图片提示词

Create a first-person low-poly fantasy adventure game screenshot in a bright stylized 3D art style, as if from a playable RPG built from a single image. The camera is at human eye level on a dirt path beside a small stream, with two player hands visible in the foreground: the left hand is casting a glowing orange fireball with sparks and warm light, and the right hand holds a simple faceted steel sword with a dark hilt. Show exactly 3 enemies: 1 large green goblin charging in the lower left foreground with a spiked club, 1 smaller goblin warrior on the path near the center holding a round shield, and 1 bulky ogre or orc boss standing inside the ruined gate in the background. The setting is {argument name="environment" default="a mountain forest valley with ruined stone fortifications"}: mossy broken castle walls on the left, wooden palisades, pine trees, gray boulders, distant angular mountains, a blue sky, and chunky white low-poly clouds. Include exactly 2 red skull banners hanging on the ruin walls, exactly 3 lit torches, 1 wooden footbridge crossing the stream, 1 treasure chest near the riverbank, 1 wooden watchtower with a small red flag on a cliff, 1 blue waterfall on the right, 1 tall pale crystal obelisk in the right foreground, and 1 glowing purple dark crystal on a stone pedestal near the stream. Add a wooden quest sign on the right that reads exactly: "DESTROY THE\nDARK CRYSTAL\nIN THE RUINS" with a small skull icon beneath. Use vibrant game-ready lighting, crisp shadows, readable silhouettes, saturated greens and blues, warm fire highlights, polygonal geometry, no photorealism, no UI overlays, no watermark. The main quest objective is {argument name="quest objective" default="destroy the dark crystal in the ruins"}, the player weapon is {argument name="player weapon" default="a steel sword"}, the magic effect is {argument name="magic spell" default="an orange fireball"}, and the enemy faction is {argument name="enemy faction" default="goblins and orcs"}.

Create a cinematic photorealistic scene of an old rural Japanese train station platform at golden hour after rain, viewed from inside the covered platform hall looking outward toward the tracks. The architecture is weathered and nostalgic: dark wooden beams, peeling plaster walls, old metal columns, large gridded windows and doors, worn benches, timetable boards, posters, and glossy wet concrete floors reflecting warm sunset light. Include exactly 4 people: 1 uniformed station attendant on the far left in silhouette updating or reading a large paper timetable on the wall, 1 elderly man sitting alone on a wooden bench near the center-left with a black rolling suitcase beside him, 1 uniformed station staff member walking away through the middle corridor carrying a clipboard or folder, and 1 young traveler standing on the platform at right with a backpack facing an approaching modern train. Add a large round analog station clock high on the left wall showing about {argument name="clock time" default="9:45"}, a brightly lit beverage vending machine near the back-left, stacked newspapers or crates in the lower-left foreground, and a single modern train arriving on the right track with headlights on. The scene should feel quiet, melancholic, and realistic, with warm orange sunset streaming from the right, distant mountains beyond the tracks, visible power lines, long shadows, soft haze, detailed textures, natural depth of field, and no added captions or watermarks. Use a wide 3:2 horizontal composition, ultra-detailed documentary photography style, shot on {argument name="camera lens" default="35mm lens"}, with {argument name="lighting mood" default="warm cinematic sunset reflections"} and a nostalgic atmosphere in {argument name="location" default="an old Japanese train station"}.

Goal: Create a clean 16:9 academic overview slide explaining {argument name="paper concept" default="GeoPT: Scaling Physics Simulation via Lifted Geometric Pre-Training"}, styled like a polished conference-deck infographic.

Canvas: Widescreen white slide, navy-blue typography and accents, minimal shadows, high-resolution scientific illustration style. Use a thin blue footer line with small circular endpoints.

Header: Large bold dark-navy title at the top: {argument name="headline text" default="GeoPT: Scaling Physics Simulation via Lifted Geometric Pre-Training"}. Directly underneath, smaller blue subtitle: {argument name="subtitle text" default="An intuitive overview of Wu et al. (2026)"}.

Big idea banner: Under the subtitle, place a rounded rectangle callout with a blue outline and very pale blue fill. On the left is a blue circular lightbulb icon. Text reads: “Big idea:” in bold blue, followed by “use abundant 3D geometry and synthetic motion to pre-train neural simulators before using scarce real simulation data.”

Main layout: A three-column pipeline with exactly 4 rows, connected left-to-right by thick blue arrows. Column headings are: “3D Geometry”, “Physics Simulation (examples)”, and “Fields (pressure / velocity / stress)”. The left side also has 4 row labels: 1) Car, 2) Aircraft, 3) Ship hull, 4) Crash / Deformation.

Row details: Row 1 shows a gray 3D car model, then a car in airflow simulation with blue streamlines and colorful vortices plus a small “t →” time marker, then a car colored with a rainbow heatmap field and a vertical color legend labeled “Pressure” with High at top and Low at bottom. Row 2 shows a gray aircraft model, then an airplane with airflow streamlines and colored wakes plus “t →”, then a rainbow heatmap aircraft with a legend labeled “Velocity” High/Low. Row 3 shows a gray ship hull model, then a ship hull moving through ocean waves with simulated water flow plus “t →”, then a rainbow heatmap ship hull with a legend labeled “Pressure” High/Low. Row 4 shows a damaged gray car body for crash/deformation, then a crash simulation with debris, deformation, and internal colored stress/field visualization plus “t →”, then a rainbow heatmap deformed car body with a legend labeled “Stress” High/Low.

Bottom explanation strip: Include exactly 3 circular icon-and-caption blocks connected by dotted blue arrows. Block 1 icon is a wireframe cube and title “Abundant geometry”, caption “Easily available 3D shapes from CAD repositories”. Block 2 icon is wind/flow lines and title “Synthetic dynamics”, caption “Scalable simulations produce diverse motion and fields”. Block 3 icon is a neural network graph and title “Better physics simulators”, caption “Pre-trained models that generalize better with limited real data”.

Footer: Centered small gray citation text: {argument name="source citation" default="Source paper: Haixu Wu et al., arXiv:2602.20399 (2026)"}.

Visual style: Crisp vector-slide composition mixed with realistic 3D renders, blue academic palette, white background, precise alignment, no clutter, no watermark, no extra rows or icons beyond the specified counts.

