GPT Image 2 prompts

[Final Product] Use the single attached full-body reference image as the sole character reference to convert the subject into a high-quality anime character wearing dignified Kyudo attire while maintaining character identity. Create one 1:1 square practical character sheet that can be used for stable reference of the person, clothing, physique, and base color scheme in subsequent video generation. Stage settings are not handled by this sheet.

[Person Reference] Image A = 1 full-body reference image. Maintain the facial features, eyes, silhouette, nose, mouth, hairstyle, hair color, bangs, parting, physique, height, shoulder width, waist position, leg length, overall proportions, and character impression from Image A. Age and gender are not fixed in text but inherited from Image A. No excessive beautification, rejuvenation, alteration, face-shrinking, physique changes, or unnatural leg lengthening.

[Anime Conversion] Regardless of whether the reference is live-action or an illustration, unify the final character into a high-quality anime style. Ensure the face, hair, body, and clothing naturally blend into a consistent anime art style rather than an ambiguous semi-realistic expression.

[Hairstyle] Basically maintain the hairstyle, color, bangs, parting, and face-framing of Image A. Hair that is short to shoulder-length should remain as is; long hair should be naturally tied at a low position on the neck only if it would interfere with archery movements. Do not add large buns, high ponytails, or ornate hair ornaments not present in the reference.

[Kyudo Attire/Equipment] Do not style as simple practice wear; arrange as dignified Kyudo formal attire suitable for competition or performance. The basic set includes a white upper garment, black to dark blue hakama, obi, and white tabi socks, carefully depicting fabric layering, pleats, sleeves, and the drape of the hakama. The chest guard (muneate) must be consistent as the same equipment across front, side, and back views, with straps and support structures continuing naturally. Do not change the presence, shape, position, or fastening method of the chest guard between the three views. **The archer's glove (yugake) and wrist gear are worn on the RIGHT HAND ONLY. The left hand does not wear a glove in any of the three views. In the front view, the character holds a Japanese bow (Wakyu) with the left hand and wears a brown to ochre yugake on the right hand. In side and back views, the yugake must remain on the same right hand; do not perform changes such as horizontal flipping, swapping the wearing hand, bare hands in the side view, or adding it only in the back view. Glove shape, color, and wrist structure must match across all three views. Do not use purple, blue, or vivid accent colors.** The bow is clearly depicted as a Wakyu and held naturally in the left hand only in the front full-body view. In side and back views, the bow is removed to make the person, hair, and clothing structure easier to read.

[Sheet Composition] Align the front full-body, side full-body, and back full-body views of the same person in the same outfit at nearly the same scale and ground line, showing them from head to toe without cropping. The front view may be slightly larger, but do not treat side and back views as small decorations. Treat the three full-body views as the main focus of this sheet, completing the character information within these views. No face close-ups.

[Color Bar] Place a single thin horizontal 4-color proportional color bar in the bottom margin consisting of white tones, black to dark blue tones, warm tones from the glove, and dark brown tones from the hair or bow extracted from the finished character. Treat it as an auxiliary element to reinforce the character's basic color scheme and do not let it stand out more than the character. The bar should be stretched long horizontally at a low height and should not look like thick color swatches or an independent UI component. The four colors should be divided to reflect the color proportions of the character as a whole; do not include color names, HEX codes, explanatory text, or icons.

[Design/Layout] Unify the design to look like the same series as the stage setting sheet. Background is light gray near white. Place the title "KYUDO CHARACTER SHEET" at the top center with thin horizontal lines on either side. Align the front, side, and back full-body views in the center and place the thin proportional color bar at the bottom. Create design quality through the balance of margins, alignment of the three views, title, thin lines, and bottom color bar without adding extra decorative elements.

[Prohibitions] Face close-ups, multiple expression variations, action pose collections, tiny side/back views, stage backgrounds, long setting text, text other than the title, text-based decorations, thick/large color swatches, independent color chips, excessive Japanese patterns, family crests/bow-and-arrow icons, purple/blue wrist accents, **wearing the glove on the left hand, swapping hands for the glove between views**, fantasy elements, equipment unrelated to Kyudo, character alteration, or changes to hairstyle, physique, clothing, or gear between views.

[Priority] 1. Character identity, 2. Consistency of three views and practical utility for video reference, 3. Maintenance of physique and body proportions, 4. Consistent Kyudo attire, chest guard, and right-hand glove across three views, 5. Natural anime conversion, 6. Clarity of basic color scheme, 7. Unified design consistent with the stage setting sheet.

[Final Product] Create one high-quality anime-style Kyudo dojo stage setting sheet that can be used as a background and spatial reference for subsequent video generation. It does not handle character identity or Kyudo clothing; it organizes the same Kyudo dojo into backgrounds for the three main directions used in videos.

[Input Items] Aspect Ratio: [1:1]. Defaults to 1:1 if not entered.

