프롬프트 · 창작 탐험

프롬프트 갤러리

실제 프롬프트와 생성 예시를 탐색하세요. 다음 창작 임무를 위한 영감을 찾고 전체 프롬프트를 무료로 복사하세요.

예시
10,974
모델
17
출처
1277
10,974 개의 프롬프트 예시 · 페이지 85 / 229
Example generated from “Watercolor Travel Poster Style — Reference Image Required”
GPT Image 2en

Transform the uploaded photo into a beautiful minimalist editorial illustration with a soft, hand-painted watercolor and gouache aesthetic. ART STYLE: Vintage travel-poster illustration, delicate watercolor washes, subtle gouache brushstrokes, visible handmade paper grain, soft pigment textures, imperfect painted edges, muted pastel colors, and a dreamy romantic atmosphere. COMPOSITION: Vertical 2:3 poster. Place the main subjects in the lower third of the frame, leaving approximately 65–70% of the upper area as clean, warm ivory paper with subtle natural texture. Keep the composition airy, peaceful, understated, and balanced. SUBJECT: Preserve the recognizable identity, hairstyle, clothing, pose, and emotional connection of the people in the reference photo. Simplify their features into elegant hand-painted shapes with minimal facial details and natural proportions. Maintain a candid, intimate, storytelling feel. BACKGROUND: Illustrate the surroundings using loose, impressionistic brushstrokes and simplified silhouettes. Famous landmarks, landscapes, trees, buildings, or cityscapes should remain recognizable but understated. Use soft navy blue, dusty blue, warm beige, muted blush pink, and gentle golden yellow accents. LIGHTING & TEXTURE: Soft natural daylight, delicate shadows, subtle color bleeding, layered watercolor washes, matte finish, and authentic textured paper. Keep details concentrated around the subjects and lower background. OVERALL MOOD: Romantic, nostalgic, peaceful, poetic, timeless, and intimate—like a page from a beautifully illustrated travel journal. STRICT RULES: Minimalist composition, generous negative space, no photorealism, no 3D rendering, no harsh outlines, no oversaturated colors, no unnecessary details, no borders, no typography, and no watermark. The final image must look genuinely hand-painted on textured paper.

Example generated from “Weeping Angel Statue in Cemetery”
GPT Image 2en

A photorealistic, vertical portrait of a weathered bronze statue of a woman sitting on a stone bench in an old cemetery. The statue is heavily oxidized with a vibrant green patina and dark brown tones. She wears a long, flowing dress that drapes over her legs and the sides of the seat. Her head is tilted back toward the sky, eyes closed, with a veil covering her hair, conveying a sense of deep sorrow or longing for release. Her hands rest gently on the edges of the stone slab behind her. The background features blurred trees with hints of autumn foliage and mossy ground, creating a moody, atmospheric setting with soft, natural lighting.

Example generated from “Marble Statue of Kneeling Woman with Rose Crown”
GPT Image 2en

A stunning, high-resolution photograph of a classical white marble statue depicting a kneeling woman in a pose of deep sorrow or prayer. The figure is draped in flowing robes and a long, translucent veil that cascades down to the base, creating an ethereal effect where the fabric appears almost like sheer silk. Resting on top of her head, over the veil, is a crown of dark red roses, providing a striking color contrast against the monochromatic stone. The statue's face is serene yet melancholic, with eyes cast downward. The lighting is dramatic chiaroscuro, spotlighting the statue from above against a pitch-black background, emphasizing the smooth texture of the marble and the delicate folds of the drapery.

Example generated from “Palmistry Infographic Layout — Entertainment Example”
GPT Image 2en

Goal: Create a vertical infographic-style palmistry guide titled {argument name="title text" default="你的个人指南"}. Canvas: A clean, modern layout with a white background and black text, divided into three main columns. The style is minimalist and educational, resembling a high-end wellness or self-improvement app interface. Layout: - Left Column: Top section has the title "你的个人指南" in large bold characters, followed by the subtitle "认识自己 · 专注成长 · 持续进化". Below this is a realistic photo of an open right hand (palm facing forward). Further down are sections for "你的手" (Your Hand) listing details like dominant hand, element, shape, and finger traits; "快速总结" (Quick Summary); "优势" (Strengths) as a bulleted list; and "关键关注点" (Key Focus Points) as another bulleted list. - Center Column: Top section shows a detailed black-and-white fingerprint scan image labeled "掌印图" (Palm Print Image). Below that is a line-drawing diagram of a hand labeled "主要掌纹图" (Main Palm Lines Diagram), with numbered lines (1-5) corresponding to specific palm lines. Underneath is a bar chart section "手型类型" (Hand Type) showing percentages for Fire, Wind, Earth, and Water elements. At the bottom is a legend for "掌丘" (Mounts) explaining their meanings. - Right Column: Contains five distinct cards detailing specific palm lines: 1. 感情线 (Heart Line), 2. 智慧线 (Head Line), 3. 生命线 (Life Line), 4. 命运线 (Fate Line), and 5. 太阳线 (Sun Line). Each card includes a description and interpretation. Below these is a final advice box "最终建议" and a notes section "备注". Visual Style: Use thin black borders for cards. Icons should be simple line art (e.g., heart, brain, shield). Text should be crisp sans-serif Chinese characters. The fingerprint image should have high contrast with visible ridges. Text Content: Ensure all Chinese labels match the provided image exactly, including headings like "手相解读", "掌印图", and the specific interpretations for each line.

Example generated from “Whale Tail Photo & Ink Illustration Diptych”
GPT Image 2en

Create a vertical diptych-style composition split into two distinct halves. The top half is a photorealistic, high-resolution photograph of a humpback whale's tail fluke rising vertically out of the deep blue ocean. The water is textured with white foam and splashing droplets around the base of the tail. In the background, there are soft-focus blue mountains under a bright sky with scattered clouds. The bottom half is a minimalist illustration on a pale blue-grey paper texture background. It features a stylized black ink drawing of the same whale tail shape, but rendered entirely using a pointillism or stippling technique (composed of thousands of tiny dots). Include specific Korean typography in the bottom section: 'No. 027' in small text at the top left, followed by a large title '먼바다로 가는 꼬리' and a smaller subtitle '먼 곳을 향한 꼬리'. At the very bottom center, include the scientific name 'Megaptera novaeangliae · 해양 포유류' in a serif font.