Create a hyper-realistic vertical close-up selfie-style photo from inside a clear plastic cup, looking upward through a circular transparent rim. A young Japanese woman, {argument name="character age" default="early 20s"}, leans very close to the camera and sips through one thick teal-green straw centered in extreme perspective, running from the bottom foreground directly to her puckered lips. She has soft fair skin, large glossy brown eyes, natural makeup, rosy cheeks, long eyelashes, and reddish-brown hair in a loose side braid with wispy bangs and flyaway strands. She wears a delicate white lace blouse with ruffled shoulders and dangling pearl earrings. The cup foreground contains bright red strawberry pieces and crushed ice, heavily blurred with sparkling highlights and condensation; the cup rim forms a translucent circular frame around her face. Background: a sunny Japanese shopping street or festival lane with traditional wooden storefronts, blue sky, leafy trees, colorful hanging banners, and one visible banner with a strawberry motif, all softly out of focus. Lighting is bright midday sunlight with backlit hair glow, lens flare, glittering bokeh on the ice, cheerful summer atmosphere. Use an ultra-wide macro perspective, shallow depth of field, realistic skin texture, crisp focus on her eyes and lips, vibrant colors, vertical 9:16 composition. Include exactly one visible straw, one main subject, one cup rim circle, and multiple strawberry chunks in the foreground; avoid extra people in focus, avoid text overlays, watermark, or cartoon styling.

Goal: Create a cinematic anime key visual for a high-school romantic comedy titled {argument name="title text" default="負けヒロインが多すぎる!"}, set in a classroom with gentle daylight and a slightly melancholic but humorous mood.

Canvas: Wide 16:9 composition, 1200×675 style framing, like a polished TV anime promotional still. Use soft realistic lighting, detailed painted backgrounds, clean character line art, subtle depth of field, and no grain or image noise.

Layout: Three teenage students sit at desks in the foreground and midground. Count exactly 3 students: 1) left foreground main heroine, 2) center boy, 3) right background girl. The camera is at desk height, angled slightly upward, with large classroom windows behind them showing a bright blue sky, distant school buildings, balcony railings, and green hills. A concrete pillar and wall occupy the right side.

Subject details: The left foreground girl is the dominant subject: a cheerful teenage girl with {argument name="main heroine hair color" default="deep navy blue"} shoulder-length messy hair, blue eyes, flushed cheeks, and a teasing smile. She wears a white short-sleeve school shirt with a blue bow and sits lazily with her cheek resting on one hand, leaning toward the viewer. In front of her are exactly 3 desk items: a pink lunch box, a pink-and-white cloth, and a small yogurt drink carton with a straw labeled in Japanese. The center boy is a nervous teenage boy with short dark brown hair, wide uncertain eyes, a sweat drop, a white school shirt, and a green tie; he looks sideways toward the blue-haired girl as if thinking things are starting again. The right background girl is a withdrawn teenage girl with {argument name="side heroine hair color" default="dark reddish-purple"} short bobbed hair and hairpins, wearing a white school shirt; she sits apart near the wall, half-hidden behind a phone or small book, looking annoyed and pretending not to care. On the right foreground desk place a dark navy school bag.

Visible text and graphic elements: Include exactly 1 main title cluster in the upper left with the large Japanese title {argument name="title text" default="負けヒロインが多すぎる!"}, small romanization “MAKEINE,” and small English subtitle “Too Many Losing Heroines!” beneath it. Include exactly 2 wall posters on the right: one larger poster on the pillar reading {argument name="poster quote" default="負けても、きっと、青春は、終わらない。"} with the small English subtitle “Too Many Losing Heroines!”, and one smaller tilted poster in the upper right reading Japanese handwriting that means “Even so, this feeling is surely not losing.” Include exactly 3 handwritten white annotation captions with curved arrows or dotted trails: near the left girl, “やっぱ私は負けヒロインなんだよね〜”; near the boy, “また始まった…”; near the right girl, “別に気にしてない”.

Visual style: Modern Japanese anime film still, delicate shadows, cool blue classroom atmosphere, warm skin tones, expressive faces, natural hair highlights, detailed desks and windows, crisp but painterly background. The mood should feel like a promotional scene from a school comedy-drama about “losing heroines.”

Constraints: Keep the composition clean and cinematic, preserve exactly the counted characters, posters, desk items, and annotation captions, avoid extra characters, avoid extra random text, avoid watermark or logo beyond the specified title graphics.

{
  "prompt_structure": {
    "subject_description": {
      "demographics": "Young woman, pale complexion",
      "style_archetype": "Goth, E-girl, Alt-fashion, Y2K Grunge",
      "hair": "Long, dark black, messy layered waves, curtain bangs framing the face, voluminous",
      "face": {
        "eyewear": "Black thick-rimmed cat-eye glasses",
        "makeup": "Heavy black winged eyeliner, defined eyebrows, matte mauve/brown lipstick",
        "piercings": "Septum ring"
      }
    },
    "attire": {
      "top": "Black strapless tube top, shiny tight-fitting material (latex or spandex look)",
      "bottom": "Black faux-leather pleated mini skirt",
      "midriff_detail": "Cut-out strappy details (whale tail aesthetic) visible at the hips between the top and skirt",
      "accessories": [
        "Black leather choker with metal hardware and O-ring",
        "Silver chain necklace with a cross pendant",
        "Wide black belt with double rows of silver studded grommets",
        "Beaded bracelet on left wrist"
      ]
    },
    "specific_details": {
      "nails": "Extremely long, coffin-shaped acrylic nails, white tips with elaborate 3D nail art/charms",
      "phone_case": "iPhone with a large, deco-den style case featuring a 3D Hello Kitty figure and embellishments",
      "cables": "White phone charging cord hanging vertically across the torso"
    },
    "environment": {
      "setting": "Domestic bathroom interior",
      "background_elements": [
        "Beige/tan tiled walls",
        "White towels hanging on hooks (left)",
        "Corner of a white washing machine (bottom left)",
        "Chrome toilet paper holder and stand (right)",
        "Shower niche with blue and green bottles (background right)"
      ],
      "lighting": "Warm artificial overhead bathroom lighting, casting soft shadows"
    },
    "composition": {
      "type": "Mirror selfie",
      "framing": "Three-quarter body shot, waist-up focus",
      "pose": "Standing with hip slightly cocked, holding phone in right hand, left arm relaxed by side"
    }
  }
}

Goal: Create a widescreen anime promotional key visual for {argument name="anime title" default="Makeine: Too Many Losing Heroines!"}, set in a bright high-school classroom during daytime, with cinematic soft lighting and detailed hand-drawn animation style.

