图片提示词

diagram:
  title: "{argument name="diagram title" default="The Cycle of the Generative AI Agent Development Swamp"}"
  style:
    character: "simple_editorial_cute"
    tone: "light_like_original"
    line: "soft_round"
    shading: "flat_minimal"
    palette: "muted_editorial"
    background: "light_beige_or_white"

  rule:
    first_character_only: "Only the first person is different"
    others_same_character: "All others are the same person"
    anxiety_requires_tech_terms: true
    dialogue_tone: "Maintain the short, light phrasing of the original diagram"

layout:
  intro_track: "The upper section is a straight line (left to right), then drops down to join the loop"
  loop_track: "Arranged as a loop (ring), with arrows curving clockwise along the circumference"

nodes:
  # ===== Upper Section: Straight Introduction =====
  - id: A
    zone: "intro"
    position: "upper_left"
    actor: "Different person (Senior Engineer)"
    visual: "Pointing at a laptop"
    item: "AI Agent Demo Screen"
    says: "It'll make you happy."
    arrow_to: { id: B, shape: "straight", direction: "right" }

  - id: B
    zone: "intro"
    position: "upper_center"
    actor: "Protagonist (Same person)"
    visual: "Waving lightly"
    says: "Even a simple agent..."
    arrow_to: { id: C, shape: "straight", direction: "right" }

  - id: C
    zone: "intro"
    position: "upper_right"
    actor: "Protagonist (Same person)"
    visual: "Executing sample code"
    item: "LangChain / CrewAI Sample"
    arrow_to: { id: D, shape: "straight", direction: "down" }

  # ===== D: Awakening Point =====
  - id: D
    zone: "connector"
    position: "right_outer_before_loop"
    actor: "Protagonist (Same person)"
    visual:
      face: "Eyes wide open (cute)"
      effect: "Small flash of inspiration line"
      posture: "Leaning forward"
    state: "Awakening"
    says: "Agents, this is insane..."
    micro_text: "(Wakes up here)"
    arrow_to:
      id: E
      shape: "gentle_curve"
      direction: "down_into_ring"

  # ===== Loop Start (Must rotate from here) =====
  - id: E
    zone: "loop"
    position: "loop_right_upper"
    actor: "Protagonist (Same person)"
    visual: "Holding head in hands (cute)"
    state: "Anxiety/Irritation Begins"
    says: "Irritated"
    thought_terms:
      - "Prompt Design"
      - "Tool Selection"
      - "Memory Design"
    thought_text: "It's not working as intended..."
    arrow_to: { id: F, shape: "curved_along_ring", direction: "clockwise" }

  - id: F
    zone: "loop"
    position: "loop_right_lower"
    actor: "Protagonist (Same person)"
    visual: "Surrounded by code and logs"
    state: "Irritation Settles In"
    says: "In a state of irritation"
    callout: "Loop! Runaway! Hallucination!"
    thought_terms:
      - "RAG"
      - "Function Calling"
      - "Tool-Use"
    arrow_to: { id: G, shape: "curved_along_ring", direction: "clockwise" }

  - id: G
    zone: "loop"
    position: "loop_bottom"
    actor: "Protagonist (Same person)"
    visual: "Covered in sticky notes and diagrams"
    says: "The body gradually develops tolerance, and usage increases."
    items:
      - "Agent Architecture Diagram"
      - "Role Assignment Memo"
      - "State Transition Flow"
    arrow_to: { id: H, shape: "curved_along_ring", direction: "clockwise" }

  - id: H
    zone: "loop"
    position: "loop_left_lower"
    actor: "Protagonist (Same person)"
    visual: "Works well, small fist pump"
    state: "Temporary Satisfaction"
    says: "Desire is temporarily satisfied"
    arrow_to: { id: I, shape: "curved_along_ring", direction: "clockwise" }

  - id: I
    zone: "loop"
    position: "loop_left_upper"
    actor: "Protagonist (Same person)"
    visual: "Thinking again"
    state: "Obsession/Thought Occupation"
    says: "Thinking about the agent

Goal: Create a clean 3D character design turnaround sheet for a dark fantasy humanoid bird-shaman character, suitable for modeling and rigging reference.

