GPT Image 2 prompts

Browse GPT Image 2 prompts for fashion portraits and graphic compositions. Compare how each prompt describes its subject and the way the image is arranged.

探索提示词

9,926 条提示词

Profile / Avatar - Jirai-Kei Girl iPhone Selfie — output
Comic / Storyboard - Dramatic Green Lantern Comic Portrait — output
Product Marketing - Minimalist Fashion Editorial Portrait — output

Product Marketing - Minimalist Fashion Editorial Portrait

yusra.GPT Image 2
A full-length editorial fashion portrait of a young {argument name="ethnicity" default="East Asian"} woman seated gracefully on a {argument name="staircase material" default="dark walnut wooden staircase"} inside a quiet, minimalist European-style interior. She sits several steps above the bottom of the staircase, facing directly toward the camera with an elegant, composed expression and relaxed posture. Her hands are gently placed together on her lap, knees positioned naturally beneath the voluminous skirt. She has long, straight-to-softly-wavy jet-black hair, parted slightly to one side, cascading over one shoulder and down the front of her dress. A large oversized black satin bow hair accessory is positioned on the side of her head, creating a striking black-and-white contrast. Her makeup is extremely natural and refined: porcelain-like luminous skin, subtle blush, softly defined dark eyes, naturally shaped brows, and muted rosy lips. Her expression is calm, innocent, delicate, and almost doll-like without looking exaggerated. She is wearing an {argument name="dress style" default="ethereal ivory-white formal ballgown"} made from lightweight, slightly textured fabric. The dress has a fitted bodice with a defined waist, a simple elegant rounded square neckline, and dramatic short puff sleeves with substantial gathered volume around the shoulders and upper arms. The waist is cleanly cinched, transitioning into an enormous, sculptural full skirt. The skirt is the focal point of the outfit: extremely voluminous, layered, asymmetrical and high-low in construction, with soft cascading folds and overlapping layers of ivory fabric. The front of the skirt rises enough to reveal her lower legs and shoes while the sides and back extend dramatically outward and downward across the staircase. The fabric pools naturally over several wooden steps, creating an elegant flowing silhouette. The skirt should feel luxurious but not overly ornate, with no embroidery, no sequins, no lace, no visible patterns, relying entirely on shape, volume, texture, and draping. Beneath the raised front of the skirt, she wears black pointed-toe pumps with delicate black bow embellishments. The shoes are partially concealed by the dress but clearly visible between the folds. The staircase is made from rich dark brown polished walnut wood, featuring broad wooden steps, traditional turned wooden balusters, a substantial handrail, and a large decorative newel post in the foreground. The staircase rises behind her and creates strong architectural symmetry. The surrounding walls are warm off-white/cream, completely minimal and undecorated, with a pale natural wooden floor visible at the bottom. The composition is centered and symmetrical, with the woman occupying the visual center while the staircase creates strong leading lines around her. The camera is positioned at approximately eye level to slightly below eye level, capturing her entire body and the dramatic spread of the gown. The framing is vertical, fashion-editorial, approximately 4:5 portrait ratio. Lighting is soft n
使用此提示词 ↗
Product Marketing - High Fashion Magazine Editorial — output

Product Marketing - High Fashion Magazine Editorial

Aniya IntelligenceGPT Image 2
Create a premium, hyper-realistic high-fashion editorial magazine cover using my {argument name="reference face" default="uploaded female face"} as the ONLY identity reference for the main female subject.

Preserve my exact facial identity and natural appearance — facial structure, skin tone, eyes, eyebrows, nose, lips, cheekbones, jawline, facial proportions, natural skin texture, and overall likeness. Keep me instantly recognizable. Do not beautify, reshape, feminize, age, de-age, or alter my facial features.

Recreate the overall composition, fashion aesthetic, lighting, typography placement, and editorial mood of the reference image, while creating an original design.