Canvas: 16:9 landscape composition, about 1170×660 px feel, with a shallow cinematic depth of field. The camera sits at desk height, looking slightly upward across a classroom table toward three students, with large windows and concrete pillars in the background. Outside the windows, show a hazy blue sky, distant buildings, and green hills.

Layout: Place the main blue-haired girl large in the left foreground, the boy in the center middle ground, and the red-haired girl on the right background. A wooden desk spans the entire bottom edge. Add the title logo in the upper left corner and promotional handwritten captions around the characters. Keep the right side darker and more shaded than the left.

Characters: Include exactly 3 students. 1) Left foreground girl: a cheerful teenage girl with long deep-blue hair, soft bangs, blue eyes, flushed cheeks, and a playful smug smile. She wears a white short-sleeved school shirt with a blue bow tie and small decorative charms at her chest. She rests her cheek on one hand, leaning casually over the desk and looking pleased. 2) Center boy: a teenage boy with short dark brown hair, pale skin, and a worried awkward expression, wearing a white short-sleeved school shirt and a dark green necktie. Add small sweat drops on his face and have him glance toward the blue-haired girl. 3) Right background girl: a small, withdrawn teenage girl with short reddish-purple bobbed hair and hairpins, seated apart near a gray pillar. She hides behind a pale phone or book, eyes narrowed as if pretending not to care.

Desk objects: Include exactly 4 visible desk items in the foreground: a pink lunch box or tray in front of the blue-haired girl, a crumpled pink-and-white cloth at the far left front, a small strawberry yogurt drink carton with a straw near the center, and a dark navy school bag in front of the red-haired girl.

Background and posters: Show exactly 2 wall posters on the right side. Poster 1 is vertical, pale gray, taped to a pillar behind the boy, with Japanese slogan text and small English subtitle “Too Many Losing Heroines!” near the bottom. Poster 2 is a darker angled poster in the upper right corner with white Japanese slogan text and a thin blue underline.

Visible text: Upper left title logo should combine Japanese title text with small English text: “MAKEINE / Too Many Losing Heroines!” Use blue and pink title colors. Add exactly 4 handwritten white caption callouts: one near the blue-haired girl on the left reading {argument name="left caption" default="やっぱ 私は負けヒロイン なんだよね〜"}, one near the boy reading {argument name="boy caption" default="また 始まった…"}, one near the red-haired girl reading {argument name="right caption" default="別に 気にしてない"}, and one at the upper right reading {argument name="upper right slogan" default="それでも、好きな気持ちは、きっと、負けじゃない。"}. Add subtle arrows or dotted guide lines pointing from captions to the characters.

Visual style: High-quality modern Japanese anime film still, delicate linework, soft cel shading, realistic classroom lighting, gentle bloom, blue-gray shadows, warm sunlight from the left windows, expressive faces, polished official key art look.

Constraints: Use exactly 3 characters, 4 desk items, 2 posters, and 4 handwritten caption callouts. Keep all text legible and placed like a promotional anime poster, with no extra characters, no watermark, and no photorealism.

Create a square-ish vertical children's book illustration of a cute cartoon {argument name="animal character" default="panda bear"} standing centered in a simple outdoor landscape. The character should be full body, front-facing, short and round with a large beige face, black circular ears, black eye patches, sleepy half-lidded eyes, small black eyebrows, a tiny black nose, and a small open mouth that makes the expression look tired, dazed, or just about to snap. Use a thick rough black outline around the body and facial features. The panda has a black torso and limbs with a large cream belly patch, stubby arms hanging at the sides, and small gray claws visible on both hands and both feet. The background is a bright cyan-blue sky with exactly two white blocky pixel-like clouds, one in the upper left and one in the upper right. The ground is flat orange-yellow dirt with scattered small rocks, tiny grass tufts, and rounded green bushes on both sides near the horizon. Style it like a hand-painted acrylic or gouache illustration on textured canvas, with visible brush strokes, slightly imperfect edges, bold flat colors, simple shapes, and a charming handmade children's storybook feel. No text, no watermark, no extra characters.

Create a highly detailed fantasy anime illustration of {argument name="character name" default="a young elven girl"} forming a magical contract with an ancient grimoire. The scene is set in a dark, cathedral-like library at night, with towering shadowy bookshelves, blue moonlight filtering through tall arched windows, tiny floating dust motes, and a mystical atmosphere. The character is shown from the waist up on the right side of the composition, with porcelain skin, long flowing {argument name="hair color" default="silvery blonde hair"}, large luminous blue eyes, delicate pointed elf ears, and an awestruck expression as she gazes upward toward the magic. She wears an ornate white and deep navy robe with intricate gold embroidery, filigree patterns, translucent sleeves, a blue gemstone pendant at her chest, jeweled hair ornaments, and small sapphire crystal accessories. In her hands is one large open ancient grimoire, dark blue-black leather cover, thick aged pages, gold metal corner fittings, chains, filigree, and a central round blue gem on the cover; two dangling blue crystal charms hang from chains along the lower left edge of the book. From the open pages bursts warm golden-white magical light, sparks, and vertical streams of glowing runes. Above the book, create one huge circular magic sigil made of concentric rings, arcane glyphs, and geometric sacred-star patterns, glowing in warm gold; include exactly four prominent jewel-like nodes around the circle plus a bright central star point. The sigil should look suspended in the air, semi-transparent, with unreadable fantasy runes rather than real language. Use dramatic contrast between cool blue moonlit background and warm golden spell light, cinematic depth of field, volumetric light beams, sparkling particles, ultra-detailed fantasy concept art, delicate anime-realistic rendering, 4:5 vertical composition, no watermark, no modern objects, no readable text.

