, being cluttered, or low quality.

Create a dramatic vertical anime-style fashion illustration of an elegant adult woman standing on a nighttime subway platform, with her face intentionally obscured by a soft rectangular gray blur for anonymity. She has {argument name="hair color" default="glossy black wavy bob hair with warm brown highlights"}, wears a {argument name="outfit" default="form-fitting ivory ruched midi dress, light beige trench coat draped open, pale wide-brim hat, black gloves, pearl earrings, and a delicate necklace"}, and carries one structured black handbag with a small gold chain charm in her right hand. The scene contains exactly one train on the left, arriving or passing with two bright round headlights and glowing destination text; exactly two background commuters on the right side, one suited figure partly cropped in the foreground and one smaller figure farther back; and exactly two illuminated overhead signs, a large red neon sign cropped at the top reading {argument name="top sign text" default="A Trai"} and a smaller orange advertisement sign beneath it with mostly unreadable stylized lettering and a circular portrait icon. Use a low-angle three-quarter composition emphasizing the woman's full-body silhouette and flowing coat, glossy tiled floor reflections, motion blur on the train, dark overhead cables, cinematic rim lighting, warm station lights against deep blue-black shadows, highly detailed manga linework, polished digital painting, urban noir atmosphere, sharp clothing folds, elegant high-fashion mood, portrait 9:16 aspect ratio.

Create a vertical 2:3 cinematic fantasy character-design poster for a guardian spirit called a Stand, blending ethereal Chinese typography with soft fashion photography and mythic creature design. The main subject is {argument name="character name" default="Higanbana · Wu"}, a young woman seated on a rough stone ledge beside a misty pond or stream, wearing a white ruched floral mini dress with thin straps and an oversized off-white knit cardigan slipping from one shoulder. Her hair is medium brown, loosely tied in a low messy bun with soft strands around the face; pose is elegant, quiet, slightly melancholic, hands resting on the ledge, legs angled forward. Hide or softly blur the face if needed, keeping the mood anonymous and poetic. Behind and above her right shoulder, create the Stand: a large translucent white dragon-deer / fox-dragon guardian spirit with an elongated delicate snout, flowing feather-like mane, branching antler or horn shapes, veil-like fins, silk ribbons, dangling gold chains and tassels, and many small white flowers blooming along its head and body. It should feel like fog, petals, and moonlight given animal form: beautiful, fragile, protective, slightly uncanny, not cartoonish. The spirit curves around the woman as if guarding her, occupying the upper-right half of the poster. Environment and lighting: a shaded waterside garden with drooping willow branches, dark green foliage, reflective water, floating mist, translucent gauze ribbons, tiny glowing petals or fireflies, and warm morning backlight filtering through haze. Use a palette of ivory, pearl white, champagne beige, pale blush, muted moss green, smoky gray, and soft gold. Overall atmosphere: serene, romantic, spectral, elegant, dreamlike, with a faint sense of grief and distance. Design the poster as an information-rich Stand profile sheet with refined white serif typography and thin decorative rules. Include exactly 6 text blocks: 1) upper-left small label in Chinese and English, “替身設定 / STAND NAME”; 2) large title “彼岸花・霧” with smaller romanization “Higanbana · Wu”; 3) a short poetic Chinese paragraph beneath the title; 4) mid-left section titled “特殊能力” with ability name “花影迷離” and romanization “Hua Ying Mi Li,” followed by a concise ability description; 5) bottom-left section titled “設計關聯” with several short bullet-like Chinese lines and exactly 7 circular color swatches in ivory, cream, warm beige, dusty pink, gray-green, deep moss, and taupe; 6) bottom-right quote box containing “「你看到的美好，也許只是我編織的夢。」” plus a smaller boxed profile headed “少女檔案” with brief attributes such as personality, image, aura, and preferences. Add one narrow vertical Chinese text panel on the right side over the spirit, framed by thin lines. Text should be decorative and readable enough to evoke a luxury editorial character sheet, without overwhelming the image. Visual style: high-end fantasy editorial poster, ultra-detailed, soft-focus photorealism mixed with delicate illustration, translucent layers, bokeh, shallow depth of field, fine ornamental linework, elegant Chinese book-design aesthetics, no harsh neon colors, no modern UI, no watermark. Use {argument name="main title" default="彼岸花・霧"} as the largest title, {argument name="ability name" default="花影迷離"} as the ability heading, and make the Stand feel like {argument name="guardian spirit motif" default="a white mist dragon-deer covered in flowers"}.