Canvas: Wide horizontal canvas, 16:9 aspect ratio, light warm gray studio background with no props, no shadows beyond soft contact shading, centered full-body figures with generous spacing.

Layout: Show exactly 4 full-body character views in a straight row, all in a rigid T-pose with arms extended horizontally: 1 front view, 2 left side profile view, 3 back view, 4 right side profile view. Keep all four views the same height and scale, evenly spaced, feet aligned on the same baseline.

Character details: The character is {argument name="character archetype" default="a dark fantasy bird-masked tribal shaman"} with a slim humanoid body, black fitted under-suit or dark skin-tight body layer, birdlike plague-doctor mask or long beak helmet, and a crown of thin red-orange feather or tendril plumes emerging from the head. The face is hidden by the mask; the front mask has a long pale bone-colored beak, dark eye openings, and carved ornamental markings. The back view shows a darker head with red tendrils visible from behind.

Outfit: The costume has exactly these visible main clothing pieces: 1 long open sleeveless olive-green robe or coat with split front and pointed tails, 2 short pale green sleeves over the upper arms, 3 black fitted forearm guards or gloves, 4 black leggings, 5 open-toe sandals, 6 layered dark feather collar around the shoulders, 7 pale leaf-like feather necklace around the neck, 8 circular bronze ring ornaments at the waist and knees, 9 hanging strap details at the hips, 10 a vertical laced panel down the upper back. The robe color is {argument name="robe color" default="muted olive green"} with weathered fabric texture and faded salmon-pink circular patterns on the sides and lower back. Use bronze, bone, leather, and dark charcoal materials.

Pose and anatomy: T-pose character sheet, arms perfectly straight and level, palms facing down or slightly inward, legs straight, feet shoulder-width apart. The figure should feel game-ready and sculptural, with consistent proportions across all views.

Visual style: High-quality 3D render, realistic stylized fantasy design, matte materials, subtle worn fabric texture, crisp details, neutral orthographic camera, no dramatic lighting, no text labels, no grid, no watermark. Emphasize symmetry, readable silhouettes, and production-art clarity. Use {argument name="render style" default="realistic stylized 3D game character turnaround"}.

Create a polished pastel anime livestreamer illustration in a cozy bedroom studio at night, featuring one cute catgirl VTuber named {argument name="character name" default="Lua Nyang"} seated close to the camera behind a desk. She has very long silky {argument name="hair color" default="pink"} hair flowing across her shoulders, two fluffy pink cat ears, large sparkling pink eyes, a shy open-mouthed smile, flushed cheeks, and a soft glossy anime complexion. Her pose is leaning forward with one hand touching her cheek, looking directly at the viewer in a playful affectionate way. Dress her in a frilly pink-and-white kawaii outfit: one oversized white bow on her head, a braided headband, one off-shoulder pink cardigan sleeve visible on each arm, a heart-pattern sailor-style collar, ruffled trim, a central ribbon bow, and small bow accents. Use a dreamy pink lighting palette with bloom, soft depth of field, lens sparkle, and high-detail gpt-image-2 anime rendering. The desk foreground contains exactly three main items: one pink mechanical keyboard, one white mouse on a pink mat, and one clear cup of pink soda with a straw and cute decoration. On the left side place exactly one black streaming microphone with a round pop filter, pink bow, and small cat charm. In the room background include exactly one neon sign on the left reading {argument name="neon sign text" default="루아냥"} beside a simple glowing cat outline and heart, one plush display shelf with small cute plushies, one large round cat plush chair or cushion on the right, one window showing a crescent moon and dreamy night sky, and several framed pastel pictures on the wall. Keep the whole composition romantic, fluffy, and idol-like, with a soft pink-purple ambiance, crisp character detail, blurred cozy background, no watermark, no extra text beyond the neon sign.