Create a candid photorealistic vertical smartphone photo of a young Japanese woman, {argument name="character name" default="unnamed"}, resting outside a convenience store during the day. She sits on a low concrete curb at the edge of an asphalt parking lot, framed from head to sneakers with the top of her head near the upper left and her legs filling the lower half. She has shoulder-length slightly messy black hair with wispy bangs, natural makeup, and a quiet downcast expression as she looks at a smartphone held partly out of frame in her left hand. She wears exactly four visible outfit pieces: a pale cream open cardigan, a light blue ribbed camisole, a loose gray knee-length skirt, and white sneakers. Her pose is casual and unguarded, knees apart as she sits, holding exactly one plastic bottled cafe latte in her right hand; the label should read {argument name="drink label" default="カフェオレ"}. Beside her on the ground to the lower right are exactly two bags: one beige canvas tote bag and one crinkled white plastic shopping bag. In the background, show a beige convenience-store exterior wall, exactly two metallic utility/trash units, exactly one white standing promotional flag with red Japanese lettering, exactly one blue bicycle-parking sign, painted white parking lines, and rough curb texture. Use natural daylight, soft shadows, realistic skin texture, shallow documentary-style composition, slight handheld smartphone perspective, muted everyday colors, and no glamour posing. The mood should feel like an ordinary fleeting moment of taking a break in front of a convenience store. Avoid studio lighting, fantasy elements, extra people, exaggerated anatomy, or overly polished fashion photography.

Create a widescreen 16:9 anime key visual for a romantic school comedy titled {argument name="title text" default="負けヒロインが多すぎる!"}. Scene: three Japanese high school students sit at desks in a bright classroom during lunch, with large windows, concrete pillars, balcony railings, blue sky and distant buildings/mountains outside. Soft cinematic afternoon lighting, shallow depth of field, detailed modern TV-anime rendering, expressive faces, delicate line art, natural skin shading, slightly melancholic but comedic atmosphere.

Characters: exactly 3 students. 1) Foreground left: a cheerful blue-haired teenage girl with long messy navy-blue hair, blue eyes, flushed cheeks, wearing a white short-sleeve school shirt and blue bow tie; she leans on one hand with a smug, teasing smile, looking relaxed and confident. 2) Center: a black-haired teenage boy in a white short-sleeve school shirt with a green necktie, sitting behind her, looking sideways with a worried awkward expression and a small sweat drop. 3) Right: a short maroon-haired teenage girl with side hair clips, wearing a white school shirt, half-hidden behind a desk and holding a patterned book or phone near her face, pretending not to care while looking annoyed or embarrassed.

Foreground props: exactly 4 prominent desk items: a pink lunch box in front of the blue-haired girl, a crumpled pink-and-white cloth, a small yogurt drink carton with a straw and Japanese text, and a dark navy school bag in front of the maroon-haired girl.

Text layout: top left display the large Japanese title {argument name="title text" default="負けヒロインが多すぎる!"} in hand-painted blue and pink lettering, with small English subtitle beneath: "MAKEINE Too Many Losing Heroines!" Add exactly 5 handwritten white annotation captions around the characters: left beside the blue-haired girl: {argument name="left annotation" default="やっぱ私は負けヒロインなんだよね〜"}; near the boy: {argument name="boy annotation" default="また始まった…"}; near the right girl: {argument name="right annotation" default="別に気にしてない"}; on a poster in the back: {argument name="poster slogan" default="負けても、きっと、青春は、終わらない。"}; top right on the wall: "それでも、好きな気持ちは、きっと、負けじゃない。" Include small curved arrows and dotted leader marks pointing from some annotations to the characters.

Background details: include exactly 2 visible wall posters with Japanese inspirational copy and small English title lines reading "Too Many Losing Heroines!" Keep the composition like an anime promotional still: the blue-haired girl dominates the left foreground, the boy is centered behind her, the maroon-haired girl is smaller on the right, with desks forming a horizontal foreground line. Avoid extra characters, avoid photorealism, avoid distorted hands, and keep all text legible.

Goal: Create a dreamy vertical promotional key visual for an anime-style summer limited parfait figure of {argument name="character name" default="Nozomu"}, designed like a collectible figurine living inside a pool-themed dessert parfait.

Canvas: Vertical 3:4 poster, bright pastel resort atmosphere, high-resolution glossy product photography mixed with delicate anime illustration. Background is a softly blurred outdoor swimming pool at a luxury summer resort with turquoise water, white parasols, lounge chairs, palm greenery, blue sky, and glowing bokeh.

Main subject: A cute seated anime girl figurine centered slightly right, smiling and looking at the viewer. She has {argument name="hair color" default="pale pink hair with aqua-blue tips"}, sparkling blue-pink eyes, translucent glossy skin, and a joyful summer expression. Her pose is seated with both long legs forward, one hand raised near her face in a playful gesture. She wears a frilly pastel sailor-inspired swimsuit dress in white, pink, and aqua, with off-shoulder ruffle sleeves, a large pink chest bow, blue corset lacing, flower accents, translucent frills, ankle-strap platform sandals, hair clips, a small flower hair ornament, and shell/star accessories.

Parfait figure base: The character sits on a large transparent shaved-ice-style clear base shaped like a splashing pool parfait. Use glittering crystalline ice, translucent aqua and pink syrup gradients, star-shaped crystals, water-splash arcs, glossy resin texture, and a circular clear display stand at the bottom. The dessert base should feel like a collectible figure display, not a real food photograph.

Counted decorative elements to include: 1 miniature pastel parasol inserted into the parfait on the left, 1 metal dessert spoon angled upward on the right, 1 pink-and-clear inflatable swim ring at the lower right, 2 seashell decorations near the front right, 1 large pink hibiscus flower near the left cup area, 1 clear parfait glass/cup shape partially visible behind the legs, 1 small acrylic display plaque at the bottom right, and many tiny star crystals embedded in the ice but keep the foreground star crystals visibly distinct.

Text layout: Place large glowing Japanese title text at the upper left reading {argument name="main Japanese title" default="のぞむ"}. Under it, add the English subtitle “SUMMER PARFAIT FIGURE”. Below that, place a pink ribbon label reading {argument name="ribbon text" default="住み着きました"}. Under the ribbon, add the flavor line “＊ POOL PARFAIT Ver. ＊”. On the small bottom-right plaque, repeat four centered lines: “のぞむ”, “SUMMER PARFAIT FIGURE”, “住み着きました”, and “POOL PARFAIT Ver.”. Use elegant serif lettering for English, soft rounded glowing lettering for Japanese, white and pastel blue text with a magical glow.