[Basic Policy] Give top priority to practicality and reproducibility, unifying all three views as the same Kyudo dojo. Match the wood, pillars, floor, walls, ceiling, roof, target area, light direction, time of day, and color tone; do not change to different facilities, buildings, or designs. Fix the style as high-quality anime background art with clear morning to midday natural light and a quiet, calm Japanese atmosphere. Do not draw people.

[3-View Composition] 1. Top Main View: A standard view of the entire Kyudo dojo looking from the shooting area toward the target area. Use the top section largely to create a high-frontality composition where the line of sight from the shooter's position to the target, sense of distance, depth, and architectural structure on both sides are clear. 2. Bottom Left View: A side background that appears behind the person when filming the shooter from the side. Face the camera almost directly toward the side wall of the shooting area, creating a flat background centered on wooden floors, plaster walls, vertical pillars, and horizontal beams. Secure a wide blank wall surface where a person can be placed sideways. Do not include compositions showing a corridor receding, strong perspective, targets, windows, or recessed central wall structures. 3. Bottom Right View: A front background that appears behind the person when filming the shooter from the front. Make the symmetry high and provide a recessed rectangular blank wall section in the center surrounded by pillars and beams. Ensure enough space in the center to place a person front-facing, and keep the left and right simple with plaster walls and wood. Do not include windows, hanging scrolls, frames, or text decorations. Distinguish it clearly from the bottom-left view by providing a shallow depth in the center rather than a simple flat wall.

[Text Elements] Do not include any text within the stage space. Do not place hanging scrolls, frames, signs, tags, posters, slogans, calligraphy, or text-based decorations. Use only the title "KYUDO STAGE SETTING SHEET" at the top center of the sheet, and do not include explanatory labels or numbers within each view.

[Layout] Place the title and thin horizontal lines at the top center. Place one main view in the top section and the side and front backgrounds in two equal-sized columns below. The sheet background is light gray near white. Arrange the margins and keep decorations to a minimum.

[Prohibitions] Architectural changes between the three views, turning views into different facilities, corridor/diagonal perspective compositions in the bottom-left view, making bottom-left and bottom-right views identical, placing targets on the side walls, people, windows, readable text, text-based decorations, atmospheric cuts of unknown purpose, excessive Japanese-style presentation, overly dramatic backlighting, dark lighting, or adding numerous extra views.

[Priority] 1. Consistency of the three views as the same Kyudo dojo, 2. The relationship between the shooting area and target area in the top view, 3. Flat side background for the bottom-left, 4. Front background with shallow central depth for the bottom-right, 5. Clear structural difference between the bottom left and right, 6. Reproducibility for video reference, 7. Sense of organization as a shared sheet.

{argument name="aspect ratio" default="9:16"} aspect ratio. A Japanese woman in her {argument name="age" default="early 20s"}. {argument name="body type" default="Chubby"} build. Fair-skinned round face. A cute smile. Features vary significantly with each generation. Photographed as a portrait. Poses, compositions, and shooting locations are randomized for each generation. Soft, transparent light as if shot in natural light. No strong artificial lighting or harsh shadows.

Subject:
{argument name="subject" default="Walking with a cloud umbrella under summer skies"}

Subject details:
A full-body shot of a woman holding a white cloud like an umbrella canopy under a vivid blue sky. The person stands small in the center of the lower half of the frame, making the vast sky and cloud the protagonists.

Character/Expression:
Dark brown straight shoulder-length hair parted to the side, with ends fluttering to the left in the wind. Profile view with brown eyes looking up at the high-held umbrella, a gentle expression with naturally closed lips. Slender build, bright natural skin, face rendered with modest detail appropriate for the distance.

Clothing/Pose:
{argument name="clothing" default="Light blue short-sleeved cropped blouse and a matching light flared mini skirt"}, with white strappy heel sandals. A cream-colored thin woven bag worn diagonally from the right shoulder. Standing at the edge of an embankment with legs lightly crossed, extending her right arm straight up to hold an umbrella with a thin black shaft and white curved handle. Left arm is relaxed at her side.

Background/Lighting:
{argument name="background" default="A clear deep blue sky occupying most of the frame"}. A single horizontal white cloud with cotton-like thickness and fine ridges floats above her head, with the black umbrella shaft connecting to its lower center. White cumulus clouds at lower levels, misty blue-green mountains, and blurred long grass in the foreground. Hard direct sunlight from the sun at the top right brightly illuminates the white clothes and cloud, creating distinct shadows.

Composition/Camera:
3:4 aspect ratio. A low-angle full-body shot with the camera placed near the ground looking up. The subject is slightly to the right of center, the cloud umbrella is in the top center, leaving wide blue sky negative space above and to the sides. The whole body and cloud are captured, with the diagonal embankment crossing the bottom. Focus on the person and the cloud; foreground grass is blurred. Natural compression of a medium telephoto lens.