Example generated from “Photo-to-Watercolor — Two-Recipe Pack (Toddler Preview) — Reference Image Required”
GPT Image 2en

Create a charming photorealistic-to-watercolor transformation using the reference image. Show a cute toddler wearing a yellow sun hat, a light blue checkered romper, yellow socks, and dark sandals, crouching curiously and interacting with three white ducks on a quiet country road. Include lush green grass along the roadside, distant residential buildings, soft daylight, and a peaceful countryside atmosphere. In the lower half, transform the same scene into a delicate hand-painted watercolor illustration on textured cream-colored paper. Preserve the toddler’s pose, outfit, the ducks’ positions, and the original composition. Use soft watercolor washes, fine pencil outlines, subtle pastel blue and sage-green accents, loose brushstrokes, and minimal hand-drawn grass details. Elegant minimalist storybook aesthetic, warm paper texture, generous negative space, heartwarming mood, high detail, vertical 4:5 composition. Create a beautiful photorealistic-to-watercolor transformation using the reference image. Show an adorable fluffy golden-colored dog sitting in a lush green meadow, looking directly at the camera with bright expressive brown eyes, a happy open mouth, and its pink tongue hanging out. Capture its soft, fluffy fur, rounded ears, bushy tail, and friendly personality in warm natural sunlight, with a softly blurred countryside background. In the lower half, transform the same dog into a delicate watercolor painting on textured cream-colored paper. Preserve the dog’s facial features, fur color, expression, and overall silhouette. Use loose golden-beige watercolor washes, delicate pencil outlines, soft pigment splashes, subtle sage-green grass strokes, and light blue brush accents. Minimalist handmade illustration, charming children’s storybook style, natural paper grain, soft pastel palette, generous negative space, cozy and heartwarming atmosphere, vertical 4:5 composition.

Example generated from “Length Comparison”
GPT Image 2en

A horizontal triptych comparison image with a dark navy background, divided into three equal vertical panels. Each panel displays a photorealistic 3D render of a fantastical 'Space Library' scene, illustrating the difference in prompt complexity. Left Panel: Labeled "5 WORDS" at the bottom in white sans-serif text. The image shows an exterior wide shot of a grand, ornate library building floating in deep space. It features a large glass dome revealing a galaxy interior, surrounded by nebulae and stars, with a reflection on a water-like surface below. Center Panel: Labeled "25 WORDS" at the bottom. The image is an interior perspective looking out through a massive curved window at a giant ringed planet (resembling Saturn). Inside, there are tall wooden bookshelves lining the walls, a cozy leather armchair, a small table with a lamp, and a patterned rug on a polished floor. Right Panel: Labeled "100 WORDS" at the bottom. The image is a close-up interior detail shot featuring a wooden desk with an open book, a brass banker's lamp, and a globe. Behind it, large arched windows look out onto the same ringed planet and starry sky, with warm lighting contrasting against the cool blue space background.

Example generated from “Hand-drawn Infographic Prompt”
GPT Image 2zh

两个纵向相连的圆角矩形面板由极粗、略带起伏的黑色手绘轮廓线锚定,白色卡扣横跨面板边缘,构成醒目的成对模块与清晰的纵向节奏。内容依托格式塔邻近性与版面分区进行组织,每个面板内部划分为稳定的左图右文双栏,左右视觉比重均衡,标题与信息在右栏自上而下阅读;整体视觉重心落在两块面板的交界处及各自醒目的大标题上,页眉区域保持独立,放置 logo 与主标题。图文通过并置与衬底相互衔接:左侧预留给角色或主题插画,置于不规则的圆角浅色干净底座之上;右侧以醒目的大字引领多行说明文字,并在文本旁搭配小型线描符号以提示内容。文字层级始终分明,大标题粗重醒目,说明文字较细且行距宽松,小字与符号形成节奏点;字形结构端正易读,字怀清晰、收笔利落、字面紧凑,与手绘图像相得益彰,无论文字多少均维持该层级;文字不足时,则根据主题补充次要小字。文本边缘平整清晰,在手绘线条之外形成干净的排版层。留白涵盖面板间距、图像底座周围的衬底留白、文本行间的阅读间隙以及页眉与正文之间的停顿;各处均保持宽裕,使粗边框、插画与标题得以清晰凸显。插画内容由所附主题提供,在左栏内采用完整、集中、略微倾斜的手绘构图,主体轮廓简洁,动态细节轻盈,承重部位端正;内容密度约占半幅画面,主要集中在各面板的左栏,其余部分为白色或浅色圆角底座、平整的面板色块及底部留白,不生成任何场景。采用复合技法:以连贯、略带抖动的粗黑线条勾勒边框与角色;以干净平滑的数字平涂铺展大面积面板与局部插画色彩;以利落的文字排版压住信息区;整体呈现出由手绘线条、电脑平涂与清晰排版合成的插画编排效果。配色保持高明度基调,色彩鲜活、饱和而不刺眼,最浓烈的色面落在成对面板内,并与角色局部色彩相呼应;暖调浅色背景与不同明度的浅色色面相互搭配,以强烈的黑白线文对比稳定层次。整体温暖明快、亲切有活力,宛如哑光数字色块与密集马克笔线条的结合;色调随主题变化,但明暗、粗细、饱和度、偏暖倾向以及欢快亲切的心理感受不受主题影响,始终保持这一性格。整体气质轻盈、友好且有序。所附主题、准确文案、用途与宽高比优先;上下面板分组、左图右文、粗黑手绘边框、图像衬底、清晰的文字层级以及高明度暖厚配色方案保持不变。 ------ 主题:水循环科普海报

Example generated from “Kiyomizu-dera Travel LP with Anime Mascot”
GPT Image 2en

Goal: Create a comprehensive, high-density vertical travel landing page (LP) for Kiyomizu-dera in Kyoto. Canvas: Vertical aspect ratio (approx. 4:5), featuring a vibrant mix of anime illustration and realistic photography with a clean white background accented by orange and purple brushstrokes. Layout: 1. Header: Top navigation bar with logo 'あぽろ塾 APOLLO JUKU' on the left, menu items (Search, Model Course, Spot, Gourmet, Access) in center, and an orange CTA button '今すぐ旅に出る' on right. 2. Hero Section: Large title '清水寺 Kiyomizu-dera' in bold black serif font. Subtitle '京都・東山の絶景と歴史をめぐる'. A central anime character (orange hair, cat ears, streetwear jacket with 'KYOTO' text) pointing right, accompanied by a tuxedo cat. Background shows a panoramic view of Kyoto with autumn foliage and pagodas. Floating handwritten text: '見たことのない京都に、きっと出会える。' and 'ここにしかない景色がある。'. Location info block: '京都・東山', '早朝散策がおすすめ', '春は桜、秋は紅葉が人気'. World Heritage badge and coordinates '34.9945, 135.7843'. 3. Highlights Section ('見どころ'): Horizontal scrollable cards showing photos of '清水の舞台', '音羽の滝', '三年坂・二年坂', '京都の街並みを一望', '世界遺産'. Each card has a photo, title, and short description. 4. App & Features Section: Left side features a smartphone mockup displaying a map app interface with pins for Kiyomizu-dera, Otowa Waterfall, etc. Center column lists 'アプリの主な機能' (App Main Features) with icons: Spot Search, Model Course, Favorites, Offline Map. Right side is 'モデルコース' (Model Course) with numbered steps 01-04 showing route thumbnails. 5. Seasonal Recommendations ('おすすめの季節'): Four distinct seasonal blocks (Spring, Summer, Autumn, Winter) with corresponding landscape photos and brief text descriptions. 6. Access Section ('アクセス'): Bottom right area listing transport options (Train, Bus, Car) with specific station names and walking times. Includes a small silhouette of a pagoda and 'Kyoto' script. Visual style: Modern Japanese web design aesthetic. High contrast between the colorful anime mascot and the serene real-world photography. Use of orange (#FF5722 approx) and purple accents. Clean sans-serif typography for body text, elegant serif for headings. Text content: Preserve all Japanese text exactly as shown in the image, including section headers like '見どころ', 'モデルコース', 'アクセス', and the specific location names.