Goal: Create a dark cinematic anime-style “guardian spirit / Stand profile” poster for {argument name="character name" default="幻想郷の夜・绯"}, combining a realistic seated airline-uniform woman in the foreground with a huge supernatural winged female entity behind her. Canvas: Vertical poster, approximately 2:3 aspect ratio, dark navy/black palette, thin decorative border lines and corner ornaments, premium game-character dossier layout. Use dramatic rim lighting, red glowing accents, storm clouds, and a bright circular airplane-window-like opening on the right showing sunset sky and clouds. Main subjects: Foreground subject is one seated adult woman, face deliberately covered by a plain rectangular blur/censor block. She wears a pale blue short-sleeve flight-attendant shirt, dark navy skirt, red/navy neck scarf, black waist belt, wing-shaped name badge, and sits in an airplane seat with one visible armrest and a small control panel. Behind her is exactly one enormous guardian spirit: a dark elegant humanoid female monster with long black hair, fragmented black feather/metal wings, thorny tendrils, chains, red gemstone lights, black armor-like dress, pale face, and a dangerous calm expression. The spirit should feel protective, ominous, graceful, and predatory. Layout and text sections: Use a polished character-sheet composition with exactly 8 visible information zones: 1 top-left title block, 1 left-middle Stand setting text block, 1 center-right quote block, 1 lower-left ability evaluation radar chart, 1 lower-left design relation box, 1 lower-right special ability block, 1 bottom-right file/personality box, and 1 bottom color palette row. Visible typography: Mix elegant serif English labels with East Asian display text. Include the top-left label “STAND NAME”, the large title {argument name="stand name text" default="幻想郷の夜・绯"}, and the romanized subtitle “Gensokyo no Yoru · Fei”. Add Chinese dossier-style copy in the other sections. Include a quote on the right: {argument name="quote" default="「航線由我指引，目的地由我決定。」"}. Include the lower-right ability title {argument name="ability title" default="夜航領域"} with romanization “Ye Hang Ling Yu”. Include a bottom-right file box headed “檔案 / FILE”. Counted graphic elements: The radar chart must have exactly 6 labeled axes: destructive power, growth potential, precision, stamina, range, and speed. The color palette row must contain exactly 6 circular swatches: dark navy, light blue, white, crimson red, muted gold, and near-black brown. The guardian design should include multiple chain arcs but one dominant chain crossing the upper body, several red glowing nodes, and scattered black feather shards. Visual style: Ultra-detailed dark fantasy illustration, semi-realistic anime rendering, glossy metallic feather textures, smoky atmosphere, cinematic depth of field, high contrast, red ember highlights, occult fashion editorial mood, no cartoon simplification. Keep the foreground human realistic and the guardian spirit more painterly/anime. The poster should resemble a premium RPG/JoJo-like Stand ability dossier. Constraints: Do not add extra characters, do not leave the face visible, do not use bright cheerful colors, keep all text integrated into the poster design, and maintain the exact 8 information zones, 6 radar axes, and 6 palette swatches.

Goal: Create a dark gothic fantasy “stand / guardian spirit” character setting poster for {argument name="character name" default="Lin Wan Tang"}, combining a seated fashion portrait with an eerie humanoid guardian spirit named {argument name="stand name" default="Bad Apple!! Shadow"}. The design should feel like a luxury occult character dossier, emotional, mysterious, dangerous, and elegant. Canvas: Vertical poster, 4:5 ratio, dark blue-gray and black palette with warm beige typography, thin decorative border lines, subtle film grain, scratches, pearls, chains, sparkles, and faint mystical diagrams. Use a cinematic editorial composition with high detail and layered shadows. Main layout: On the left foreground, show the stand master: a young woman seated on a black leather chair, legs crossed, wearing a loose beige knit sweater with large black heart shapes, sheer dark stockings, and pointed black high heels. She has long wavy dark brown hair falling over her shoulders, a quiet melancholic pose, and an intentionally blurred or obscured face. Behind and to the right, towering over her, show the guardian spirit: a tall slender feminine humanoid doll-creature with a porcelain mask face, long wild black hair flowing like smoke, a black heart motif on the face and chest, pale stitched skin, black glossy heart-shaped bodice, skeletal mechanical arms, pearl and chain ornaments, translucent shredded black veils, and a thin rapier-like weapon held downward. The guardian’s silhouette should be both graceful and grotesque, like a cursed bridal puppet. Text and typography: Use elegant serif and small-caps typography mixed with delicate Chinese dossier text. Include large Chinese name text on the upper left, romanized name beneath it, and the label “STAND MASTER.” Include a quoted line in Chinese-style typography near the left side. On the upper right, place the heading for the stand name with the large text “Bad Apple!! • Shadow” and a short subtitle about the stand’s existence as a tangible embodiment of the silent heart realm. Along the bottom center, place the large spaced title “BAD APPLE!! • Shadow” and the subtitle “SILENT HEART REALM.” Keep text decorative and legible enough to suggest a character sheet, but do not overcrowd the poster. Information panels and counted elements: Include exactly 7 visible character profile rows in a small rectangular panel on the left: birthday, height, zodiac sign, temperament, traits, stand type, and stand attribute. Include exactly 6 small color palette swatches below that panel, in dark navy, charcoal, beige, pearl gray, muted brown, and steel blue. Include exactly 1 radar chart at the lower left with 6 labeled ability axes. Include exactly 3 rectangular ability-visualization thumbnails along the bottom left: a black heart core, a circular time-gear diagram, and a shadowy cocoon or veil form. Include exactly 4 circular detail callouts on the right side, each connected to the guardian by fine lines: porcelain face mask, black heart chest pattern, mechanical needle or rapier structure, and shadow silk ribbons. Include exactly 3 small heart-bullet ability descriptions in the lower right text block. Include exactly 1 ornate tarot-card-like inset illustration in the lower right showing a black heart trapped in a delicate frame of chains and thorns. Visual style: Ultra-detailed dark fantasy concept art blended with fashion editorial photography, gothic lolita occult aesthetics, Art Nouveau dossier framing, JoJo-like stand design energy, porcelain doll horror, black-heart symbolism, smoky hair and fabric motion, soft spotlighting from above left, cool shadows, high contrast, intricate linework, elegant poster typography. Constraints: The final image must look like a finished character setting sheet rather than a simple portrait. Maintain exactly one human stand master and exactly one guardian spirit. Keep the guardian much taller than the seated woman. Use the theme {argument name="symbolic theme" default="silent heart realm, rotten sweetness, dangerous tenderness"}. Use the primary palette {argument name="color palette" default="black, smoky navy, pearl gray, beige, and muted gold"}. Avoid bright colors, modern UI styling, extra characters, weapons besides the single rapier, and any readable watermark.