Texture/Style:
High-definition photorealistic photo. Transparent summer colors, clean blue and white palette, soft hair and thin fabric in the wind, and the fluffy three-dimensional feel of the cloud. Refreshing and slightly fantastical, but the lighting on the person and scenery is realistic.

Negative:
Don't change the cloud umbrella to a normal fabric umbrella; don't use eye-level views or close-ups of the person.

A '{argument name="subject" default="sheep"}' being attacked by a '{argument name="monster" default="harpy"}'

Blueprint for a robot from the '{argument name="organization name" default="Metropolitan Police Heavy Equipment Task Force Robot Unit"}'

Create an ultra-realistic candid street-style photograph of a {argument name="subject" default="young man in his early 20s"} as a reference image walking through a {argument name="setting" default="modern upscale grocery market"}. He is photographed from behind and slightly from the side as he looks over his shoulder toward the left. He has short, naturally messy dark brown hair and a slim athletic build. He is wearing an {argument name="outfit" default="oversized dark navy Acne Studios hoodie with large white text on the back reading \"Acne Studios Stockholm 1996\", a matching dark navy baseball cap, and a black backpack worn over both shoulders"}. The hoodie has a relaxed oversized fit with natural fabric folds. The setting is a large premium supermarket with fresh produce displays, flowers, shoppers, escalators and industrial ceiling details. Warm indoor lighting mixed with natural ambient light. Candid photography, realistic skin and hair texture, natural body proportions, subtle motion, authentic smartphone camera quality, shallow depth of field, realistic background detail, understated luxury streetwear aesthetic, no artificial or CGI appearance. Vertical composition, 4:5 aspect ratio.

{
  "title": "The Man Who Sold Tomorrow",

  "image_description": "A cinematic neo-noir scene set in a {argument name="city style" default="futuristic city"} during heavy rain. A mysterious businessman stands alone beneath a glowing holographic billboard 500 stories above the city. Countless neon lights reflect across the wet rooftop surface. Flying vehicles drift silently through dense fog between skyscrapers. The man wears a perfectly tailored {argument name="clothing" default="black overcoat"} and leather gloves. In his hand is a small {argument name="prop" default="golden pocket watch"}. The watch displays tomorrow's date. His expression is calm but deeply troubled, as if he knows something nobody else does. Lightning briefly illuminates the city skyline behind him.",

  "subject": {
    "character": "Mysterious businessman",
    "appearance": "Elegant black overcoat, leather gloves, sharp facial features",
    "expression": "Melancholic, thoughtful, secretive",
    "pose": "Standing at the edge of a skyscraper rooftop"
  },

  "environment": {
    "location": "Cyberpunk megacity rooftop",
    "weather": "Heavy rain",
    "atmosphere": "Fog, neon reflections, distant flying vehicles"
  },

  "story_element": {
    "twist": "The pocket watch shows tomorrow's date because he already knows exactly how the world ends."
  },

  "cinematography": {
    "camera_angle": "Low angle",
    "shot_type": "Medium wide shot",
    "lens": "85mm anamorphic lens",
    "depth_of_field": "Shallow depth of field"
  },

  "lighting": {
    "primary": "Neon city lights",
    "secondary": "Lightning flashes",
    "effects": [
      "Volumetric fog",
      "Rain reflections",
      "Cinematic glow"
    ]
  },

  "color_palette": [
    "Blue",
    "Cyan",
    "Amber",
    "Black"
  ],

  "mood": [
    "Mysterious",
    "Thrilling",
    "Emotional",
    "Cinematic"
  ],

  "quality": {
    "style": "Denis Villeneuve inspired sci-fi cinema",
    "realism": "Ultra photorealistic",
    "resolution": "8K",
    "film_grain": true,
    "IMAX_quality": true,
    "award_winning_cinematography": true
  }
}

Create a highly detailed, cinematic black-and-white portrait of a young woman, centered and facing the camera with a calm, confident expression. She has {argument name="hair style" default="long, voluminous wavy hair"} with soft curtain bangs framing her face. Preserve realistic and delicate facial features, natural proportions, expressive eyes, and subtle lips.

Render the entire portrait using a high-contrast monochrome halftone particle technique, inspired by vintage screen-printing, stippled ink illustrations, and retro comic-book artwork. Use thousands of tiny black-and-white dots to form the facial details, hair, skin, and clothing. Add strong directional lighting with bright white highlights and deep black shadows for a dramatic three-dimensional effect.

The woman is wearing a vintage-style blouse with a {argument name="collar type" default="large white Peter Pan collar"} and vertically pleated fabric. Surround her with a {argument name="background color" default="pure black background"}, subtle glowing white particles, tiny floating geometric fragments, dust-like specks, and scattered points of light. Add a luminous white rim light around her hair and shoulders.

Ultra-detailed facial rendering, precise halftone dots, sharp edges, dramatic contrast, artistic monochrome aesthetic, cinematic composition, high clarity, professional poster design, 8K detail, symmetrical portrait framing, sophisticated editorial artwork, no text, no watermark, no inset image, no extra people.