Pose & styling: A sophisticated female fashion model shown in an elegant three-quarter/profile pose, looking slightly toward the left. She wears a dramatic {argument name="accessory" default="oversized wide-brim sculptural hat"} and a luxurious flowing fashion outfit with a high draped collar/cowl around the neck. The outfit should feel avant-garde, elegant, and couture.

Color palette: Deep burgundy, crimson red, wine red, and black, with subtle tonal gradients. Use dramatic cinematic lighting with soft highlights across the face and strong shadows for a mysterious luxury-fashion atmosphere.

Background: Minimalist deep burgundy editorial backdrop with a smooth gradient transitioning toward near-black at the bottom. Clean, sophisticated, premium fashion-magazine aesthetic.

Typography: Create a large elegant serif magazine title at the upper left reading “FASHION”, with sophisticated editorial typography. Add:
“NEW FASHION”
“NEW COLLECTION”

Include a small editorial text box reading:
“MORE THAN 20 BEST BODY OUTFIT”

Near the lower center, add the headline:
“SECRETS OF”
“BEAUTY”

Add a small handwritten-style tagline beneath:
“IT’S ALL ABOUT FASHION”

Typography should look professionally typeset, balanced, refined, and integrated naturally into the composition. Avoid random or misspelled text.

Photography style: ultra-realistic professional fashion photography, luxury magazine editorial, cinematic studio lighting, realistic skin texture, detailed fabric, subtle shadows, high dynamic range, sophisticated composition, premium print-quality finish.

Composition: vertical magazine-cover layout, subject positioned toward the right side with generous negative space on the left for typography. Keep the face clearly visible beneath the hat brim. Strong visual hierarchy and polished Vogue/Harper’s Bazaar-inspired editorial sophistication.
使用此提示词 ↗
Profile / Avatar - Editorial Portrait with Retro Headphones — output

Profile / Avatar - Editorial Portrait with Retro Headphones

MaercihGPT Image 2
{
  "input": {
    "subject_reference": "{argument name="subject reference" default="{{UPLOAD_SUBJECT_IMAGE}}"}",
    "pose": "{argument name="pose" default="{{LEAVE_DEFAULT}}"}"
  },
  "prompt": {
    "type": "photorealistic_editorial_portrait",
    "subject": {
      "source": "uploaded_reference_image",
      "instruction": "Use the uploaded reference image as the single source of truth for the subject. Preserve the person's identity, facial structure, proportions, natural features, skin tone, hairstyle, and overall appearance. Do not replace or redesign the subject."
    },
    "composition": {
      "framing": "tight close-up portrait",
      "camera_angle": "straight-on, eye-level",
      "pose": "{argument name="pose" default="{{POSE}}"}",
      "subject_position": "centered in frame",
      "camera_distance": "close to the subject"
    },
    "styling": {
      "expression": "calm, confident, subtly intense gaze toward camera",
      "skin": "realistic glowing skin with natural texture, subtle highlights and visible pores",
      "makeup": "minimal natural editorial makeup"
    },
    "wardrobe": {
      "description": "simple dark gray tank top with thin straps"
    },
    "accessories": {
      "headphones": "large retro over-ear headphones, glossy lime-green earcups with black padded headband and black ear cushions"
    },
    "pose_details": {
      "hands": "both hands resting naturally one over the other beneath the chin",
      "foreground_object": "large bright red circular object occupying the lower foreground"
    },
    "environment": {
      "background": "minimal warm beige studio backdrop",
      "setting": "clean professional studio portrait"
    },
    "lighting": {
      "style": "soft controlled studio lighting",
      "direction": "front-facing diffused light",
      "quality": "soft shadows with gentle facial highlights",
      "look": "high-end beauty editorial photography"
    },
    "camera": {
      "style": "professional fashion photography",
      "lens": "85mm portrait lens",
      "depth_of_field": "shallow but natural",
      "focus": "tack-sharp eyes and facial features"
    },
    "visual_quality": {
      "realism": "extremely photorealistic",
      "detail": "high micro-detail, realistic hair strands, natural skin texture",
      "color_grading": "warm neutral editorial tones",
      "finish": "premium magazine beauty campaign"
    },
    "negative_prompt": [
      "identity change",
      "different person",
      "altered facial structure",
      "plastic skin",
      "over-smoothed skin",
      "AI-looking face",
      "excessive makeup",
      "distorted hands",
      "extra fingers",
      "deformed anatomy",
      "warped headphones",
      "text",
      "logos",
      "watermarks",
      "artificial beauty filter",
      "cartoon",
      "illustration",
      "low resolution"
    ],
    "aspect_ratio": "3:4"
  }
}
使用此提示词 ↗
Social Media Post - Cinematic Stadium Sports Portrait — output
Product Marketing - One-Prompt Object Recolor Transformation — output
Social Media Post - Candid Late-Night Kitchen Snapshots — output