Example generated from “Swiss Editorial Diptych with Texture Zones — Reference Image Required”
GPT Image 2en

Create one vertical 3:4 diptych for each uploaded photo, divided into equal top and bottom halves with a seamless join. No divider, frame, or collage. Swiss editorial design, experimental photography, and vintage print aesthetics. TOP HALF — ORIGINAL PHOTOGRAPH Preserve the original identity, pose, scene, composition, perspective, lighting, and colors. Allow only subtle color grading and background extension. No text. BOTTOM HALF — SEVEN EXPERIMENTAL TEXTURE ZONES Repeat the entire original photo. Overlay seven asymmetrical rectangular zones of different sizes directly onto the image pixels, crossing non-critical details. Maintain visual continuity inside and outside every zone. Never crop, zoom, shift, duplicate, or redraw the image. Keep all important focal features visible. SEVEN NUMBERED TECHNIQUES 01. Reeded Glass: Vertical optical refraction. 02. Black-and-White Engraving: Fine etched lines and crosshatching. 03. American Stipple: Black-and-white microdot texture. 04. Vintage Halftone: CMYK printing dots. 05. Pixel Mosaic: Square pixel blocks. 06. Transparent Overlay: 2–4 translucent overlapping layers. 07. Duotone Screen Print: Two-color screen-printing effect. DESIGN DETAILS Add hairline borders, tiny corner squares, and fine leader lines to identify each rectangular zone. Label every technique with its corresponding number and English name. Keep all areas outside the seven zones photographic and realistic. STRICT REQUIREMENTS No logos, watermarks, extra images, or unnecessary decorative elements. Preserve the original photograph's integrity and maintain a clean, sophisticated editorial layout. OUTPUT: One high-definition 3:4 vertical poster with equal top and bottom halves.

Example generated from “Anime Character Sheet: Crying Girl”
GPT Image 2en

A high-resolution anime character design sheet for a 3-year-old girl, rendered in a soft pastel aesthetic. The layout is divided into sections: Top Left: A large close-up of the crying girl with brown bob hair and red eyes, wearing a pink floral dress, reaching out her hands. Label '3歳' (3 years old) and '災害シーン用キャラクターシート' (Character Sheet for Disaster Scenes). Top Right: Four smaller expression panels labeled '泣く' (Crying), '怯える' (Frightened), '驚く' (Surprised), and '普段の笑顔' (Usual Smile). Bottom Left: A full-body turnaround showing three views labeled '正面' (Front), '側面' (Side), and '背面' (Back). She wears a tiered ruffled dress with a cherry blossom pattern, white socks, and pink sneakers. Bottom Right: Detail close-ups labeled '瞳ディテール' (Eye detail), '口ディテール' (Mouth detail), '髪飾り' (Hair accessory - flower clip), '服ディテール' (Clothing detail - bow), '靴ディテール' (Shoe detail), and '生地の柄' (Fabric pattern).

Example generated from “Cyber Stage Character Concept Sheet”
GPT Image 2en

Create a comprehensive character concept sheet for {argument name="character name" default="Cyber Stage Yuna"}, featuring a futuristic idol warrior design. The layout should include: 1. **Main Portrait**: A full-body illustration of the character in an elaborate cyberpunk stage outfit with holographic elements, crystal accessories, and flowing fabric accents. 2. **Outfit Showcase**: Multiple views (front, side, back) showing different angles of the costume design. 3. **Expression Sheet**: Nine facial expressions including grinning, confident, tech-savvy, stage trouble, serious, determined, wink, joyous, and smirk. 4. **Detail Close-ups**: Sections highlighting boots & neckline, shoulder armor, back detail, eye detail, and hair detail. 5. **Materials & Textures**: Swatches showing holographic silver, pearl white mesh, electric blue crystal, and neon lavender accents. 6. **Accessories & Weapons**: Display of halo crown, crystal lance variations, holographic cape, and soundwave earpiece. 7. **Combat Poses**: Silhouette illustrations showing dynamic action stances. 8. **Weapon Effects**: Attack animations including morning strike, piercing thrust, and ultimate digital harmony effects. 9. **Concept Overview**: Key elements section with color palette swatches and thematic icons. 10. **Style Notes**: Description emphasizing futuristic glow, stage idol energy, techno & dreamy aesthetics. 11. **Character Summary**: Role as idol/healer, element as hydro/digital, personality as confident/iconic. 12. **Special Quote**: "When melody becomes your blade, even the tides will kneel." The overall aesthetic should blend cyberpunk technology with ethereal fantasy elements, using a color scheme dominated by deep blues, purples, silvers, and glowing accents. Include a five-star rating indicator and maintain professional game art documentation standards throughout.

Example generated from “Quiet vs Abandoned Train Station Comparison”
GPT Image 2en

A split-screen comparison of two train station scenes at dusk, separated by a thin vertical line. Left panel: A quiet, well-maintained train station named "RIVERDALE" with warm yellow lighting under the platform canopy, a wooden bench, a round clock on the wall, clean concrete platform with yellow tactile paving edge, curved railway tracks leading into the distance, signal lights glowing red, and a vibrant sunset sky with orange and purple hues over distant hills. Right panel: An abandoned, dilapidated train station named "RAVENSBROOK" with peeling paint, broken windows, overgrown weeds and grass between cracked concrete platforms, rusted railway tracks, weathered wooden support beams, dark and eerie atmosphere, similar sunset sky but more muted and cloudy. Bottom labels: "QUIET" in white text on dark background beneath left image, "ABANDONED" in white text on dark background beneath right image.

Example generated from “Haiku About Neglected Wine Cellar”
GPT Image 2en

the abbot has neglected the wine cellar and the rats content to eat rancid grapes have the run of the place

Example generated from “Chainsaw Man Makima Character Design Sheet — Reference Image Required”
GPT Image 2en

Goal: Create a detailed anime character design sheet for Makima from Chainsaw Man. Canvas: Vertical layout, clean white background with black text and line art. Layout: Central full-body illustration of the character. Surrounding panels detail specific costume elements, physical traits, and profile information. Subject details: Makima stands in the center wearing a long-sleeved white button-down shirt tucked into high-waisted black trousers, a black necktie, and a long black coat draped over her shoulders like a cape. She has red hair braided at the back with bangs framing her face, golden eyes with distinct rings, and black leather shoes. Her expression is calm and composed. Sections: - Top Left: Title "Makima" (真纪真) and subtitle "Character Setting / Costume Breakdown". - Top Right: Profile box listing Affiliation (Public Safety Devil Extermination Bureau), Identity (Divine of Control), and Ability (Control). - Left Column (Costume Breakdown): Panels for Shirt, Necktie, Suit Jacket, Suit Pants, Leather Shoes, and Historical Reference (showing photos of women in suits). - Right Column (Physical & Props): Panels for Hairstyle (front/side/back views), Eye Color (close-up of golden eye), Weapon/Chain (illustration of chains), Historical Reference (chains/handcuffs), Accessory (earring/hand gesture), Prop (contract document and stamp), Color Palette (swatches of red, beige, black, gold), and Keywords (Control, Rationality, Calm, Mystery, Danger). Visual style: High-quality anime illustration mixed with technical fashion sketching. Clean lines, precise shading on clothing folds, and elegant typography. Text content: Include bilingual labels (English/Japanese) for each section header as seen in the reference image.