Visual style: Ultra-cute Japanese anime merchandise illustration, pastel blue and pink palette, transparent resin, glass, shaved ice, sparkles, lens bloom, soft sunlight, glossy highlights, delicate details, magical summer collectible figure advertisement. Add small doodle stars and sparkles around the title. The overall mood is refreshing, sweet, luxurious, and limited-edition.

Constraints: Keep the figure fully visible from head to sandals, preserve the exact poster-like composition with text on the upper left and plaque on the lower right, avoid dark colors, avoid horror or realism, no extra characters, no watermark, no messy unreadable typography.

Change her facial expressions, various angles of close up shots

Create a photorealistic lifestyle portrait of {argument name="subject" default="a stylish adult East Asian woman"} sitting on a soft bed or window-seat beside a tall apartment window, working on a silver laptop with a visible white Apple-style logo. She has a short wavy {argument name="hair color" default="dark brown"} bob, black rectangular eyeglasses, large black over-ear headphones, natural makeup with soft red lipstick, and a calm confident expression while looking toward the camera. Her left hand lightly adjusts the glasses near her temple and her right hand rests on the laptop keyboard. She wears an oversized white deep V-neck long-sleeve lounge shirt with relaxed folds and ribbed cuffs, styled as cozy indoor workwear. Composition is a vertical 3:4 medium close portrait from upper thighs to head, subject centered slightly right, laptop in the lower foreground, window frame on the left, blurred city buildings outside, and a warm minimalist bedroom background on the right with one small potted plant and one softly glowing candle or lamp. Use soft natural daylight from the window mixed with warm ambient interior light, shallow depth of field, realistic skin texture, cinematic muted tones, gentle shadows, high-detail fabric wrinkles, no text overlays, no watermark, no extra people.

Create a wide cinematic fantasy anime illustration of exactly six beautiful nymph maidens resting beside a clear turquoise spring in an enchanted garden. The scene is sunlit, lush, and highly detailed, with dappled light filtering through huge ancient trees, colorful wildflowers in the foreground, mossy stones, a golden ornate gazebo with a blue domed roof in the midground, cascading waterfalls on the right, and a distant fairytale castle with blue spires rising above the forest. Use a polished 3D anime figure style with soft skin, large expressive eyes, delicate faces, long silky hair, and elegant translucent Greek-inspired chiffon dresses with gold trim. The six nymphs are: 1) a foreground blonde nymph on the far left, seated close to the viewer with one knee raised, wearing a white draped dress with gold straps; 2) a lavender-haired nymph seated under the gazebo area, wearing a light purple dress, one leg crossed over the other; 3) a blue-haired nymph seated on a rock near the center, wearing a pale blue dress and looking forward calmly; 4) a pink-haired nymph standing near the water, wearing a flowing pink dress and lifting one hand in a gentle greeting; 5) a green-haired nymph kneeling by the spring, wearing a pale green dress; 6) a silver-white-haired nymph seated on the far right among flowers, wearing a lavender-white dress and touching her chest lightly. The mood should be serene, magical, and idyllic, like {argument name="scene title" default="Nymphs Resting by the Spring No. 3"}. Emphasize {argument name="art style" default="highly polished fantasy anime 3D illustration"}, {argument name="lighting" default="warm golden morning sunlight with sparkling highlights"}, {argument name="setting" default="enchanted forest spring with waterfalls, gazebo, and distant castle"}, and {argument name="color palette" default="pastel dresses, emerald foliage, turquoise water, and gold architecture"}. Full-body composition where possible, graceful poses, refined anatomy, bare feet, no text, no watermark, no modern objects, panoramic 16:9 aspect ratio.

Create a polished anime-style vertical portrait illustration of a cute kemonomimi maid girl leaning toward the viewer in a bright Victorian room. The subject is a young cat-eared girl with {argument name="hair color" default="silvery ash gray"} twin-tail hair, soft curled side locks tied with small black ribbons, large fluffy cat ears with pale fur and pink inner ears, and big {argument name="eye color" default="pinkish purple"} sparkling eyes. She has a small gentle smile, rosy cheeks, a soft childlike face, and glossy highlights on her hair and eyes. She wears a classic black-and-white maid outfit: black long-sleeved dress, white rounded collar, black bow at the neck, white frilled apron, puffed white shoulder frills, white cuffs, and a large white ruffled maid headband. Pose her in a close-up three-quarter view, upper body filling most of the frame, leaning forward with her head slightly tilted and one hand extended toward the camera in a friendly wave with five visible fingers. Place her in front of tall sunlit window panels with cream curtains, carved dark wooden trim, faint patterned wallpaper, and soft greenery visible outside. Use warm morning backlight streaming through the windows, delicate rim light on the ears and hair, bright high-key highlights, soft shadows, shallow depth of field, and a clean highly detailed Japanese anime illustration style with smooth linework, vibrant pastel lighting, and no text or watermark. Composition should be vertical 3:4, intimate and cute, with the character centered and cropped around the waist.

Goal: Create a vertical Japanese manga page about a cute girl regretting an overly short haircut after visiting a hair salon, in a soft detailed anime style.

Canvas: Portrait manga page, approximately 4:5 aspect ratio, with thin black gutters and clean white borders. Use exactly 6 panels: top row has 2 equal panels, middle has 2 full-width horizontal panels stacked, bottom row has 2 equal panels.

Character details: Main character is {argument name="character name" default="a young girl"} with large teal-green eyes, flushed cheeks, and {argument name="hair color" default="light chestnut brown"} hair cut into a very short tousled bob/pixie cut with soft bangs and a tiny ahoge curl. She wears a white frilly blouse, a large black-and-white striped ribbon bow at the collar, brown suspender skirt/overalls, and a small brown shoulder bag in the outdoor panel. Her expressions move from shock to embarrassment, sadness, forced composure, and regret.

Setting and visual style: A cozy bright hair salon with warm wooden furniture, round mirror, salon chair, shelves of bottles, plants, soft daylight, and delicate watercolor-like anime rendering. Use pastel colors, thin clean line art, expressive shoujo manga faces, gentle highlights in the hair and eyes, and subtle screentone/shading. Speech and thought bubbles are white with black outlines and Japanese text.