Subject: {argument name="subject" default="Coral flower dress"}

Main Subject:
A vertical photograph capturing a single woman lying on a light pink floor from directly above, with massive coral and pale beige dress fabric spreading across the entire screen like petals. The right thigh is visible through a deep slit.

Person/Expression:
The upper body lies diagonally from the top left to the bottom right of the frame, face turned toward the camera above, staring straight with brown eyes, a calm expression with slightly parted glossy coral lips. Small oval face and thin chin, rounded almond-shaped eyes, thin parallel eyebrows, slender bridge of the nose and a small tip, thin upper lip and thick lower lip. Pink blush and eyeliner at the outer corners of the eyes, bright glowing skin. Long light brown hair spread in a fan shape to the left side of the head with thin bangs and large waves.

Clothing/Pose:
Strapless long dress exposing all shoulders. The chest is wrapped in fine horizontal pleats of pale beige, with a thick coral cloth band crossing diagonally from the right chest, gathering at a large flower ornament at the front waist. Layers of translucent fine vertical pleated fabric, with one leg coming out from a deep slit on the right hip. Left arm bent overhead with the forearm on the forehead, right arm resting on the waist with fingers overlapping. Both legs crossed and cut off at the bottom.

Background/Light:
The entire background is a smooth {argument name="background color" default="pale mauve pink"} floor. Coral, pale pink, and beige chiffon fabric waves and spreads toward the bottom left, top right, and right edge. A firm but soft main light hits from the top left, creating a bright gloss on the arms, face, chest, and legs, and casting clear shadows on the fine pleats of the fabric.

Composition/Camera:
3:4 vertical, fashion portrait near full body from a top-down vertical bird's-eye view including head to crossed legs in frame. Face is near the top left, exposed right thigh at bottom right, wide drapes fill the surrounding margins. Focus on the face and flower ornament, with deep depth of field showing the entire fabric clearly.

Texture/Style:
High-quality photorealistic live-action photography. High-definition depiction of translucent chiffon, fine pleats, 3D flower ornaments, skin, and hair, coordinated in {argument name="color scheme" default="analogous colors from coral to mauve"}.

Negative:
Do not turn the top-down overhead view into a front-facing standing pose; do not omit the wide drapes and deep slit.

Use the provided reference image as the main visual reference. Create a highly photorealistic close-up portrait of a {argument name="subject" default="young woman"} with long, straight, silky {argument name="hair color" default="dark-brown"} hair, parted naturally in the center and falling loosely over her shoulders. Preserve the same facial structure, soft features, natural eyebrows, almond-shaped dark eyes, delicate nose, smooth fair skin, and soft pink-red lips. She is looking toward the camera with a gentle, subtle smile and a calm, elegant expression.\n\nShe is wearing a simple {argument name="clothing" default="light-gray sweatshirt"} with a relaxed fit. Capture her from the shoulders up in a slightly turned three-quarter pose, with her face angled gently toward the camera. Use a warm, cozy indoor background with soft brown and beige tones, naturally blurred with shallow depth of field.\n\nPhotorealistic skin texture, realistic hair strands, natural facial details, soft warm ambient lighting, subtle highlights, realistic shadows, cinematic portrait photography, 85mm lens look, shallow depth of field, high detail, pro…

{\n "CITY_NAME": "{argument name="city name" default="London"}",\n "SUBJECT": "{argument name="subject description" default="a person"}",\n\n "prompt": "Recreate the reference image as a premium editorial city poster. Replace the original person with SUBJECT while preserving SUBJECT's identity, facial features, hairstyle, hair color, skin tone, and natural appearance. SUBJECT must be standing in the SAME pose, SAME body orientation, SAME framing, SAME scale, and approximately the SAME position as the reference subject.\\n\\nReplace the original city with CITY_NAME and rebuild the surrounding skyline, landmarks, architecture, bridge, and urban photography specifically for CITY_NAME. Replace all city-specific labels, coordinates, location information, and captions accordingly.\\n\\nCRITICAL TYPOGRAPHY COMPOSITION: The oversized CITY_NAME typography must be physically layered OVER the standing character, exactly like the reference. Do NOT place all of the city text behind the subject. The large CITY_NAME letters must pass across the subject's body and overlap her clothing and lower torso, visibly covering portions of her. At the same time, some portions of the typogra…"

A striking young woman with {argument name="hair" default="long tousled dark-brown hair"}, {argument name="eye color" default="luminous blue-gray eyes"}, glossy natural skin, and an athletic hourglass build stands confidently. Blend gritty urban biker fashion with elegance : a {argument name="jacket style" default="distressed deep-red leather jacket"} worn open over a fitted black cropped top, paired with sleek black high-gloss pants,. Her pose is relaxed and self-assured, one hand lifting her hair, direct gaze toward camera. Low-to-mid angle fashion portrait, waist-to-thigh framing, shallow depth of field. Warm golden backlight pours through,, leather, skin highlights, floating dust, and fabric. Semi-realistic anime realism, refined facial features, cinematic editorial, luxurious textures, sensual but powerful, crisp focal subject, softly glowing background.