Example generated from “Point Nemo Infographic Layout — Unverified Sample Labels”
GPT Image 2en

Goal: Create an educational infographic explaining {argument name="topic" default="Point Nemo, the most remote place on Earth located in the ocean"}. Canvas: Vertical portrait layout with a deep navy blue background and a subtle starry night sky texture at the top. The style is clean, modern, and informative, using gold and white text for high contrast. Layout: 1. **Header:** Large serif title "THE MOST REMOTE PLACE ON EARTH IS IN THE OCEAN" (with "IN THE OCEAN" highlighted in light blue). Below it, a small introductory paragraph defining Point Nemo as the location farthest from land, mentioning the distance of 2,688 km (1,671 miles). 2. **Central Map:** A stylized world map centered on the South Pacific Ocean. Continents are rendered in muted earth tones (greens, browns) against dark blue oceans. Labels include NORTH AMERICA, SOUTH AMERICA, EUROPE, ASIA, AFRICA, AUSTRALIA, and ANTARCTICA. 3. **Data Visualization:** A bright yellow dot marks "POINT NEMO" in the ocean south of Easter Island. Three dashed yellow arrows radiate from this point to the nearest landmasses, labeled with distances: "2,688 km (1,671 miles)" to Antarctica, "2,754 km (1,711 miles)" to South America, and "2,966 km (1,843 miles)" to Australia. 4. **Info Box (Left):** A rectangular box with a compass icon containing the coordinates "48°52.6'S 123°23.6'W" and the nickname "oceanic pole of inaccessibility." 5. **Bottom Section:** A section titled "WHY IS IT INTERESTING?" with three columns featuring icons and text: - Icon: Satellite dish. Text: "It's the farthest point from any land, making it the most isolated place on the planet." - Icon: Space station/satellite. Text: "Spacecraft and the International Space Station use it as the safest place to reenter or drop objects." - Icon: Ship. Text: "It's so remote that the closest humans are usually astronauts — not sailors." 6. **Footer:** A concluding statement next to a globe icon: "Point Nemo shows us that even in our well-explored planet, there are places more remote than anywhere on land." Visual Style: Professional infographic design. Use clear sans-serif fonts for body text and serif fonts for headings. Ensure the map projection clearly highlights the isolation of the central point.

Example generated from “Simulated Thermal Camera Image”
GPT Image 2en

A realistic thermal imaging photograph of a person standing in an indoor environment, captured with a FLIR camera. The subject is wearing a long-sleeved shirt and pants, rendered in the classic rainbow color palette where hot areas (face, hands, feet) appear white/yellow/red and cooler areas (clothing, background) appear blue/purple. In the top left corner, there is digital overlay text reading '2026-10-10 14:23:51' and 'CAM 01 THERMAL'. In the top right corner, the white 'FLIR' logo is visible. On the far right edge, a vertical temperature scale bar ranges from 20.0°C at the bottom to 40.0°C at the top. The background shows a dimly lit room with a plant on the left and a door frame on the right.

Example generated from “Tsing Hua Arts Festival — Historical Poster Study”
GPT Image 2en

Create a split-screen graphic design poster for the '2025 Spring x Tsing Hua Arts Festival'. The left side features a clean white background with vibrant, flat vector-style illustrations of flowers including a large daisy, pink lilies, blue hydrangeas, and purple lilacs. Vertical Chinese text on the far left reads '2025 春之清華藝術節' alongside English text 'Spring x Tsing Hua Arts Festival'. The bottom left contains four distinct event categories: '音樂會暨大師班', '創作展覽', '藝術工作坊', and '研討論壇', each with specific dates (4/15-5/28) and small QR codes. The right side displays a soft-focus, dreamy version of the same floral arrangement against a warm, hazy cream background. Large serif typography at the top reads '2025 Spring x Tsing Hua Arts Festival', accompanied by smaller text 'ART CULTURE PEOPLE IN BLOOM'. The bottom right shows the dates '4.15 — 5.28' and the logo 'TSING HUA UNIVERSITY'.

Example generated from “Japanese Game Dev Banner”
GPT Image 2en

A horizontal banner image with a light beige background. On the left side, there is Japanese text in purple and yellow: '制作記録 #2 敵キャラ編' (Production Record #2 Enemy Character Edition), followed by large bold text '生成AIの「あるある」がモンスターに!' (Generative AI's "Common Occurrences" become monsters!), and smaller text below 'ジェネコの冒険 ~ノイズの大迷宮~ 敵キャラ制作の2日間' (Geneko's Adventure ~ The Labyrinth of Noise ~ Two Days of Enemy Character Creation). To the right of the text, there are two small chibi characters: a girl with brown hair wearing glasses and a school uniform holding an open book, and a blue cat-like creature with pink ears reading over her shoulder. The rest of the image is filled with a collage of various colorful monster designs arranged in rows. Top row includes a shiny green sphere with eyes, a muscular blue furry beast with glowing purple eyes, a purple insectoid robot with green wings, and a multi-limbed robotic figure. Middle section features a red beetle-like creature with warning signs on its shell, a wizard character wearing a rainbow-colored pointed hat and robe holding a staff, a white moth-like creature with blue eye patterns on its wings, a black spider-tank hybrid with red lights, and a pink abstract shape resembling a face or symbol. Bottom row shows a green goblin-like creature with glowing yellow eyes, a brown spiky turtle-like monster, a pale humanoid figure with blank facial features, a split-faced doll character (half pink/half grey) holding a knife, and a purple blob monster with angry eyes.

Example generated from “Colossal MPD Robot in Rainy Tokyo”
GPT Image 2en

A colossal, bipedal police robot towering over a rainy Tokyo street. The robot is constructed from heavy industrial machinery parts, featuring exposed gears, hydraulic pistons, and thick black cables connecting segmented white armor plates. It has no human face; instead, its head is a rectangular sensor array with multiple camera lenses. Red emergency lights flash on its shoulders and chest. The text "警視庁 MPD" (Metropolitan Police Department) is stenciled in black on the white armor plating of its torso and thighs. In the foreground, wet asphalt reflects the red and blue lights of surrounding police cars and traffic signals. Several small human police officers in dark tactical gear stand near the robot's feet, emphasizing its massive scale. A blue road sign indicating directions to "Tokyo Sta." and "Ginza" is visible on the right. A helicopter hovers in the grey, overcast sky above.