Social Media Post - Candid Late-Night Kitchen Snapshots

DeepBlue深藍GPT Image 2
Vertical realistic mobile phone snapshot, late night residential kitchen, {argument name="subject" default="adult female"} opening the refrigerator looking for food; the whole thing is like a life moment suddenly captured by a roommate standing next to them with a mobile phone.
Random number of people, random positions, random refrigerator content, random actions, random shots; the person may be squatting and rummaging through the refrigerator, bending over to check, staring at a drink, looking back at the camera, blocking the refrigerator light with a hand, or holding a bunch of snacks.
Random shots, internal perspective of the refrigerator, ultra-close-up wide angle, low angle, side-rear snapshot, door gap occlusion, half-face crop, refrigerator light overexposure, mirror reflection, slight motion blur; random selection.
The refrigerator randomly displays {argument name="items" default="drinks, leftovers, eggs, plastic boxes, sauces, half a watermelon, take-out boxes, fresh-keeping bags, strange snacks"}; each photo only highlights a few real items.
The kitchen background randomly displays a sink, garbage can, magnetic stickers, dining table, shopping bags that haven't been put away, washed cups, and hanging towels.
The cold white light of the refrigerator becomes the main light source, the surrounding environment is dark, real mobile phone automatic exposure, high local exposure on the character's face, slight noise and compression, real life texture.
Focus: Refrigerator content, character actions, lens position, random expressions; allow for a slightly absurd but reasonable object, keeping the overall look real.
Signature at the bottom left "● DeepBlue".
使用此提示词 ↗
E-commerce Main Image - Stirred Iced Honey Latte Close-Up — output
Social Media Post - Miniature Fever Patient in Water Glass — output

Social Media Post - Miniature Fever Patient in Water Glass

Mehwish kiranGPT Image 2
Create a highly realistic cinematic photorealistic night bedside scene with a surreal miniature person inside a clear glass of water. The subject is a sick young {argument name="character gender" default="woman"} sitting curled up at the bottom of the glass, tiny in scale, wearing dark navy satin pajamas with white piping, bare feet visible, looking tired and feverish with one cheek resting on one hand. Preserve natural human proportions and detailed facial expression. Add a white fever-cooling patch on the forehead and slightly messy long wavy {argument name="hair color" default="brown hair"}. The glass is centered on a warm wooden bedside table, filled with clear sparkling water containing many tiny air bubbles around the subject, with realistic refraction, distortion, highlights, and thick glass rim/base. Surround the glass with exactly 5 bedside illness items: 1 blue-and-white medicine box on the left, 1 blister pack of round pills in the lower left, 1 loose white capsule near the front left, 1 digital thermometer in the lower right displaying {argument name="thermometer reading" default="38.7°C"}, and 1 crumpled white tissue on the right. Add 1 dark blue ceramic mug with gold crescent moon and star pattern behind the glass on the right. Background includes 1 warm table lamp with beige fabric shade on the left, softly glowing, plus a blurred alarm clock and plant in the dark room. Use shallow depth of field, warm amber lamp light mixed with cool night shadows, cinematic vertical composition, macro realism, ultra-detailed textures, cozy sick-day atmosphere, no extra people, no text except the thermometer digits, no watermark.
使用此提示词 ↗
Comic / Storyboard - Photo to Cozy Crayon Drawing Comparison — output