Ultra-premium vertical fashion-tech poster, Cannes-level collectible campaign. {argument name="model" default="Young male/ female model"}, waist-to-mid-thigh framing, slightly low angle, calm confident pose, head slightly raised and tilted, elongated neck. No celebrity likeness. {argument name="background" default="Dense flat cobalt-blue background"} with ultra-thin technical grid, registration marks, subtle print wear and silkscreen traces. Mixed-media fashion illustration: expressive ink contours, graphite, dry brush, selective marker fills, visible sketch lines, tactile pen pressure, textured art paper. {argument name="accents" default="Luxury mustard-yellow print accents"}, grainy paper, subtle ink bleeding, natural print imperfections, sharp yet organic lines, rare collectible screen-print finish, 8K. No photorealism, cartoon, anime, clutter, props, extra people, heavy typography, or generic commercial layout.

{argument name="brand name" default="[BRAND NAME]"} → Premium Energy Drink Concept

Transform {argument name="brand name" default="[BRAND NAME]"} into a futuristic, premium energy drink brand using its existing logo, colors, typography, shapes, proportions, and visual personality as the core creative DNA.

Do not simply place the logo on a generic can. Reimagine the identity into a distinctive energy drink with a custom can silhouette, intelligent label system, signature energy graphics, flavor identity, premium typography, metallic accents, embossing, realistic aluminum texture, and condensation.

Create a sophisticated 3:4 vertical product-design board featuring one dominant photorealistic hero can in 3/4 view, plus front, side, back, angled, and detail views. Include a matching multipack/box and refined design callouts for logo integration, color palette, flavor concept, materials, energy graphic system, and label structure.

Use high-end commercial product photography, controlled reflections, dramatic but realistic lighting, soft shadows, and a clean futuristic studio environment. Make every design decision specific to {argument name="brand name" default="[BRAND NAME]"}

Create a premium contemporary fashion editorial poster using the uploaded subject as the main character. Preserve the subject's identity, facial structure, natural skin tone, hairstyle, hair color, body proportions, and recognizable appearance accurately.

The BACKGROUND must be generated automatically based on the overall creative direction. Do not require the user to define the background. Automatically create a sophisticated environment, studio setting, architectural space, abstract backdrop, or atmospheric scene that best complements the subject, wardrobe, pose, and editorial concept. The background should feel intentional, premium, realistic, and visually cohesive.

The subject must follow the user-defined POSE naturally and convincingly. Automatically generate the wardrobe, accessories, makeup, hair styling, lighting, color palette, camera angle, composition, and overall art direction to suit the generated concept. These elements should not require user input.

Use the user-defined TEXT as the primary oversized editorial typography. Integrate it directly into the composition like a professional fashion magazine cover. The typography should be bold, modern, clean, and large-scale, with intentional spacing and layering. The subject may overlap or partially obscure the typography where appropriate to create depth. Ensure the text is correctly spelled and clearly readable.

Do not add random promotional copy, prices, dates, sale information, product lists, fake brands, logos, watermarks, or unnecessary small text. Only add minimal secondary typography if specifically required by the visual concept.

Create a high-end fashion editorial aesthetic with cinematic realism, realistic skin texture, natural anatomy, authentic fabric materials, sophisticated lighting, subtle film grain, controlled contrast, premium color grading, and strong visual hierarchy. Avoid generic AI aesthetics, plastic skin, excessive beauty retouching, distorted anatomy, extra fingers, duplicated limbs, warped clothing, messy typography, or visual clutter.

The final composition should feel like a professionally art-directed contemporary fashion magazine/editorial campaign. The subject is the focal point, the user-defined TEXT is the dominant graphic element, and everything else—including the background—is generated automatically to support the concept.

Create a premium, modern high-fashion CHARACTER CONCEPT ART SHEET in a {argument name="layout" default="16:9 widescreen"} layout on a {argument name="background" default="pure white"} background. THE ATTACHED REFERENCE IMAGE DEFINES THE ART STYLE — treat it as a strict style bible and reproduce its exact rendering

Create 10 separate ultra-cinematic Pixar 3D animation quality storyboard frames for a 15-second superhero origin story of {argument name="superhero name" default="IGNIS"} — a {argument name="subject description" default="young male fire"} superhero. Style: Pixar/Disney 3D animation quality, vibrant colors, dramatic cinematic lighting, emotional

Create a premium cinematic character bible sheet for a {argument name="profession" default="basketball player"} named {argument name="character name" default="JAY"}. Use the uploaded character sheet as the strict visual reference for his exact face, body, proportions, hair, skin tone and outfit. Do not change his appearance.