Panel-by-panel layout and content:
1. Top-left panel: Interior salon mirror scene. Show the girl from behind sitting in a salon chair, looking into a round mirror. In the mirror her front-facing reflection is shocked with wide eyes and open mouth after seeing the very short haircut. Add large blue punctuation marks “!?” near the left side and a spiky surprise symbol above the reflection.
2. Top-right panel: Close-up front portrait against a dark bluish vertical emotional background. She looks teary and stunned, cheeks pink, lips trembling. Add one thought bubble with {argument name="short haircut thought" default="み、短すぎる…"} and small vertical trembling text “うる…”.
3. Middle first full-width panel: Side view in the salon. The girl sits looking uncertain while an adult female hairdresser with brown hair in a neat low bun smiles and claps encouragingly. The hairdresser wears a white shirt with dark suspenders/apron. Add sparkle icons and a speech bubble saying {argument name="hairdresser compliment" default="お似合いですよ！"}.
4. Middle second full-width panel: The girl alone in the salon, centered, looking embarrassed and about to cry, with the round mirror and salon shelves in the background. Add a thought bubble saying {argument name="embarrassed thought" default="こんなはずじゃ…"}.
5. Bottom-left panel: Exterior street scene outside the salon entrance with plants and warm storefront details. The girl stands with her shoulder bag, touching her short hair with one hand, eyes downcast and uncomfortable. Include a chalkboard sign that reads exactly: “Ciel Hair / Open 10:00 / Close 19:00 / Thank you!” Add a thought bubble saying “鏡見るのつらいな…”.
6. Bottom-right panel: Dreamy regret panel with dark purple-blue background. The short-haired girl in the foreground touches her head and looks worried. Behind her, show a glowing imagined version of herself with long twin tails tied with black ribbons, smiling confidently. In the lower right, show exactly 2 loose black hair bows floating/fallen as symbolic objects. Add sparkles around the imagined long-haired version and a thought bubble saying {argument name="final regret thought" default="やっぱり切りすぎたかも…"}.

Constraints: Use exactly 6 panels, exactly one recurring main girl, one hairdresser, one salon sign, and 2 loose black bows in the final panel. Keep all visible dialogue in Japanese as specified. Do not add extra panels, extra characters, watermarks, logos, or narration boxes.

Goal: Create a black-and-white Japanese manga-style educational comic page about sunscreen, steroid ointment, ultraviolet rays, and skin inflammation, with only a few teal accent fills. The comic should look like a printed newspaper health explainer with bold panel borders, screentone shading, halftone gradients, hand-drawn ink lines, vertical Japanese dialogue, and slightly crowded text layout.

Canvas: Square 1:1 page, high-contrast monochrome manga art, thick black outer border and thick gutters. Use teal accents only on clothing, umbrella, and sunscreen packaging. Keep the overall look like a generated manga infographic with dense vertical text.

Layout: Use exactly 4 panels arranged as follows: two equal top panels side by side, one wide middle panel spanning the full width, and one wide bottom panel spanning the full width.

Panel 1, top left: A cheerful young woman doctor or educator with shoulder-length dark hair, wearing a white coat over a teal shirt, raises one index finger as if correcting a misconception. Radiating emphasis lines behind her. Include two vertical speech bubbles. Text: 「それも誤解です！」 and 「紫外線も皮膚への刺激の１つなので」.

Panel 2, top right: A young girl with a short dark bob and teal sweatshirt wipes or shields her forehead, looking worried under strong sun. Add a small handwritten sound effect 「まぶしい〜」. Place a vertical title box on the right edge reading 「おまけのこぼれ話」. Include two vertical speech bubbles. Text: 「そう言えば、ステロイドは紫外線ダメだから」 and 「朝は塗らないんですよね？」.

Panel 3, middle wide: A dramatic black-and-white wildfire or mountain fire scene with wind blowing flames across a forested slope; include smoke, trees, a winding road, and a spiral wind symbol. Overlay three vertical narration text boxes. Text box 1 on the right: 「山火事に追い風が吹くように」. Text box 2 center-right: 「炎症のある皮膚に紫外線が当たると余計ダメージが大きくなりますが」. Text box 3 left: 「それとステロイド自体はあまり関係がありません！」 plus smaller side note 「（ステロイド外用は１日２回が基本です。）」.

Panel 4, bottom wide: The same girl now happily uses sun protection: she holds a teal parasol above herself, wears a light hat, and has sunscreen bottles nearby labeled “UV” and “SPF 50 PA+”. Add handwritten labels near objects: 「日焼け止め」 near the sunscreen, 「日傘」 and 「帽子」 near the umbrella and hat. Include three vertical narration blocks: right block 「おそらくワセリンを塗ると日焼けするというイメージから来るのでしょうが そもそもそれ自体は明確な論拠はありません。」, center-left block 「紫外線対策はもちろん大切なので、」, far-left block 「皮膚に炎症がある時は、適切に紫外線対策をしましょうね！」. Add a small handwritten note at lower right: 「塗ってみて刺激のないものを使ってネ!!」.

Visual style: Japanese manga health-education comic, rough ink texture, expressive but simple characters, screentone dots, grayscale printing, dense vertical lettering, slightly imperfect hand-drawn borders, clean white gutters, teal accent color around #68cfc6.

Constraints: Include exactly 4 panels and exactly the listed visible text elements. Do not add extra characters beyond the educator and the girl. Do not use full color beyond the teal accents. Make the composition feel information-dense and slightly cramped, like an AI-generated medical explainer comic. Main topic can be customized as {argument name="comic topic" default="misconceptions about steroid ointment, ultraviolet rays, and sunscreen"}; educator accent clothing color {argument name="accent color" default="teal"}; top-right title label {argument name="side title" default="おまけのこぼれ話"}; sunscreen label {argument name="sunscreen label" default="SPF 50 PA+"}; handwritten reminder {argument name="reminder note" default="塗ってみて刺激のないものを使ってネ!!"}.

Goal: Create a dramatic black-and-white manga page in a gritty seinen style about two anime girls confronting each other over borrowed money, with intense stormy atmosphere, debris, ink splatter, and cinematic low-angle composition.

Canvas: Tall vertical manga page, approximately 9:16 aspect ratio, monochrome grayscale screentone printing, high contrast blacks, detailed line art, speed-line impact effects, no color.

Layout: Use exactly 2 stacked panels separated by a diagonal white gutter. The top panel occupies about 45% of the page and shows a hooded girl standing on the edge of a cracked rocky cliff, viewed from below. The bottom panel occupies about 55% and shows the second girl reclining or sitting against broken ground and rubble, viewed close-up from a low frontal angle.