Goal: Create a vertical comparison image showing the same white Toyota Celica GT-Four in three stacked photographic panels, comparing three realism styles: a real car, a real car photographed to look like a miniature, and a plastic model photographed to look like a real car.

Canvas: Portrait 3:4 composition, 768 × 1024 style, divided into exactly 3 equal horizontal panels separated by thin white divider lines. No captions or overlay text outside the car details.

Subject: A {argument name="car model" default="white Toyota Celica GT-Four ST205"} shown from a front-left three-quarter angle, low stance, white multi-spoke wheels, large rear wing, hood vent, round twin headlamps, front bumper fog lights, side decal reading “GT-FOUR,” and a Japanese license plate with the visible number {argument name="license plate number" default="22-31"}. The car should remain visually consistent across all three panels.

Setting: Nighttime urban industrial road scene in {argument name="location style" default="Japan"}, wet asphalt after rain, reflective puddles, concrete pillars, chain-link fencing, bridge or elevated roadway structures, and warm streetlights receding into the background with bokeh.

Panel 1: Top panel shows the real full-size car photographed naturally. Sharp, realistic automotive photography, crisp body details, accurate proportions, deep reflections on wet pavement, moderate background blur, cinematic night lighting.

Panel 2: Middle panel shows the same real car but photographed to look miniature. Use tilt-shift depth of field, stronger blur at foreground and background, slightly toy-like scale illusion, softened details, shiny wet road reflections, and a more diorama-like feel while still using the same viewpoint.

Panel 3: Bottom panel shows a plastic scale model of the same car photographed to look like a real full-size car. Keep the car highly realistic but subtly model-like in the surface finish and fine details; use shallow depth of field, realistic reflections, convincing low-angle automotive composition, and the same wet night street environment.

Visual style: High-quality photorealistic comparison, cinematic contrast, glossy white paint, warm orange and cool blue night tones, realistic lens bokeh, no people, no extra vehicles, no logos or watermarks. Maintain exactly 3 panels and exactly 1 car per panel.

Create a photorealistic high-angle smartphone selfie of an adult {argument name="character ethnicity" default="East Asian"} woman in a bright modern bedroom, looking directly into the camera with a soft friendly smile. She has {argument name="hair color" default="warm light brown"} hair styled in a loose low side bun with wispy bangs and face-framing strands, natural glowy makeup, brown eyes, and smooth skin. She wears a fitted {argument name="top color" default="cream white"} ribbed spaghetti-strap crop tank top, an open sheer white button-up shirt, a delicate thin necklace with a tiny pearl pendant, and high-waisted blue denim shorts. The camera is held above her at arm’s length with a wide-angle phone perspective, creating a casual selfie composition with her upper body centered and one extended arm visible in the foreground. Background: soft daylight bedroom interior with neutral walls, wardrobe doors with dark handles on the left, a bed with white bedding on the right, a mirror or glass panel edge, and gray carpet. Lighting is natural window light, bright and flattering, with shallow depth of field, realistic skin texture, subtle fabric translucency in the overshirt, crisp details, and no text or watermark. Use a vertical 3:4 portrait frame, candid social-media photo style, polished but not over-edited.

Goal: Create a cinematic character design sheet for a {argument name="character name" default="21-year-old Indigenous Mesoamerican female warrior"}, showing her costume, face paint, hairstyle, and multiple angles for concept-art reference.

Canvas: Wide horizontal 3:1 character sheet on a neutral medium-gray studio background, high-resolution photorealistic fantasy concept art with soft cinematic lighting and sharp fabric/leather detail.

Layout: Use exactly 11 visible views arranged across the canvas: 4 small stacked portrait studies on the far left, 1 large central front-facing three-quarter body portrait, 5 smaller full-body turnaround/action reference views in the middle-right, and 1 large close-up face portrait on the far right.