Create a highly realistic cinematic commercial-style portrait scene featuring a {argument name="main subject" default="young South Asian woman"} sitting casually behind a polished wooden table in a warm, cozy indoor room. She has long dark brown hair, natural makeup, soft expressive eyes, and wears a simple black short-sleeve T-shirt; she leans forward with her arms resting on the table, smiling gently and looking down toward the tabletop. On the table in the foreground, place exactly 4 key objects: 1 large beige ceramic mug on the left filled with steaming hot coffee, 1 upright dark brown NESCAFÉ 3-in-1 instant coffee sachet leaning near the mug with clear readable vertical “NESCAFÉ” branding and “3 in 1” text near the bottom, 1 tiny miniature version of the same woman sitting on the tabletop beside the sachet with arms crossed, wearing a dark green jacket, brown pants, and white sneakers, and 1 small scattered pile of coffee granules near the mug. The miniature woman should appear realistic and doll-sized, leaning back against the sachet, eyes closed with a calm confident expression. Use warm amber lamp light from a blurred table lamp in the background on the left, shallow depth of field, realistic steam rising from the mug, glossy wood reflections, cozy evening atmosphere, soft cinematic bokeh, dramatic commercial lighting, ultra-detailed skin texture, realistic proportions, vertical 9:16 composition, no extra people, no extra text beyond the coffee sachet branding.

Goal: Create a clean 12-frame advertising shot map for a Korean strawberry milk soda can, designed like a polished product-commercial storyboard on a white background.

Canvas: 16:9 horizontal canvas, 720p feel, divided into exactly 12 panels arranged in a 4 columns by 3 rows grid. Use very thin pale blue divider lines between panels and around the outer border. High-key studio lighting, crisp product photography, glossy liquid simulation, mostly white negative space, blue-red-pink brand palette.

Product and branding: Feature a slim aluminum can of {argument name="product name" default="LOTTE Milkis Strawberry"}. The can is white and silver with condensation droplets, bold chunky blue “Milkis” lettering, small black “LOTTE” logo near the top, red “Strawberry” text, a small strawberry icon, pink powdery gradient patches, red crescent-like swoosh graphics, and a tiny blue line-art umbrella girl character printed on the label. Add small Korean text near the logo where appropriate. Keep the label legible but allow slight perspective distortion.

Recurring character and liquid motif: Use one tiny blue line-art umbrella girl mascot, drawn like a simple ink illustration with a skirt and parasol, appearing in multiple panels as if she is surfing or floating through the ad. Use sparkling translucent pink strawberry soda liquid with many bubbles, occasional white creamy milk ribbons, fresh strawberries with green leaves, and a clear glass filled with pink fizzy drink.

Layout: Include exactly 12 discrete storyboard panels, listed left to right, top to bottom:
1. Extreme close-up of the wet can label, cropped vertically, showing large blue “Milkis,” red “Strawberry,” strawberry icon, condensation beads, and the umbrella girl printed on the can.
2. Angled close-up of the can label, slightly cleaner and brighter, the umbrella girl appears larger near the red Strawberry text, with soft pink gradient on the can.
3. Full upright can centered in a white studio space, silver pull tab visible at top, label fully readable, tiny blue steam-like flourish above the can opening.
4. Diagonal can tilted upper left to lower right, open top visible, a dynamic arc of translucent pink soda splashing out to the right with suspended bubbles.
5. Wide minimalist panel with only the left edge of the can cropped into frame and the blue umbrella girl floating in open white space above.
6. Pink soda wave sweeps diagonally across the panel with many bubbles; the tiny umbrella girl rides above the wave holding her parasol.
7. Fresh red strawberry with green leaves riding on the pink soda wave; the umbrella girl stands on top of the strawberry like a surfer.
8. Creamy white milk ribbon twists through the pink soda splash; a strawberry sits in the flow and the umbrella girl balances on it.
9. Wide shot with the can small and cropped at bottom left, pink soda stream and bubbles flowing diagonally, a strawberry suspended mid-splash, umbrella girl drifting nearby.
10. Extreme graphic close-up of the large blue “Milkis” logo letters filling the frame, with pink bubbles at the bottom and the tiny umbrella girl climbing or floating over the letters.
11. Clear glass close-up filled with pink fizzy soda, bubbles rising; a strawberry floats inside and the umbrella girl overlays the glass as if parachuting in.
12. Final product packshot: full can standing beside a tall clear glass of pink soda and one fresh strawberry at the base, white background, clean commercial composition.

Visual style: Premium beverage advertisement storyboard, photorealistic can, glass, fruit, bubbles, and splashes combined with a hand-drawn blue mascot illustration. Bright whites, glossy highlights, realistic refraction in liquid, crisp macro detail, airy Korean soft-drink commercial aesthetic.

Constraints: Use exactly 12 panels in the 4x3 grid. Do not add hands, people, extra packaging, price tags, captions, watermarks, or additional products. Keep all panels visually connected by the same white background, pink strawberry soda, blue umbrella mascot, and consistent Milkis can design.