Comic / Storyboard - Photo to Cozy Crayon Drawing Comparison

underwoodGPT Image 2
Goal: Create a vertical 3:4 comparison image showing a warm original-photo scene on top and its cozy crayon-drawing interpretation on the bottom.

Canvas: Vertical image, approximately 900 x 1200 px, split into exactly 2 horizontal panels of equal width. The top panel is a soft, sunlit realistic photo; the bottom panel is an off-white paper illustration with wax-crayon texture.

Top panel: Show {argument name="character description" default="a young woman with long dark hair"} sitting sideways on a narrow old balcony railing outside a weathered apartment doorway. She is relaxed and slightly introspective, looking down at an acoustic guitar while playing it. She wears {argument name="outfit" default="a loose patterned short-sleeve shirt, dark shorts, white socks, and chunky white sneakers"}. The guitar is a glossy honey-brown acoustic cutaway guitar with a dark tortoiseshell pickguard. Include iron balcony bars, leafy green vines, a cream stucco wall, an open aged door, and a dim interior room in the background. Use golden late-afternoon sunlight with dappled leaf shadows across her legs, guitar, wall, and floor; dreamy lens softness, cozy summer atmosphere, shallow depth of field.

Bottom panel: Recreate the same subject as a loose hand-drawn crayon sketch on textured cream paper. The woman, pose, guitar, balcony railing, and shoes should be simplified but recognizable, drawn with sketchy overlapping crayon strokes in warm yellow, teal blue, coral red, green, and charcoal gray. Keep the style childlike yet stylish, soft and cozy, with visible wax grain and imperfect outlines. Add exactly 7 decorative crayon elements around the main figure: 1 simple yellow sun in the upper left, 2 small yellow flowers on green stems on the left, 1 blue musical note near the center-right, 1 short gray handwritten phrase reading {argument name="quote text" default="Better Days Ahead"}, and 2 small yellow flowers on green stems in the lower right. Add a few loose green and blue ground scribbles below the seated figure.

Text and branding: The only readable phrase inside the drawing is {argument name="quote text" default="Better Days Ahead"}, set in a small gray typewriter-like handwritten style to the right of the illustrated guitar player. Add a subtle white watermark reading {argument name="watermark text" default="Promptsref.com"} in the bottom-right corner.

Constraints: Use exactly 2 panels, with the top photorealistic and the bottom crayon-drawn. Do not add extra people, extra guitars, or additional readable text. Preserve the warm sunny mood and the cozy hand-crafted crayon aesthetic.
使用此提示词 ↗
Game Asset - Isometric Voxel River Temple City — output

Game Asset - Isometric Voxel River Temple City

ashenGPT Image 2
Create a detailed isometric voxel diorama of an ancient stepped temple complex beside a river, rendered like a miniature block-built game scene on a rectangular base. The setting should feel like {argument name="civilization style" default="an ancient Mesopotamian or Mesoamerican river temple city"}, with warm tan stone, terracotta roofs, carved columns, and symmetrical ceremonial architecture.

Canvas: Wide 4:3 composition, orthographic isometric view from above, centered on a floating rectangular terrain slab against a plain warm beige background. Use soft studio lighting, ambient occlusion, gentle shadows, and a polished low-poly voxel aesthetic.