Example generated from “Abandoned Road Under Bridge”
GPT Image 2en

Create a photorealistic image of an abandoned road at night. The scene should be dark and eerie, with the road curving gently into the distance. On the left side, there is a tall metal utility pole with wires extending from it. A concrete retaining wall runs along the right side of the road, covered in patches of moss and overgrown weeds. Above the wall, there is a large concrete bridge structure supported by thick pillars. The ground is wet, reflecting some light, and scattered with dead leaves and debris. Sparse tufts of dry grass grow through cracks in the asphalt. The overall atmosphere is desolate and mysterious, typical of urban decay.

Example generated from “Giant Robot Police Officer in Tokyo”
GPT Image 2en

A cinematic, photorealistic wide shot of a bustling Tokyo street at dusk. The central subject is a massive, bipedal police robot standing approximately three stories tall. The robot features an industrial design with heavy-duty black mechanical joints and white armored plating. It has a rectangular head unit equipped with multiple camera lenses and red emergency lights on its shoulders. The text "警視庁 MPD" is clearly printed in white on the robot's chest plate and shoulder guards. In the foreground, a drunk Japanese salaryman in a dark suit and loosened tie is stumbling backward while holding a beer can in one hand and taking a selfie with his smartphone in the other. To the right, two human police officers wearing dark uniforms with "警視庁 MPD" on their backs stand near a metal guardrail, looking up at the machine. The background includes warm yellow streetlights, blurred pedestrians, a blue road sign pointing to Nihonbashi, and the illuminated Tokyo Tower rising against a purple and orange twilight sky.

Example generated from “Close-up Portrait with Matchstick Light”
GPT Image 2en

Create an ultra-realistic, cinematic close-up portrait of a beautiful young woman with long, messy, flowing dark hair, thick eyebrows, long eyelashes, and slightly parted lips. Her eyes are gently closed, with a calm, dreamy, emotional expression. She is holding a lit matchstick near her face with her right hand. The warm orange and golden flame light illuminates one side of her face, nose, lips, neck, and fingers, while the other half remains in deep, soft shadows. Loose strands of hair frame her face naturally. Dark moody atmosphere, dramatic chiaroscuro lighting, realistic skin texture, detailed facial features, artistic photography, warm amber tones, soft shadows, shallow depth of field, fine details, 4K quality, vertical portrait, 9:16 aspect ratio. Preserve natural facial proportions and create an elegant, mysterious aesthetic.

Example generated from “Giant Robot Traffic Control”
GPT Image 2en

A cinematic, photorealistic night scene in Tokyo featuring a colossal bipedal police robot towering over a busy intersection. The robot is constructed from heavy-duty black and white mechanical armor plating with exposed hydraulic joints and pistons. Its head consists of a horizontal sensor array with multiple camera lenses instead of a face, and it has glowing red warning lights on its shoulders. The text "警視庁 MPD" (Metropolitan Police Department) is printed in white on the robot's chest, shoulder pads, and legs. The robot is actively directing traffic, with one hand raised high in a stop signal and the other extended to the side. In the foreground, wet asphalt reflects the city lights, where several orange and white traffic cones are placed. A human police officer wearing a dark blue uniform, a white helmet, and a reflective vest with "MPD" on the back stands with his back to the viewer, holding a glowing red baton. To the left, a white Japanese police car with flashing red and blue lights is partially visible. The background features a dense urban skyline with illuminated skyscrapers and the iconic Tokyo Skytree glowing blue against the night sky. Overhead highway signs indicate directions for Ginza, Ueno, and the Shuto Expressway.

Example generated from “Cinematic Fried Chicken Food Advertisement”
GPT Image 2en

GOLDEN CRUNCH | THE ULTIMATE FOOD CINEMA. Create an ultra-photorealistic, world-class cinematic food advertisement featuring a beautiful young adult woman with naturally expressive features, soft wavy brunette hair, and a cozy premium beige knit sweater, seated in an elegant, warmly lit modern dining room. Capture the irresistible split-second moment as she brings a perfectly golden, extra-crispy fried chicken drumstick toward her open mouth, her expression filled with genuine excitement and anticipation. In the foreground, a luxurious white ceramic platter overflows with freshly fried chicken pieces featuring exceptionally detailed, rugged crunchy breading, rich golden-amber tones, delicate crisp flakes, and subtle natural oil highlights. **CINEMATIC EXCELLENCE:** Shot on an ARRI Alexa 65 with an 85mm cinematic portrait lens, shallow depth of field, exquisite focus on the woman's face and hero chicken piece, creamy optical bokeh, sophisticated warm tungsten practical lighting, soft directional key light, delicate rim lighting across her hair, natural skin pores, realistic hands, anatomically correct fingers, lifelike eyes, authentic textures.

Example generated from “Sister of Battle Rising”
GPT Image 2en

Her blessed chainsword lies shattered, she is bleeding, surrounded and outnumbered. Yet beneath the burning gaze of Saint Celestine, this Sister of Battle rises with her boltgun roaring and the Emperor's name on her lips.

Example generated from “Horus Heresy Final Duel Scene”
GPT Image 2en

The final duel of the Horus Heresy. Aboard the Vengeful Spirit in orbit above a burning Terra, the Emperor of Mankind faces his favored son, Warmaster Horus, now fully consumed by Chaos.

Example generated from “Lady Death Fighting Demons”
GPT Image 2en

Brian Pulido’s Lady Death (Hope) cutting through several horned demons and lesser fiends.

Example generated from “Silhouette on Mountain Peak at Cosmic Sunset”
GPT Image 2en

A majestic, surreal landscape featuring a lone silhouette of a person standing on a jagged rocky peak in the foreground. The figure has one arm raised triumphantly toward the sky. Above them, the horizon curves dramatically like the edge of a planet, separating a sea of clouds below from an epic sky above. The sky is split into two distinct zones: the upper portion is a deep blue starry night with swirling nebulae and constellations, while the lower band glows with intense golden-orange sunset light, creating a fiery contrast. Below the curved horizon, layers of fluffy white and grey clouds drift among distant mountain ranges, illuminated by the warm glow of the setting sun. The overall atmosphere is awe-inspiring, cinematic, and fantastical, blending elements of space exploration with terrestrial beauty.

Example generated from “Realistic Anime Warrior in Medieval Battle”
GPT Image 2en

A hyper-realistic, cinematic action shot of a young Japanese female warrior in the midst of a chaotic medieval battle. She has long black hair tied back with two gold hair clips and wears a modified dark blue sailor-style school uniform (seifuku) reinforced with heavy leather armor on her shoulders and arms. Her face is dirty with soot and blood splatters, showing an intense, determined expression as she lunges forward. In the foreground, blurred zombie-like creatures are attacking. The background depicts a burning village with wooden structures, smoke rising into the sky, and other soldiers fighting amidst the chaos. The lighting is dramatic and gritty, emphasizing the realism of the scene.