Subject details: The warrior is a young Indigenous Mesoamerican woman with olive-tan skin, intense pale hazel-green eyes, strong brows, full lips, and an alert, controlled warrior expression. Her {argument name="hair color" default="black"} hair is intricately braided into many tight braids with a center part, long length, braided side sections, and loose baby hairs at the hairline. She wears subtle dark obsidian war paint: a smoky horizontal mask-like band across the bridge of the nose and under both eyes, with faint red accent marks around the eyes in some portrait studies. Add natural skin texture, light sweat or dew on the close-up face, and realistic pores.

Costume: Dress her in a dark-gray fitted leather warrior tunic with sculpted paneling, high neck, layered chest armor, ribbed seams, long sleeves in muted gray woven fabric, dark bracers, black fitted pants, and dark boots. The shoulders have dramatic carved obsidian-black feather or scale-like pauldrons with glossy jagged pieces, giving a volcanic stone armor look. Overall palette: charcoal, graphite, black obsidian, muted gray fabric, and warm natural skin tones.

Exact view count and descriptions: 1) far-left top small frontal head-and-shoulders portrait looking forward; 2) far-left second small three-quarter portrait looking to the viewer's right; 3) far-left third small three-quarter portrait with sterner expression and red eye accents; 4) far-left bottom small side-profile portrait facing right; 5) large central front-facing half-to-three-quarter body view with hands on hips; 6) middle-right rear full-body turnaround view showing long braids down the back; 7) middle-right side full-body view facing left; 8) lower middle top-down full-body view looking down at the head and shoulders; 9) lower right top-down full-body view in a slightly different stance; 10) far-right large cinematic close-up of the face, cropped from forehead to upper chest; 11) include the central figure as the main hero view, larger than all reference views.

Visual style: Photorealistic AAA game character concept sheet, realistic materials, detailed leather stitching and armor paneling, clean studio presentation, subtle depth of field on the close-up, dramatic but controlled lighting, no weapons visible, no text labels, no watermark, no extra characters. Emphasize the {argument name="mood" default="alert warrior stance"} and the cultural-fantasy warrior identity without adding modern accessories.

hyper real vertical amateur iphone shot of a young woman around {argument name="age" default="21yo"} {argument name="setting" default="playing colourful arcade games at the beach at night"} laughing with girl friends, supermodel aesthethics, night time, subtle blurry shot - old money clothes but {argument name="outfit" default="tight summer outfit with crop top, denim shorts and high heels stilettos"}

Use 100 percent face refranace given strictly Portrait of a {argument name="subject" default="young Asian woman"} standing by the ocean, looking back over her shoulder at the camera with a soft, confident gaze. She has long brunette hair styled in a low ponytail with face-framing strands, wearing gold hoop earrings. She's dressed in a {argument name="outerwear" default="black-and-white vertical striped oversized blazer"} layered over a modest {argument name="top" default="high-neck magenta top"}. Soft natural daylight, ocean and sandy shore blurred in the background, shallow depth of field, fashion-editorial style, warm skin tones, crisp detail on fabric texture, professional beauty photography.

A medium close-up cinematic portrait of a {argument name="subject" default="young handsome man"} with {argument name="glasses" default="round dark glasses"} and messy dark hair with a neatly trimmed beard, holding a {argument name="camera" default="black DSLR camera"} with both hands. He wears an open black button-down shirt over a neutral grey-beige t-shirt, standing in a dark, atmospheric photography studio. Soft, natural side-lighting enters from a window on the right, illuminating his face and camera with gentle highlights and deep shadows. In the softly blurred background, studio softboxes and camera equipment add depth. Shot on a 50mm lens, shallow depth of field, muted color palette, high editorial fashion portrait quality.

Photorealistic professional photograph of two beautiful women with {argument name="hair color" default="long brunette hair"}. Both wear {argument name="clothing" default="highly reflective red form-fitting catsuits"} with front zippers. The woman on the left holds a black tactical-style rifle across her body. The woman on the right holds a black handgun at her side and is positioned lower beside her, one hand resting near the other woman’s high heel. Both look toward the camera with confident expressions. {argument name="background" default="Dark brick wall background with a graffiti colored written 'Keor' on the wall"}, dramatic rim lighting outlining the body, face, high heels and clothing textures. Highly detailed textures, realistic skin, sharp focus, cinematic lighting, professional fashion/action photography style, 8k resolution.