Main layout: Place one large central stepped ziggurat temple slightly left of center, with many stacked terraces, vertical column-like details, a flat pale roof, and small golden ornaments along the top. In front of it, add a lower entrance platform with columns and stairs. To the right of the main temple, include one smaller stepped temple shrine; behind it, include one even smaller tower shrine near trees. On the left foreground, create a rectangular ceremonial pool with bright blue water, stone edging, and tiered steps around it. Along the entire front edge, create a broad turquoise river with tiled water texture and a stone promenade with stairs descending to the water.

Counted visible elements: exactly 3 stepped temple structures: 1 large central ziggurat, 1 medium front-right ziggurat, and 1 small rear-right shrine tower. Exactly 1 rectangular pool. Exactly 2 small wooden boats on the river, each with a tiny mast or pole. Exactly 5 terracotta-roofed rectangular buildings or market stalls around the plaza: 2 long buildings on the left side, 2 buildings on the right side, and 1 small building in the front-left area. Exactly 4 striped canopy market stalls in the plaza: 1 blue-and-white canopy, 1 teal-and-white canopy, 1 red-and-white canopy, and 1 yellow-and-white canopy. Exactly 5 palm trees distributed around the complex, plus exactly 4 broad leafy green trees. Add scattered tiny voxel people, small benches, planters, reeds, crates, and decorative posts, but keep them secondary.

Visual style: Charming voxel art, crisp block geometry, tiny modular bricks, warm earthy palette of sand, clay, ochre, cream stone, muted green vegetation, and bright cyan water. The temple should be majestic but toy-like, with no photorealistic textures, no outlines, and no text. Emphasize the {argument name="mood" default="lowkey beautiful peaceful miniature world"} feeling, with clean composition and high detail.
使用此提示词 ↗
Social Media Post - Real Woman with Illustrated Twin — output

Social Media Post - Real Woman with Illustrated Twin

Laraib Fatima‎GPT Image 2
Create a highly realistic editorial-style fashion photograph of one real young woman standing beside one hand-drawn illustrated girl who looks like her miniature visual twin. The scene is vertical, full-body, shot against a plain textured gray concrete wall with a rough gray pavement floor, soft natural daylight, muted colors, and a clean minimal street-style composition. Place the illustrated girl on the left, about half the height of the real woman, and the real woman on the right, both facing slightly toward each other with gentle smiles. The real woman is a young East Asian woman with {argument name="hair style" default="long slightly wavy dark brown hair with wispy bangs"}, round thin eyeglasses, and a calm shy expression, looking down toward the illustrated girl. She wears {argument name="main outfit" default="a cobalt blue chore jacket over a cream sweater with navy horizontal stripes, a long cream pleated skirt, white socks, and burgundy leather loafers"}. Add a small round tan crossbody bag with a narrow strap worn diagonally across her torso; her hands are tucked casually into the jacket pockets. The illustrated girl should be a charming hand-drawn childlike counterpart with soft pencil-and-watercolor texture, slightly oversized head, rosy cheeks, round glasses, long messy dark hair, and the same outfit details: blue jacket, striped cream shirt, long cream skirt, white socks, burgundy loafers, and tan crossbody bag. Make the illustration appear physically present in the same space, aligned with the ground plane, with subtle contact shadow and believable scale, while preserving visible sketch outlines and painterly shading. Use a realistic 35mm editorial photo look, natural proportions for the real woman, crisp fabric detail, no text, no logos, no extra people, no props beyond the single tan bag on each figure.
使用此提示词 ↗
Product Marketing - Luxury KFC Fashion Ad — output