Example generated from “Minecraft Character in Sunflower Field”
GPT Image 2en

A hyper-realistic, cinematic close-up shot of a Minecraft-style character standing in a vibrant field of sunflowers during a golden sunset. The character is constructed from blocky voxels, featuring dark skin, a full black beard, and wearing a green cap with a white t-shirt and green pants. He is holding a large, blocky bee gently against his face. Surrounding him are numerous other flying bees with translucent wings, rendered with realistic motion blur and depth of field. In the background, there is a pixelated rainbow arching across an orange sky filled with blocky clouds, a distant tree line, and a small white pickup truck on a dirt road. The lighting is warm and volumetric, creating a nostalgic and peaceful atmosphere.

Example generated from “Surreal Cloud Coffee Scene”
GPT Image 2en

Create an ultra-realistic, cinematic, photorealistic surreal photograph of a handsome young South Asian man sitting casually on a gigantic fluffy white cloud hundreds of meters above a sprawling modern city, peacefully drinking a cup of coffee while commercial airplanes fly beneath and around him. The man is in his mid-20s, with thick slightly messy wavy black hair, warm brown skin, expressive deep-brown eyes, and a neatly trimmed beard and mustache. He wears an oversized black technical jacket over a clean white T-shirt, dark wide-leg cargo trousers, futuristic black sneakers, a silver chain, and a black smartwatch. His expression is calm and relaxed, as if sitting on a balcony above the world. The enormous cloud looks physically realistic—soft, dense, voluminous and detailed, with individual vapor textures and subtle sunlight passing through its edges. The man sits naturally on its surface, one leg slightly hanging over the edge, holding a realistic ceramic coffee cup in one hand. Far below, a vast modern city stretches toward the horizon with skyscrapers, roads, tiny vehicles, parks and realistic atmospheric haze. Several passenger airplanes fly through the sky at different distances and scales, creating an incredible sense of depth. One airplane passes relatively close beneath the cloud while others appear tiny near the horizon. Golden-hour sunlight illuminates the scene from the side, creating warm highlights on the man's face, clothing and cloud while casting soft natural shadows. The sky transitions from bright blue near the horizon to deeper blue overhead, with subtle wispy clouds scattered across the atmosphere. Use an epic wide-angle composition showing the enormous scale difference between the man, cloud, airplanes and city below. Strong foreground-to-background depth, realistic atmospheric perspective, natural reflections, physically accurate lighting, detailed skin texture, realistic fabric fibers, realistic hair strands, cinematic color grading, HDR, subtle film grain, shallow depth of field on the character with the distant city naturally softened. Photographed like a high-end cinematic editorial campaign, 35mm lens, f/2.8, ultra-detailed, 8K UHD, photorealistic, natural proportions, realistic human anatomy, believable physics, dramatic scale, surreal but completely convincing. No cartoon appearance, no illustration, no CGI look, no plastic skin, no distorted face, no extra fingers, no duplicate people, no deformed airplanes, no floating objects, no text, no watermark.

Example generated from “Cinematic Multi-Door Fantasy Scene Prompt”
GPT Image 2en

Ultra-realistic, cinematic 8K HDR, high-fashion editorial photograph of a strikingly handsome young man in his early 20s, with thick naturally wavy jet-black hair, strong masculine eyebrows, intense deep dark-brown eyes, a sharply defined sculpted jawline, subtle natural stubble, and realistic skin texture with visible pores. CONCEPT & COMPOSITION: A breathtaking, surreal cinematic scene featuring a handsome young man standing confidently at the center of a gigantic, mysterious hall surrounded by dozens of magnificent doors of different shapes, sizes, materials, and architectural styles. Every door opens into a completely different world, creating a spectacular visual contrast between reality, fantasy, and the unknown. The young man wears a perfectly tailored black outfit consisting of a luxurious black high-neck shirt, a long elegant black coat, tailored black trousers, and polished black boots. His thick, naturally wavy hair is slightly tousled, and his intense eyes gaze directly toward the camera with a calm, confident, mysterious expression. One hand rests gently on the handle of a partially opened door, as if he is about to step into another dimension. THE DOORS & THEIR WORLDS: Surrounding him are dozens of enormous, beautifully detailed doorways, each revealing a unique and breathtaking environment: 🌌 Cosmic Dimension: A glowing galaxy filled with colorful nebulae, distant planets, stars, and swirling cosmic energy. 🏔️ Frozen Kingdom: A magical winter landscape with snow-covered mountains, frozen waterfalls, and gently falling snow. 🔥 Kingdom of Fire: A dramatic volcanic world with flowing rivers of lava, burning skies, and glowing embers. 🌸 Cherry Blossom Paradise: A peaceful Japanese landscape filled with pink cherry blossoms, a traditional Japanese temple, and a beautiful red bridge. 🏰 Dark Fantasy Realm: A mysterious gothic castle surrounded by a dark enchanted forest and a sky illuminated by lightning. 🌆 Cyberpunk Future: A futuristic city with towering skyscrapers, glowing neon signs, flying vehicles, and vibrant blue and purple lights. 🌿 Enchanted Forest: An extraordinary magical forest with gigantic ancient trees, glowing flowers, floating golden particles, and mystical waterfalls. ☁️ Heavenly Dimension: A magnificent world above the clouds featuring golden palaces, radiant sunlight, and endless white clouds. 🏜️ Lost Desert Civilization: An ancient golden city surrounded by desert dunes.

Example generated from “Steampunk Last Gear Prompt”
GPT Image 2ja

"The Last Gear" 機械都市は停止した。 私は最後に残った歯車をそっと拾い上げた。 もしこれが噛み合えば、 きっと、この世界は再び動き出すだろう。 ⚙️💧✨

Example generated from “Giant Police Robot at Night Rescue”
GPT Image 2en

A cinematic, photorealistic night scene depicting a massive bipedal police robot standing in the center of a muddy disaster response zone during heavy rain. The giant robot is white and black with industrial paneling, featuring red emergency lights on its shoulders and head, and markings that read "警視庁 MPD" (Metropolitan Police Department) on its chest and arms. Its head consists of multiple camera lenses arranged horizontally. Surrounding the robot's feet are numerous first responders: police officers in dark uniforms with "MPD" on their backs, firefighters in orange gear, and paramedics in blue vests with Japanese text. To the left is a mobile command unit truck displaying screens with data maps. In the background, there is a large pile of rubble or debris illuminated by floodlights, with excavators visible atop the mound. Several drones hover in the sky, casting bright spotlights down onto the wet ground. The atmosphere is tense and high-tech, with reflections on the wet pavement and rain streaks visible against the dark sky.