{argument name="subject" default="subject"} in this style={argument name="transformation" default="transformation"} - Preserve the subject’s recognizable structure, but reinterpret it as a stylized editorial composition where drawing and miniature object-world logic coexist. - Convert some forms into sparse grainy line illustration while converting other forms into soft pastel miniatures or still-life objects. - Reinterpret texture-rich or detailed areas as small clusters of plants, props, vessels, terrain fragments, or decorative micro-world motifs. - Convert smooth surfaces into matte pastel color fields, while using graphic linework to describe edges, volume, and texture selectively. - Reduce complexity into a limited color system with one dominant graphic ink tone and a few quiet pastel support colors. - Replace deep realism with poster-like clarity, paper-print tactility, and curated tabletop arrangement. - Allow the subject to feel partially drawn and partially physically staged, creating a deliberate ambiguity between illustration and object. - Use negative space and sparse composition to keep the image clean, modern, and highly readable. </transformation> <shape_language> - Favor simple containers, rounded objects, leaves, stems, pebbles, tabletop silhouettes, miniature clusters, and clean abstract forms. - Use a mix of flat contour areas and soft low-relief sculptural accents. - Keep repetition rhythmic and decorative. - Let small details feel intentional and editorial rather than busy. </shape_language> <rendering_rules> - Use visibly grainy printed or crayon-like marks for lines and texture accents. - Keep volumetric forms soft, simplified, and pastel. - Limit the palette aggressively for a designed, print-friendly look. - Use shallow spatial depth with subtle diorama hints. - Make the final result feel like an art print, a magazine illustration, and a miniature set all at once. </rendering_rules> <finish> - Graphic limited-palette editorial hybrid. - Printmaking texture fused with soft pastel mini-world staging. - Smart, modern, collectible, and quietly distinctive. </finish> <avoid> - Avoid photoreal rendering. - Avoid deep cinematic perspective or strong realism. - Avoid overstuffing with too many tiny details. - Avoid using too many colors; keep it controlled and designed. </avoid>

First, analyze **the source image (flower, clothing, pattern, etc.)**.

Instruction to ChatGPT:

"Convert this image into a JSON prompt, including size, color, shape, and details."

Instruction: "Metaprompt Goal: Create a prompt for NanoBananaPro that draws the specified texture onto a UV deployment map. Perform a process similar to Substance Painter, generating the texture by referencing image 2 against the UV map. Based on this JSON, create a prompt to change the color by referencing an additional image. Aim to create a prompt template in JSON format that generates a skirt texture on a UV map. Clearly instruct NanobananaPro to create the color based on this image. Generate a single prompt JSON according to a predefined JSON schema: ..."

Four sequential vertical portrait images (tall 9:16 aspect ratio) telling a continuous romantic story of a young Chinese couple strolling on a charming European street, photorealistic, cinematic lighting, highly detailed faces with authentic Chinese features, soft natural skin texture, elegant casual clothing.

Panel 1 (top): Early morning golden light, the Chinese couple walking side by side on a quiet cobblestone street in a historic European old town, smiling and holding hands, classic European architecture with flower boxes and cafe signs in the background, soft mist in the air.

Panel 2: Mid-morning, the couple stops in front of an old bookstore with large glass windows, the man points excitedly at a book while the woman laughs, warm sunlight reflecting on the cobblestones, authentic Chinese facial features, natural expressions.

Panel 3: Afternoon, the couple sits closely on a wooden bench under a blooming tree on the European street, sharing an ice cream, the woman resting her head on his shoulder, soft bokeh of European passersby and colorful buildings in the background, tender and intimate mood.

Panel 4 (bottom): Golden hour evening, the Chinese couple walking away down the same European street hand in hand, the man carries a small bouquet of wildflowers, warm orange sunlight casting long shadows, romantic and peaceful atmosphere, cinematic depth of field.