Goal: Create a realistic screenshot of a Chinese social media post on a smartphone, showing a viral AI-generated illustrated story card about an ancient palace dancer and an emperor, with engagement metrics circled in red.

Canvas: Vertical smartphone screenshot, 9:16 aspect ratio, white app interface background, clean mobile UI. The status bar at the top shows 11:22, 4G signal, and battery icon.

Top app header: Include a back arrow on the left, a small circular female profile avatar, the account name {argument name="account name" default="菲菲_AI智能体"}, a rounded pink outline follow button labeled “关注”, and a share arrow icon on the right.

Main image post: A beige comic-style image card centered below the header, with a small dark rounded page indicator at the top-right reading “2/10”. The card contains exactly 2 stacked illustrated panels, each bordered in black, separated by large cream caption bands with bold Chinese brush-calligraphy text.

Panel 1: Ancient Chinese imperial palace scene in soft muted watercolor comic style. On the right, a stern emperor in ornate golden robes sits on a dragon-carved throne under curtains, with palace architecture behind him. In the center foreground, a graceful female dancer in pale flowing robes performs with her back partially turned. On the left, exactly 5 seated female musicians in traditional robes play instruments, including pipa and string instruments. The caption band below reads “在玉皇大帝眼中”, with “玉皇大帝” emphasized in dark red and the rest in black.

Panel 2: Frontal dramatic portrait of the same dancer standing in a palace hall, looking sad and vulnerable, wearing pale white-blue hanfu, ornate hair accessories, and delicate makeup. Behind her are exactly 2 armored palace guards, one on each side, plus a blurred group of palace women in the background. Cool blue-gray lighting creates a solemn atmosphere. The caption band below reads “她仅仅只是一名舞女”, with “舞女” emphasized in dark red and the rest in black.

Post text below image: Show a carousel dot row with exactly 5 dots, the second dot red and the other 4 light gray. Then display the Chinese caption text in two lines: “人还是原本那个人，在不同人眼里价值截然不” and “人还是原本那个人，在不同人眼里价值截然不同”. Beneath it, show hashtags in blue-gray text: “#人与人的区别 #AI教程 #AI工具 #人生的价值 #ai关键词 #提示词分享 #AI学习 #AI绘图 #AI漫画 #AI新手村”.

Bottom interaction bar: Include a rounded comment input field labeled “说点什么...”. To the right, show exactly 3 engagement items: a heart icon with {argument name="like count" default="2072"}, a star/favorite icon with {argument name="favorite count" default="876"}, and a comment bubble icon with {argument name="comment count" default="110"}. Draw a thick hand-drawn red oval around these 3 metrics only.

Visual style: The app UI should look like a real screenshot, while the embedded artwork should look like polished AI comic illustration with ink outlines, soft watercolor shading, muted palace colors, and cinematic ancient Chinese drama atmosphere.

Constraints: Keep all visible Chinese text legible and accurately placed. Do not add extra panels, extra engagement icons, watermarks, or unrelated UI elements.

Goal: Create a humorous 4-panel meme comic about {argument name="character name" default="an anxious introvert"} being forced into increasingly social situations, in a polished digital cartoon style.

Canvas: Landscape 4:3 comic page with four equal panels in a 2×2 grid, thick black gutters, warm saturated colors, expressive faces, soft evening lighting, slightly exaggerated Pixar/anime-inspired character design, clean readable speech bubbles and meme captions.

Main character: A pale, skinny young adult introvert with messy black hair, huge round anxious eyes, dark hoodie, dark plaid pajama pants or black pants depending on panel, awkward slouched posture, and a constantly overwhelmed expression. Keep the same character recognizable across all 4 panels.

Panel 1, top left: Cozy dim bedroom or apartment at night. The introvert stands alone holding a black mug labeled “LEAVE ME JUICE,” wearing gray cat slippers. Background details include a bookshelf, string lights, a beanbag or chair with a pillow that says “NO THANKS,” a framed wall sign reading “SOCIAL BATTERY 100%,” and a phone-style notification card in the foreground. Thought bubble text: “Perfect. A calm day with zero human interaction.” Notification text: “REMINDER: Team-building retreat starts now!” Bottom yellow meme caption: “introvert difficulty setting: impossible”.

Panel 2, top right: Chaotic networking or team-building party packed with smiling coworkers around the terrified introvert. Use purple party lighting, string lights, banners, confetti, drinks, name tags, and exaggerated extrovert energy. A banner in the background reads “CONNECT! COLLABORATE! COMMUNICATE!” Add background exclamations: “BLA BLA BLA!”, “HAHA!”, and “COOL!” The introvert stands in the center in a black hoodie with a red name tag that says “HELLO MY NAME IS NOPE.” A blond extrovert leans in with a huge smile and speech bubble: “Let’s go around and share a fun fact!” The introvert’s small speech bubble: “I once microwaved tea.” Include 4 visible coworker name tags: “HELLO MY NAME IS TARA,” “HELLO MY NAME IS PEOPLE PERSON,” “HELLO MY NAME IS DAVE,” and “HELLO MY NAME IS CHATTY KAT.”