Example generated from “Summoning Magic Comedy Prompt — Reference Image Required”
GPT Image 2ja

添付画像のキャラクターが、壮大な召喚儀式によって思いもよらないものを呼び出してしまうコミカルなシーンを描いてください。 [キャラクター] 顔立ち、髪型、基本となる配色、性格的特徴はそのまま維持してください。表情やポーズは、このシーンに合わせて変更します。 [召喚されたもの] 具体例や候補は提示せず、今回のためにまったく新しいものを考えてください。壮大な儀式との強烈なギャップを生み出し、絵を見ただけで何が出現したのか一目でわかり、「なぜそれを召喚した?」とツッコミたくなるようなものにします。キャラクターの外見や衣装に直接関連するものだけに限定せず、意外性のある組み合わせを取り入れてください。 [シーンの構成] 召喚されるもの、そのサイズや状態、キャラクターのリアクション、周囲への影響を、それぞれ別の視点から決めていきます。その後、それらを 1 つのわかりやすい出来事としてつなぎ合わせます。キャラクターのリアクションを単なる「驚き」に固定せず、表情や動きを通じて召喚されたものとの関係性が伝わるようにしてください。 [演出] 儀式の演出は本格的かつ壮大なものにし、出現したものとの温度差を利用してユーモアを生み出します。メインのオチは 1 つに絞り、それを引き立てるための小道具やサブキャラクターを配置してください。テキストによる説明には頼らず、キャラクター・召喚されたもの・リアクションが一目で読み取れる構図にします。なお、この会話内で同じプロジェクトの画像がすでに生成されている場合は、そこで使用したものやオチを再利用せず、別の方向性で制作してください。

Example generated from “Smug Character Sitting and Pointing — Reference Image Required”
GPT Image 2en

Using the provided reference image, regenerate the character in a new pose and context. Keep the character's appearance (pink hair, glasses, cap, hi-vis vest) unchanged. **Transformation Instructions:** 1. **Pose & Expression:** Change the character from standing to sitting casually in an office chair with legs crossed. Update her expression from angry/shouting to a smug, confident smirk. She is now pointing directly at the viewer with one hand while holding a can up in the other. 2. **Dialogue:** Replace the original speech bubble text with "SKILL ISSUE." in bold, black capital letters. 3. **Props & Details:** - Swap the "Steel Reserve" can for a generic white "Lite" beer can. - Add red wired earphones hanging from her ears. - Place a small grey cat plushie on her chest/vest area. 4. **Background Changes:** - Modify the poster on the left wall to feature a simple line-drawing of a cat face with X eyes above the text "SAME SHIT DIFFERENT DAY". - Add a new poster on the right wall that reads "CANS CATS PLANES ETC.". - Include a mug on the desk that says "BAD PLANS TODAY".

Example generated from “Wanted Poster Character Prompt — Reference Image Required”
GPT Image 2ja

添付画像のキャラクター [名前: {argument name="character_name" default="Unknown"}] を、「たぶん無実なのに、なぜか不思議な罪で連行されている」ようなコミカルな指名手配風・逮捕風のビジュアルにしてください。 容疑はランダムに決めてください。内容は軽妙で楽しく、ユーモアたっぷりに仕上げてください。 構図は、身長計の前に立っているポーズか、軽く取り押さえられている瞬間にしてください。さらに、周囲には「証拠品」として奇妙な小物を配置してください。2 カット(正面と横顔)でも、1 枚の画像でも構いません。実在する警察や国のマークは使用せず、完全にフィクションのユーモアにとどめてください。後から容疑を書き込めるよう余白を残しておいてください。

Example generated from “Anime Girl Pitching Baseball”
GPT Image 2en

A vertical webtoon-style illustration of a female anime character pitching a baseball. The character has long silver hair, cat ears, and purple eyes. She wears an orange baseball cap with a logo, a white jersey with 'SynClub' written on it, white pants with orange stripes, and a black glove. Her right hand is extended forward, releasing a glowing baseball surrounded by swirling purple energy and dark fox-like spirits. In the background, there's a stadium filled with spectators under a blue sky with clouds. On the left side, there are three smaller panels: one showing her face in shock, another showing her full body from a distance, and a third showing a close-up of the purple energy. At the top right, there's a speech bubble with Japanese text. At the bottom, large stylized Japanese text reads '狐火トルネード・タイプ3'. The overall style is dynamic and energetic, typical of sports manga.

Example generated from “Debonair Man with Bowling Gear Prompt”
GPT Image 2en

a man appears wearing a brown trench coat looking debonair despite his green bowling ball and multi-colored bowling shoes

Example generated from “Cute Penguin Girl Comic Strip”
GPT Image 2en

A four-panel vertical comic strip featuring a cute chibi girl character wearing a fluffy black penguin costume with a large pink satin bow tie. The art style is hyper-realistic 3D render (Pixar/Blender style) with soft lighting and shallow depth of field. Panel 1: Medium shot of the girl walking on a sunny park path, smiling happily. Speech bubble text: "いい天気だな〜♪" (Nice weather~). Panel 2: Rear view of the girl looking at a silver metal trash can in the grass. Speech bubble text: "……あれ?なんだろう" (...Huh? What's that?). Panel 3: Extreme close-up of the girl's face with wide blue eyes and an open mouth, looking surprised. Speech bubble text: "おっきい!" (It's big!). Panel 4: Side view of the girl sniffing the trash can curiously. The can has a label reading "一斗缶 18L". Speech bubble text: "くんくん…" (Sniff sniff...).

Example generated from “Mad Crow Popping Yellow Balloon”
GPT Image 2en

A whimsical, hand-drawn illustration in a storybook style depicts a chaotic scene on a telephone wire. On the left, a large, frantic black crow with wide, bulging eyes and an open beak is perched on the wire, its wings spread in agitation. To the right of the crow, a bright yellow balloon is caught on the wire and has just burst, with jagged yellow shards flying outward in all directions against a backdrop of fluffy white clouds and blue sky. Below the wire, the tops of several European-style rooftops are visible, featuring red tiles and chimneys. The artwork uses textured lines and watercolor-like shading.

Example generated from “Neo-Naive Style Character Fight Illustration Prompt”
GPT Image 2zh

[角色 A] 和 [角色 B],新天真绘画 × 局外人艺术 × 原始具象绘画 × 儿童绘画风格的角色转译。 两个角色由少量不规则的几何色块构成,仅保留其最强烈的身份标识:主色调、头部轮廓、发型或耳朵、眼睛、代表性服装色块、标志性配饰。 两人正在激烈但滑稽地打斗,身体紧贴并纠缠在一起。根据角色随机设计动作,如抓脸、推头、拥抱、拉扯、扇耳光、压制或互相扑倒;手部接触清晰,肢体交叉,重心失衡,使打斗关系一目了然。 幼稚且不准确的人体解剖结构:细长的躯干,僵硬笔直的四肢,比例失调,头部倾斜,夸张的手脚,错误的透视。面部特征压缩成几个粗糙的色块,表情愤怒、震惊、龇牙咧嘴、尖叫或委屈,具有荒诞的喜剧感。 粗糙的不透明颜料平涂,干笔痕迹,边缘磨损,填充不均,颤抖的手绘轮廓,局部底色暴露;身体和服装直接以平涂色块表现。 暖灰白色或米白色纸张背景,大面积留白。两个角色位于中心,相互纠缠形成紧凑的动作团,画面专注于表达角色之间的关系。 就像一个未经训练的人凭记忆快速画出的打斗瞬间:原始、笨拙、直接、荒诞、可爱。 竖版格式,右下角有极小的 "voxcat" 签名。