Top panel subject: A slim teenage anime girl named {argument name="creditor name" default="Nagi"} stands ominously above the viewer on a jagged rock ledge. She wears a light-colored hooded raincoat or parka with large front pockets, the hood up, short light hair visible under the hood, one eye partly obscured, and a calm but threatening expression. Wind and clouds swirl behind her, with dust and small rocks floating in the air. Add two large black jagged manga sound-effect strokes near the left and lower-right edges.

Bottom panel subject: A slim teenage anime girl named {argument name="debtor name" default="Sora"} sits below among cracked stone rubble, injured and dirty but smirking defiantly. She has very long dark hair blowing across the frame, sharp eyes, scratches and mud on her face and legs, and wears a dark sporty oversized jacket with visible small text “SORA” on the chest, striped sleeve details, layered streetwear, harness-like straps, shorts, belts, and zippers. Her pose should feel relaxed and insolent despite the battle damage.

Text content: Include exactly 2 vertical Japanese speech bubbles. In the top panel, place one tall white speech bubble on the right containing: 「ソラさん—— “お金”返してください」. In the bottom panel, place one white speech bubble on the left containing: 「——ヤダね」. Use clean manga lettering and keep the bubbles vertical.

Visual style: Hyper-detailed monochrome manga illustration, dramatic lighting, textured screentone sky, cracked rock rendering, dust particles, ink splashes, heavy blacks, sharp highlights, emotional tension, catfight showdown mood, professional serialized manga page quality.

Constraints: Include exactly 2 characters, exactly 2 panels, exactly 2 speech bubbles, and exactly 2 large jagged sound-effect marks. Do not add extra panels, extra characters, color, watermarks, captions, or English dialogue.

Goal: Create a retro mid-century magazine advertisement poster for {argument name="brand name" default="ima"}, with a warm cream paper background, aged print texture, subtle halftone grain, and a nostalgic 1950s editorial illustration style.

Canvas: Horizontal 3:2 composition, generous off-white margins, balanced like an old magazine spread. Left half is typography-heavy; right half is an illustrated woman at a desk.

Layout: On the left, place a large black Chinese headline in elegant serif-style type, arranged in four stacked lines, with the brand word in red: {argument name="headline text" default="仅靠刷手机，我用 ima 搭了个自动选题库"}. Under it, add a small red dot and thin gray horizontal divider, then a smaller black subtitle: {argument name="subtitle text" default="素材丢进去，选题自己跑出来"}. In the bottom-left corner, add a tiny footer reading “ima × Ideas come out.” in muted gray.

Main illustration: On the right, draw a fashionable retro woman with pale skin, short black curled hair, red headband, red short-sleeve top, red lipstick, pearl earrings, and a wristwatch. She rests her chin on one hand while holding a smartphone in the other, looking calmly at the phone. In front of her is an open silver laptop with the brand logo “ima” centered on the back. On the desk include exactly 3 desk items: the laptop, a small notebook with a pen, and a yellow teacup on a saucer.

Floating idea stream: From behind the laptop and woman, create a swirling white paper-ribbon trail curling upward to the top-right, suggesting ideas automatically emerging. Attach exactly 8 small rectangular topic cards to the swirl, each with Chinese labels: 1) {argument name="topic card one" default="情感共鸣"}, 2) “反常识”, 3) “情绪共鸣”, 4) “热点延禧”, 5) “观点递增”, 6) “行业洞察”, 7) “行业洞察”, 8) “长期价值”. Use varied muted paper colors: beige, pale yellow, light green-gray, orange, and cream. Also include exactly 2 small blank note papers and a few tiny gold sparkle stars around the floating cards.

Visual style: Vintage Chinese magazine cover aesthetic, hand-drawn commercial illustration, ink outlines, soft watercolor-like shading, muted reds and yellows, cream background, printed paper grain, slightly faded colors, elegant yet practical productivity-ad feeling.

Constraints: Keep all visible text crisp and readable, preserve the exact number of floating labeled cards and desk items, avoid photorealism, avoid modern glossy UI elements, no extra characters, no watermark.

Create a cinematic dark-fantasy action still of {argument name="character name" default="Angela the Amazon"}, a fierce Amazon warrior woman crouching in a mossy ancient forest as if stalking an enemy. She has long, voluminous, wind-swept {argument name="hair color" default="copper-red"} curly hair, intense eyes, subtle battle-smudged makeup, and an athletic muscular build with dramatic heroic proportions. She wears ornate bronze-gold fantasy armor: a winged engraved crown/helmet with cheek guards, large hoop earrings, sculpted metal breastplate cups over dark leather straps, circular metal rings and studs, chain wraps around one upper arm, heavy engraved bracers, and a leather belt with round bronze plates. Her pose is low and forward-leaning, one arm extended back and the other gripping a long spear or polearm crossing the lower foreground, with the spear tip visible at the left edge. Set the scene in a dim, misty, old-growth woodland with twisted tree trunks, thick branches, moss-covered fallen logs, and soft green-gray atmospheric haze. Use shallow depth of field with a sharply detailed warrior and blurred background, realistic skin texture, wet highlights on armor and skin, dramatic side lighting, muted forest greens contrasted with warm bronze metal and fiery hair. Composition should be a horizontal cinematic frame, medium close-up from the waist up, character slightly left of center facing toward the left, dynamic tension, photorealistic high-budget fantasy movie look, no text, no watermark.

Create a full-body dark fantasy pixel-art character portrait of {argument name="character name" default="Anubis"}, an imposing anthropomorphic jackal-headed Egyptian god-warrior standing against a pure black background. The figure is centered, front-facing in a slight three-quarter stance, tall and extremely muscular, with charcoal-gray stone-like skin, a black jackal muzzle, tall pointed ears, sharp claws, bare digitigrade feet, and glowing {argument name="eye color" default="golden amber"} eyes. He wears ornate ancient Egyptian armor and jewelry in tarnished gold, dark teal enamel, and weathered linen. Count and include exactly 12 main visible costume and accessory elements: 1 striped pharaoh-style nemes headdress with gold trim and teal panels, 2 leafy green laurel or vine strands draped over the crown, 3 broad gold-and-teal collar necklace, 4 large gold ankh pendant at the chest, 5 smaller dark ankh marking or pendant on the abdomen, 6 stacked gold bracelets on the weapon-holding wrist, 7 gold-and-teal upper arm band, 8 ornate belt with a large central green gemstone, 9 hanging front loincloth panel with an ankh symbol and geometric Egyptian patterns, 10 tattered beige side skirt strips, 11 matching gold-and-teal forearm bracer, 12 matching gold-and-teal shin guards and knee plates. In his left hand he holds a massive cracked obsidian greatsword resting across his shoulders; the blade is jagged black stone with glowing orange lava fissures and a fiery molten tip. Add a dangling gold ankh charm from the sword hilt. The mood is regal, brutal, mythic, and intimidating. Use crisp high-detail 16-bit / 32-bit game sprite rendering, strong rim lighting, dramatic highlights on the gold, subtle lava glow from the sword, visible pixel texture, clean silhouette, no text, no border, no watermark, black empty background.