Consistent characters across all four panels, same Chinese couple, same clothing, seamless story progression from morning to evening, vertical composition, ultra-realistic, 8K detail, soft film color grading.

a stack of three {argument name="subject" default="Tomatoes, Eggs, and Sweatshirt and Ice creams"} with {argument name="liquid" default="LIQUID"} dripping down, on a white background, in the style of food photography, macro shot, studio lighting, nikon z6 ii, volumetric lighting, neutral tones.

Realistic candid photo of [person from uploaded photo] standing chest-deep in {argument name="water type" default="rough ocean water"}, surrounded by large breaking waves and white sea foam, no hat. In one raised hand, held above the water, they hold an {argument name="laptop" default="open dark grey/space-black laptop"} with a visible Apple logo, angled toward their face as if working on it. In the other hand, raised at chest height, they hold a {argument name="beverage" default="to-go coffee cup"} with a white lid and brown sleeve. Their expression is focused and slightly tired, looking down at the laptop screen. Only the head, shoulders, and raised arms are visible above the waterline; the rest of the body is submerged. The ocean is turbulent with rolling waves and whitewater in the background, cool grey-blue-green color grading, overcast diffused daylight, slightly moody and dramatic lighting. Shot from a distance with a slight telephoto compression, as if photographed candidly from the shore. Vertical composition, photorealistic, ironic "always working" meme aesthetic, no text overlays anywhere on the image.

{argument name="actor" default="Ana de Armas"} cosplaying as {argument name="character" default="Spider-Gwen"}

Ultra-realistic 8K portrait of a stylish young man FROM ATTACHED PHOTO, standing confidently with a VIP suitcase in front of an {argument name="airline" default="Emirates airplane"} at the airport. Passengers board in the blurred background. he wears an {argument name="top clothing" default="white linen shirt"}, {argument name="bottom clothing" default="black cargo pants"}, white sneakers, and aviators sunglasses. Stylish pose with one hand casually in her pocket. Realistic lighting, cinematic, and natural, capturing a modern, fashionable travel vibe. keep my face exactly the same as the uploaded photo.

{argument name="subject" default="Union member field critic"} with {argument name="equipment" default="cooling gear"}, airport ready.

Ultra-realistic studio portrait of a man, close-up headshot, centered composition, neutral expression, sharp facial details, textured skin, clean grooming, wearing a {argument name="top clothing" default="black turtleneck"} and {argument name="outerwear" default="dark tailored coat"}, modern minimalist fashion, monochrome black and white photography style, only the glasses in color ({argument name="glasses color" default="Translucent amber-orange"} frame and lenses), high contrast lighting, soft studio light, cinematic shadows, editorial fashion photography, plain light background, 85mm lens, shallow depth of field, ultra detailed, photorealistic, magazine quality, 8K.