Panel 3, bottom left: Karaoke scene split within the panel into two stacked moments. Top portion: the introvert hides behind a large potted plant while an excited woman in a colorful patterned shirt points at them, with a speech bubble saying “OMG you’re perfect for karaoke!” Her name tag reads “HELLO MY NAME IS SUNSHINE SUSAN.” Bottom portion: the introvert is forced on stage gripping a microphone, trembling with panic in front of a karaoke monitor. Neon purple stage lights, music notes, hearts, and a giant lyric screen reading “DON’T STOP BELIEVIN’.” The introvert’s speech bubble says: “Please do stop believin’.”

Panel 4, bottom right: Loud brunch/club afterparty atmosphere with a crowded table full of laughing friends, champagne flutes, pancakes, fruit, plates, coffee, and party lights. The introvert sits in the center, exhausted and dead-eyed in a black hoodie, with a mug labeled “SOCIAL OVERLOAD” and a skull icon. A happy person shouts in a speech bubble: “We’re all going clubbing after this!” A small ghost version of the introvert floats above, X eyes and halo, saying: “I’m freeeeee... (goodbye cruel world...)” A wall sign reads “BRUNCH IS THE NEW PRE-GAME.” Bottom black meme caption with mixed yellow and white text: “me: goes outside once” and “the universe: congratulations, you are now booked for the full extrovert experience”.

Constraints: Use exactly 4 panels. Preserve all listed text legibly. Make the humor clear through contrast between the introvert’s panic and the crowd’s enthusiasm. No watermark, no extra panels, no unrelated characters, no photorealism.

Create a vertical, photorealistic fine-art portrait of a regal {argument name="cat color" default="solid black"} cat sitting upright on an ornate dark purple brocade cushion in a luxurious Victorian gothic interior. The cat is the single main subject, centered slightly left, with sleek glossy fur, tall pointed ears, long whiskers, alert posture, and pale golden eyes gazing off to the left. Around its neck is an antique gold filigree necklace with a dark oval gemstone pendant, giving the cat an aristocratic, mysterious look. The setting is dim and elegant: deep plum velvet curtains with a gold tassel tieback on the left, a dark charcoal wall, a partially visible gilded picture frame in the upper right, and an ornate gold side table in soft focus behind the cat holding a brass vase with dark purple flowers and a small candleholder-like ornament. Include exactly these visible background decor elements: 1 large curtain with 1 tassel tieback, 1 gilded wall frame, 1 ornate side table, 1 brass flower vase, 1 bunch of dark purple flowers, 1 small gold ornament, and 1 extra purple cushion at the lower right. Use a moody chiaroscuro lighting style with soft directional light from the left highlighting the cat’s face, eyes, fur texture, and jewelry, while the background falls into natural shallow-depth-of-field blur. Color palette: black, deep violet, antique gold, charcoal, and muted burgundy. Render as high-end editorial pet photography, ultra-detailed fur, realistic anatomy, elegant composition, no text, no watermark, no extra animals.

Create a highly detailed fantasy anime illustration of a vast treetop city built around an enormous ancient tree, viewed from a high wooden café terrace in warm afternoon sunlight. In the foreground, show exactly 2 young anime girls seated at a small square wooden table on the balcony, calmly drinking tea and looking at each other; one has {argument name="first girl hair color" default="light brown hair"} and a white blouse with a dark capelet, the other has {argument name="second girl hair color" default="dark brown ponytail"} and a green cloak. On the table include exactly 4 visible tea-service items: 2 cups held by the girls, 1 dark teapot, and 1 small cup or bowl near the teapot. Surround them with rustic wooden railings, chairs, floorboards, leafy vines, potted plants, and dappled sun patches. Beyond the balcony, depict a breathtaking vertical fantasy metropolis made of timber, glass, brass, and greenery, with countless tiny houses, platforms, stairways, bridges, railings, lanterns, and market details layered deep into the scene. The centerpiece is {argument name="central structure" default="a colossal tree trunk wrapped by multiple spiral train tracks"}; show exactly 5 prominent spiral rail loops encircling the trunk at different heights. Include exactly 4 clearly visible tram cars: 1 on the lower-left track, 1 on the mid spiral track, 1 on the glass-roof building slope, and 1 on the upper-right rail line. Add a large glass conservatory-like building with a steep turquoise roof on the right-center, a small red-and-white circus tent near the lower center, and suspended walkways disappearing into dense foliage. Use {argument name="art style" default="intricate cinematic fantasy anime background art"}, ultra-rich linework, painterly textures, glowing sunbeams through leaves, teal-and-gold color grading, high-frequency architectural detail, magical realism, deep perspective, no text, no watermark, vertical composition.