{
  "prompt_description": {
    "subject": "A young adult Caucasian woman with a slender and toned athletic build, appearing to be in her early to mid-20s.",
    "face": "Her face is symmetrical with prominent cheekbones, a delicate nose, and full lips. Her expression is neutral and serene, with a direct gaze. Her blue eyes are piercing and alert. Her skin is smooth with a subtle, healthy glow, showing minimal texture.",
    "hair": "Long, wavy blonde hair with natural highlights, styled to appear windswept. Strands are gently blowing across her face and shoulders.",
    "clothing": "A vibrant red string bikini. The halter top has a deep V-neckline, with thin straps tied around the neck and back. The cups are triangular and tied at the center. The bikini bottoms are low-rise with thin side ties, also tied in bows, with a minimal coverage. The fabric appears to be a slightly stretchy, form-fitting material, clinging to her body. There is a subtle sheen on the fabric due to moisture or lighting.",
    "accessories": "No visible jewelry. Her fingernails are clean and appear to be unpolished or have a very subtle, neutral polish.",
    "environment": "A sandy beach with soft, light-colored sand. The ocean is visible in the background, with gentle waves lapping the shore. The water has a clear, turquoise hue. In the further background, there are out-of-focus elements of a resort or beach establishment, including lounge chairs and umbrellas, and palm trees with lush green foliage. A yellow kayak is partially visible on the sand in the mid-background.",
    "lighting": "Bright, natural sunlight, likely midday. The light is somewhat harsh, creating strong specular highlights on her skin and the wet sand. Shadows are present but not overly deep, indicating an open, sunny environment. The light casts a warm tone on the scene.",
    "atmosphere": "Relaxed, summery, and vibrant. Evokes a sense of leisure, warmth, and natural beauty."
  },
  "style": {
    "art_style": "Ultra-realistic commercial photography, beach lifestyle editorial.",
    "color_grade": "Warm and vibrant, with bright highlights and slightly saturated colors. Skin tones are natural and golden. Blues of the ocean are rich, and the red of the bikini pops.",
    "texture_quality": "Extremely sharp and detailed. High micro-contrast, showcasing the texture of skin, hair, and sand. Virtually no visible grain, suggesting a high-quality sensor and post-processing."
  },
  "camera_settings": {
    "camera": "High-end full-frame or medium format digital camera.",
    "lens": "A moderate telephoto lens, around 70-105mm, providing slight compression and flattering facial features.",
    "aperture": "f/2.8 - f/4, creating a pleasing shallow depth of field that separates the subject from the background.",
    "iso": "Low ISO (e.g., 100-200) for maximum detail and minimal noise.",
    "shutter_speed": "Fast shutter speed (e.g., 1/500s or faster) to freeze motion and "
  }
}

Goal: Create a vertical 4-panel manga comic featuring {argument name="character name" default="Suzuko-chan"}, a cheerful anime girl reacting to successful image upload in an AI playground.

Canvas: Portrait manga page, 3:4 aspect ratio, white outer margin, four equal horizontal panels stacked vertically, thick black panel borders and gutters. Clean high-resolution anime illustration with polished line art, soft cel shading, bright eyes, and pastel gradient backgrounds.

Character details: One recurring character appears in all 4 panels: a cute teenage anime girl with long flowing navy-blue hair, large sparkling blue eyes, fair skin, soft blush, a silver hair clip on the right side of her bangs, a crisp white button-up shirt, and a white sailor-style cap with a small blue “W” emblem. Keep her design consistent across all panels.

Layout and exact panel count: Use exactly 4 horizontal comic panels.

Panel 1: Pale blue gradient background with a large soft white oval glow behind the character. Medium shot of the girl centered slightly left, looking surprised with a small open mouth. On the right, place one vertical oval speech bubble with black Japanese text reading {argument name="panel 1 speech" default="MAIPlaygromで…"}.

Panel 2: Pale pink gradient background with a white oval glow. Close-up of the girl’s face filling most of the panel, smiling brightly with wide blue eyes and blush. On the right, place one vertical oval speech bubble with black Japanese text reading {argument name="panel 2 speech" default="MAI-Image-2.6が…"}.

Panel 3: Warm yellow gradient background with a white oval glow. Waist-up shot of the girl, smiling with both hands raised open in excitement. On the right, place one vertical oval speech bubble with black Japanese text reading {argument name="panel 3 speech" default="ついに画像アップロードできた！"}.

Panel 4: Pale pink background with white glow and scattered colorful confetti pieces. Close-up/medium shot of the girl winking with one blue eye open, smiling widely, giving a thumbs-up toward the viewer. On the right, place one vertical oval speech bubble with black Japanese text reading {argument name="panel 4 speech" default="とっても嬉しい！"}. Add a small signature-like caption at the bottom right reading “鈴子ちゃん”.

Discrete visible elements to include: exactly 4 panels, exactly 4 speech bubbles, exactly 1 recurring anime girl, exactly 1 thumbs-up pose in the final panel, exactly 1 two-hands-raised pose in the third panel, exactly 1 small bottom-right character-name caption, and colorful confetti only in the final panel.

Visual style: Modern Japanese webcomic/anime style, delicate facial features, glossy detailed hair, vivid blue eyes with star-like highlights, soft pastel color palette, clean black outlines, no halftone texture, no gritty effects.

Constraints: Keep all text vertical inside oval speech bubbles, preserve the Japanese text as written, avoid adding extra panels, extra characters, extra logos, or additional captions. Make the comic readable and neatly aligned.