Create a highly realistic, polished group selfie of three young women sitting together in a premium stadium viewing area, with a live sports field and a large audience visible behind them. 1. Subject Description, Hair, Skin, Expression & Pose: Three young women are positioned closely together in a natural, friendly group selfie composition, all facing the camera. The woman on the left has fair-to-light skin and medium-to-long light-brown hair styled in a high ponytail, with a few loose strands naturally framing her face. She wears a warm, subtle smile and looks directly into the camera with a relaxed, cheerful expression. Her posture is upright and slightly angled toward the other two women. The woman in the center has fair skin and auburn-brown hair pulled back neatly into a low ponytail. She has soft facial features, natural eyebrows, clear eyes, subtle makeup, and a calm, pleasant expression. She faces the camera directly with a relaxed posture. The woman on the right has fair-to-light skin and long dark-brown hair styled in a high ponytail with loose lengths falling naturally over one shoulder. She looks directly toward the camera with a composed, friendly expression. She stands slightly closer to the camera, creating natural depth within the group selfie. All three subjects should have realistic facial features, authentic skin texture, natural proportions, believable hands and fingers, and subtle individual differences in hairstyle and appearance. Their expressions should feel spontaneous and natural rather than overly posed. 2. Clothing & Accessories: The woman on the left wears an elegant dark-green sleeveless top with delicate light-colored decorative trim along the neckline, paired with dark trousers. She wears a simple fine necklace and small understated earrings. The woman in the center wears a fitted royal-blue sleeveless top paired with dark trousers and a subtle belt. She wears layered delicate necklaces, including a small cross-shaped pendant, creating a refined casual style. The woman on the right wears a simple white ribbed sleeveless top paired with relaxed blue jeans. She wears a delicate necklace and small understated earrings. Keep all clothing tasteful, casual, modern, and appropriate for a sporting-event setting. 3. Environment & Lighting: Place the group inside a premium stadium seating or hospitality area with comfortable dark upholstered seats. Through the large opening behind them, show a brightly illuminated green sports field surrounded by rows of spectators. The distant audience should be naturally blurred and indistinct, creating realistic depth without distracting from the three subjects. Include subtle stadium architecture, railings, glass panels, and illuminated advertising areas in the background without readable text or prominent logos. Use a combination of bright stadium lighting and soft ambient interior illumination. The subjects' faces should be evenly exposed with natural highlights and gentle shadows. Add subtle background bokeh from stadium lights to create an atmospheric evening-event feeling. 4. Camera & Style: Photorealistic professional event photography with the natural appearance of a high-quality smartphone group selfie. Vertical 4:5 composition, slightly wide-angle perspective equivalent to approximately 28–35mm, camera held at arm's length, natural perspective, realistic depth, sharp focus across all three faces, softly blurred stadium background, high dynamic range, cinematic but natural color grading, realistic skin tones, detailed hair strands, authentic fabric textures, subtle ambient light, balanced exposure, crisp details, professional editorial finish, ultra-high-resolution photographic realism.

Create an extraordinarily photorealistic cinematic photograph of a {argument name="setting" default="normal modern city at blue hour"}, captured from street level with a professional full-frame camera and a 35mm lens. At first glance, everything looks completely normal: realistic pedestrians walking, taxis and cars moving through traffic, storefronts, streetlights reflecting on slightly wet asphalt, office windows glowing naturally, subtle atmospheric haze. But hidden between the skyscrapers is something absolutely impossible: A {argument name="subject" default="colossal HUMAN-LIKE FIGURE"}, hundreds of meters tall, is partially visible behind the buildings. The figure must look almost completely human rather than monstrous — realistic human skin, subtle pores, tiny imperfections, faint veins, natural facial structure, realistic eyes, slightly messy hair, believable anatomy. Only portions of the enormous person are visible between the buildings, making it initially difficult to understand what the viewer is seeing. One enormous {argument name="interaction" default="hand is resting naturally against the side of a skyscraper"}, with individual fingerprints, wrinkles and skin texture visible. The enormous face is partially obscured by clouds and buildings in the far background, looking calmly down toward the city. CRITICAL: Make the giant blend naturally into the environment. The lighting on the giant must perfectly match the city's lighting, atmospheric perspective, haze, reflections and color temperature. Buildings in front of the figure should partially occlude it. Some atmospheric fog should pass between the giant and the camera. People on the street should NOT be dramatically reacting. Most pedestrians should appear completely unaware, creating an eerie sense that the viewer has discovered something impossible that everyone else somehow missed. Extreme photorealism, physically accurate lighting, realistic global illumination, natural shadows, realistic reflections, atmospheric depth, subtle lens imperfections, realistic skin microtexture, high dynamic range, cinematic depth of field, documentary photography aesthetic, believable scale, extremely detailed architecture. Composition should create an immediate visual mystery: the viewer should need a second look to realize that the shape between the skyscrapers is actually an enormous human being. No fantasy illustration aesthetic, no cartoon appearance, no CGI look, no exaggerated monster features, no glowing eyes, no text, no logos, no watermark. Make it look like an accidentally captured photograph from a real camera — so convincing that it could initially be mistaken for a genuine photograph. 8K-level detail, ultra-high resolution, natural colors, realistic photographic grain, masterpiece-level realism.