Goal: Create an atmospheric vertical “stand guardian spirit” character setting poster based on a gentle, misty feminine aura, combining a realistic fashion portrait with an ethereal mythical companion. The central concept is {argument name="stand name" default="彼岸花・霧"} with the romanized subtitle “Higanbana · Wu,” presented like a delicate fantasy character profile sheet. Canvas: Vertical 2:3 poster, cinematic editorial composition, soft gray-green palette, dreamy backlit mist, shallow depth of field, high-detail photorealism mixed with fantasy illustration. Background is a quiet riverside or pond garden with dark water, willow branches hanging from above, floating petals, haze, and glowing dust motes. Main subject: A young woman sits on a stone embankment in the lower center, elegant and calm, face intentionally obscured by a plain soft brown rectangular censor block. She has dark brown hair in a loose low bun with wispy strands. Outfit includes exactly 3 visible pieces: a white spaghetti-strap ruched mini dress with tiny pale floral print, an oversized off-white knitted cardigan slipped off the shoulders, and delicate thin shoulder straps. Pose is relaxed and graceful, one hand resting on the stone edge, legs angled forward, soft sunlight outlining her shoulders and hair. Guardian spirit: Behind and to the upper right of the woman, create one large translucent white dragon/deer-like floral spirit, long and serpentine with an elegant skull-like face, closed or lowered eyes, tall branch-like horns, feathered or veil-like fins, and flowing gauze ribbons. It should look protective, silent, and otherworldly rather than aggressive. Decorate it with exactly 5 main ornament types: white blossoms, pearl chains, dangling tassels, sheer silk veils, and fine silver branches. The spirit curves around the woman like mist, framing her without covering the text. Layout and typography: Design the poster as an information-dense character sheet with refined Chinese/Japanese editorial typography, thin white divider lines, small vertical annotation boxes, and elegant serif lettering. Use off-white text with subtle glow and high legibility. Include exactly 5 text zones: 1) top-left title block, 2) mid-left description block, 3) lower-left special ability block, 4) bottom-left design keywords and color palette, 5) right-side vertical lore panel plus bottom-right quote/profile card. Text content: Top-left title block shows “替身設定” with small English “STAND NAME,” then the large title {argument name="stand name" default="彼岸花・霧"}, and the subtitle “Higanbana · Wu.” Mid-left description reads in Chinese poetic lines about her guardian spirit being like mist and silk, circling her silently, elegant and watchful. Lower-left section title is “特殊能力,” ability name {argument name="ability name" default="花影迷離"}, romanization “Hua Ying Mi Li,” followed by short Chinese ability notes describing petals, gentle illusions, blurred distance, and disappearing footprints. Bottom-left section title is “設計關聯,” followed by exactly 5 bullet points: floral elements, soft palette, mist connection, emotional symbolism, and guardian/dream motifs. Beneath it place exactly 7 circular color swatches: white, ivory, pale cream, muted blush, cool gray, sage gray-green, and deep taupe. Right side has a narrow vertical Chinese text panel describing the spirit’s form and temperament. Bottom-right includes the quote {argument name="quote" default="你看到的美好，也許只是我編織的夢。"} and a profile card titled “少女檔案” listing exactly 5 traits: personality, image, aura, temperament, and preference. Visual style: Soft romantic dark-fantasy fashion editorial, ornate but restrained, like a luxury game character concept poster. Use luminous fog, translucent fabric, bokeh, film grain, fine paper texture, and subtle vignetting. The overall feeling should be tender, melancholic, mysterious, and protective. Constraints: Keep the woman’s face covered by a simple rectangle; do not show a recognizable face. Do not add extra characters. Maintain exactly one guardian spirit, exactly 7 color swatches, exactly 5 bullet points in the design section, exactly 5 profile traits, and exactly 5 text zones. Avoid logos, watermarks, modern UI icons, or cluttered unreadable text.