Example generated from “Opus 5.5 vs Wife GPT Comic Strip”
GPT Image 2en

Goal: Create a vertical 3-panel comic strip in an anime style, titled "Thoughts After Experiencing Opus 5.5". Canvas: Vertical layout with three distinct panels stacked on top of each other. Panel 1 (Top): - Subject: A young man with messy black hair and starry eyes looks excitedly at a cheerful girl with long orange hair wearing a sunflower-themed dress. - Text Bubble (Man): "After experiencing Opus 5.5... Wow... so charming....!!" - Info Box (Right): A yellow decorative box labeled "Claude:" containing the text "Youthful and smart, making life colorful!". - Background: Bright, sunny field of sunflowers. Panel 2 (Middle): - Subject: The same man is now hiding his face in his hands, sweating nervously. Beside him stands an elegant woman with long white hair, horns, and a cold expression. - Text Bubble (Man): "Oh no, this feels like cheating... Lost the passion of newlyweds..." - Info Box (Left): A purple decorative box labeled "Wife GPT:" containing the text "Still beautiful and elegant, but a bit cold.". - Background: Darker, indoor setting with blue curtains. Panel 3 (Bottom): - Subject: A scene at a train station platform. The man is peeking around a pillar, blushing, next to the orange-haired girl (Claude). In the background, the white-haired woman (Wife GPT) stands watching them from a distance near a train. - Signage: A blue sign hanging above reads "Transfer Station ->". - Text Bubble (Left): "Strict rules at home, can only meet secretly at the transfer station! Keep it down..." - Text Bubble (Right): "I know everything... watching quietly." Visual Style: Clean digital illustration, vibrant colors for the first panel, muted tones for the second, realistic lighting for the third.

Example generated from “Manga Infographic: AI Prompt Engineering”
GPT Image 2en

A vertical, highly detailed manga-style infographic page with a parchment-like background. The layout is divided into three main sections: a top illustration area, a middle diagrammatic breakdown, and a bottom text-heavy conclusion. Top Section: - A large, ethereal illustration of a woman with long white hair and fox ears (a kitsune) standing on a mountain peak at sunset. She wears flowing white robes. - Below her, an elderly man with long black hair tied back, wearing dark blue traditional Chinese robes, sits cross-legged holding a calligraphy brush, looking up thoughtfully. - To his right, a younger man in grey robes kneels, sketching the kitsune onto paper. - Surrounding this scene are blocks of Japanese text explaining the concept of 'Prompt' (プロンプト). - On the left side, there are four small square character portraits labeled with descriptions like 'Beautiful Woman', 'Dancer', etc., showing variations of female characters. - Large vertical Japanese text on the right reads 'Introduction to Prompts'. Middle Section: - A horizontal flow diagram titled 'Verbalizing features to compose a picture'. - It shows a sequence of six small square images connected by plus signs (+), leading to an equals sign (=) and a final result image. - The sequence includes: 1. Long silver hair, 2. Blue eyes, 3. White clothes, 4. Mountain sunset, 5. Looking back pose, 6. Soft backlighting. - The final result is the complete image of the kitsune from the top section. Bottom Section: - A large, bold Japanese headline in red and black ink style: "Listing hundreds of words loses one stroke" (百語を並べて一画を失う). - Surrounding the headline are torn pieces of paper with handwritten notes listing negative effects of over-prompting, such as '8K resolution', 'ultra-high quality', 'complex composition', 'unnatural lighting', 'film-like lighting', 'large scale', 'highest quality'. - At the very bottom, a concluding sentence in a box states that words are tools, but too many obscure the essence of the single stroke.

Example generated from “Chun-Li Kicks Giant Monster”
GPT Image 2en

A high-resolution digital illustration in a vibrant, comic-book style depicting a chaotic street fight. The scene features Chun-Li from Street Fighter on the right side, wearing her iconic blue qipao with gold trim and white boots, executing a powerful high kick that connects directly with the torso of a massive green-skinned monster. The monster resembles Blanka but is significantly larger and more muscular, with wild red hair, sharp claws, and glowing eyes, roaring in pain as it absorbs the impact. In the foreground, lying defeated on the rubble-strewn ground, is Ryu from Street Fighter, wearing his torn white gi, appearing battered and unconscious. The background shows a gritty urban alleyway with brick buildings, fire escapes, and overflowing trash cans under a dramatic, fiery orange and purple sky. Debris and dust explode outward from the point of impact.

Example generated from “Hand-Drawn Fashion Illustration Transformation — Reference Image Required”
GPT Image 2en

Transform the uploaded photo into a charming hand-drawn fashion illustration while preserving the original scene and composition. Keep the same people, number of people, pose, body positioning, facial expressions, clothing, accessories, hairstyles, objects, drinks, and overall framing from the reference photo. Do not add or remove important elements. Convert the realistic photograph into a minimal Japanese/Korean-inspired lifestyle sketch illustration, using loose but confident dark green ink linework with subtle hand-drawn imperfections. Simplify facial features while keeping each person’s recognizable hairstyle, glasses, headphones, clothing silhouette, and pose. Use a soft pastel pink background with the illustration drawn primarily in deep forest-green ink. Replace realistic textures, shadows, vehicles, buildings, pavement, and surroundings with simplified expressive linework and selective cross-hatching. Keep enough environmental details to make the original location and composition recognizable. Style: cute editorial illustration, Japanese stationery aesthetic, fashion sketch, casual slice-of-life manga drawing, imperfect pen strokes, clean negative space, minimal detailing, subtle vintage feel, handmade appearance. Color treatment: predominantly pastel pink background + dark green line art, with very minimal secondary tones if needed. No photorealism, no 3D rendering, no glossy digital painting. Important: preserve the original image’s composition, perspective, camera angle, poses, proportions, clothing, accessories, and key objects. The result should look like the exact photograph was redrawn by an illustrator, rather than redesigned into a different scene. Output: clean vertical illustration, aesthetically balanced, high-resolution, polished hand-drawn finish. Short version for image models: Turn this photo into a cute hand-drawn Japanese lifestyle/fashion sketch. Preserve the exact people, poses, expressions, clothing, hairstyles, accessories, drinks, composition, camera angle and important background elements. Redraw everything with loose imperfect dark forest-green ink linework on a soft pastel pink background, using minimal cross-hatching and simplified details. Charming editorial stationery aesthetic, casual manga-inspired illustration, handmade pen texture, clean negative space, elegant and whimsical. Do not change the composition or add/remove subjects; make it look like the original photo was directly illustrated by hand.