Transform my portrait into a whimsical {argument name="art style" default="hand-painted watercolor anime illustration"} inspired by {argument name="inspiration" default="Spirited Away"}, while preserving my exact facial features, hairstyle, skin tone, expression, and overall likeness. Soft watercolor textures, delicate ink outlines, {argument name="color palette" default="warm pastel colors"}, dreamy natural lighting, expressive anime eyes, lush magical Japanese scenery, gentle atmospheric haze, subtle fantasy details, cinematic composition, charming hand-painted imperfections, nostalgic and enchanting mood, highly detailed.

metadata:
model_context: "Corporate Tech, High-Energy Business Youtube Thumbnail mixed with Anime Art"

image_structure:
composition_type: "Diagonal Split Layout (Arrowhead Shape)"
visual_flow: "Left Text (Problem/Solution) -> Right Visual (Result/Character)"
layers:
layer_background_left:
shape: "Geometric polygon covering left 60% of screen, with a sharp arrow-point extending into the right side"
color: "Matte Black (#1A1A1A)"
texture: "Subtle carbon fiber or mesh pattern to add tech depth"
layer_typography:

location: "Centered within the left black area"
alignment: "Left-aligned or Centered"
elements:
- type: "Headline (Main)"
text: "{argument name="main headline" default="Design to Earn\nwith AI Side Jobs"}"
style: "Extremely bold, heavy Gothic font (Impact style). White text."
highlight: "Apply Bright Yellow (#FFD700) color to the word '稼ぐ' (Earn) and 'AI'"
- type: "Sub-Headline (Detail)"
text: "{argument name="sub headline" default="Complete Guide to Nano Banana Pro"}"
style: "Smaller white sans-serif text, clean and readable"
- type: "Body Text (Explanation)"
text: "{argument name="explanation text" default="From beginner to professional quality.\nCourse aiming for monetization via the shortest route."}"
style: "Small, thin grey or white text block, justified layout"
layer_cta_button:
shape: "Rounded Pill Shape"
color: "Bright Yellow (#FFD700)"

text: "{argument name="CTA text" default="Apply Now ▶"}"
text_color: "Black"
location: "Bottom left, below the text block"
style: "Flat design with slight drop shadow"
layer_visual_right:
shape: "Fills the remaining right side (behind the arrow overlap)"
content: "Japanese Kawaii Anime Girl illustration"

style: "2D Cel-shaded anime art, high detail, vibrant colors"
details:
subject: "A cute anime girl with a futuristic / cyber-fashion look"
appearance:
hair: "Pastel pink and blue gradient bob cut, with cyber hair clips"
eyes: "Large sparkling anime eyes, confident smile"
clothing: "Stylish tech-wear hoodie with yellow accents to match the background theme"
pose: "Dynamic pose, pointing finger towards the text on the left or giving a thumbs up"
background_context: "Simple light grey or white background with abstract pop patterns (dots/stars) to make the character stand out"
style_and_tone:
art_style: "Modern Corporate Tech Design mixed with Japanese Anime aesthetics"
color_palette:

primary: "Black, Yellow"
secondary: "Pastel Pink/Blue (Character), White"

Ultra-photorealistic high-fashion cover portrait of {argument name="subject" default="Sadie sink"}, wearing an jaw-dropping floor-length {argument name="dress color" default="crimson red"} silk gala gown with an exaggerated plunging sweetheart neckline, a dramatic thigh-high slit, and sculpted sharp shoulder accents, striking a powerful runway pose, looking directly into the camera with a captivating, fierce gaze, sleek wet-look Hollywood waves, vibrant high-contrast studio portraiture, deep {argument name="backlight color" default="emerald velvet"} backlighting creating a dramatic rim-light along her silhouette, forensic skin texture, micro pore detail, luxury Vogue editorial cover aesthetic, high-glamour specular highlights, full-body composition, vertical 4:5, subtle "OzairAI" watermark in the bottom right corner.