Product Marketing - Luxury KFC Fashion Ad

CaiKa AIGPT Image 2
Create a luxury fashion fast-food advertisement for {argument name="brand name" default="KFC"} with a vertical 4:5 editorial composition. A stylish young East Asian woman in a refined red-and-white uniform-inspired outfit reaches dynamically toward the camera to grab one piece of crispy fried chicken from an oversized red-and-white striped chicken bucket in the foreground. The bucket is packed with exactly 8 visible golden fried chicken pieces, with crunchy crumbs and tiny flakes flying in the air around her hand. Show the brand logo on the front of the bucket: Colonel-style face icon above the text “KFC”. The model has loose dark hair in an elegant messy updo, soft glowing skin, subtle warm makeup, red lips, and small gold hoop earrings; her expression is confident, tempting, and slightly glamorous as she looks toward the viewer. Outfit details: cream cropped jacket with rolled sleeves, deep red sleeveless top, red waist sash/belt, cream high-waisted wrap-style skirt or apron panel over red fabric, clean tailored fashion styling. Use a low-angle wide-lens perspective from near the bucket, making the fried chicken large and tactile in the foreground and the model’s reaching arm create motion and depth. Background: minimalist geometric red-and-cream studio architecture with bold red wall planes, warm sunlight, soft shadows, premium commercial photography mood. Add elegant ad copy in the upper left: headline “{argument name="headline text" default="CRAVE THE CRUNCH"}” in spaced uppercase serif letters, and smaller subheadline “{argument name="subheadline text" default="Freshly crispy. Made to tempt."}”. Color palette: rich KFC red, warm cream, golden-brown chicken, high-end glossy food styling. Lighting should be cinematic, warm, clean, and polished, with shallow depth of field, crisp details on the chicken crust, and a premium fashion magazine advertising finish. Avoid extra text, extra logos, distorted hands, or messy packaging.
使用此提示词 ↗
Social Media Post - Stockholm Travel Photo Collage — output

Social Media Post - Stockholm Travel Photo Collage

小小东GPT Image 2
Goal: Create a high-end editorial travel photography collage capturing {argument name="city" default="Stockholm"} through warm, cinematic lifestyle and landscape moments, like a premium tourism moodboard.

Canvas: Horizontal 3:2 collage, approximately 1200×800 px, with thin clean white gutters between panels and no outer border or watermark.

Layout: Use exactly 10 photo panels arranged in three rows: top row has 4 equal vertical panels; middle row has 3 panels with the center panel wider than the side panels; bottom row has 3 panels of similar width. Maintain crisp rectangular cropping and consistent white spacing.

Panel count and contents:
1. Top-left: golden sunset over the waterfront skyline, two people sitting from behind in the foreground, leafy bokeh framing the scene, church spire silhouette.
2. Top-second: close-up cafe table with a cinnamon bun and a white coffee cup labeled “Vete-Katten,” shallow depth of field, blurred pedestrians and city street behind.
3. Top-third: winter waterfront panorama with icy water, snow, pastel sky, and Stockholm City Hall tower reflected in the frozen harbor.
4. Top-right: view from a ferry deck with a red life ring labeled “STOCKHOLM,” Swedish flag waving, sparkling water and city skyline in the distance.
5. Middle-left: narrow old-town cobblestone alley with warm ochre buildings, hanging sign reading “ANTIK,” bicycles, flower pots, and a lone woman in a light coat walking away.
6. Middle-center: waterfront city view framed by soft pink cherry blossoms, a white sailboat on bright water, historic skyline and church spires in the background.
7. Middle-right: quiet interior window scene with sheer curtains, wooden chair, vase of wildflowers on a sill or table, book labeled “KINFOLK,” and waterfront skyline visible outside.
8. Bottom-left: rainy street reflection photo, wet pavement mirroring colorful historic buildings and a tall church spire, blurred walking legs at the edge, warm golden highlights.
9. Bottom-center: elevated overlook at {argument name="time of day" default="sunset"}, a lone backpacked person sitting on rocky ground, panoramic islands and waterways below, pink-orange sky.
10. Bottom-right: close-up pink hollyhock or garden flowers backlit by sun flare, soft bokeh, domed historic building and church spire blurred in the background.

Visual style: Photorealistic DSLR travel photography, Scandinavian city atmosphere, natural light, warm golden-hour tones mixed with one cool winter panel, shallow depth of field, realistic lens compression, subtle filmic color grading, clean editorial composition, crisp detail, gentle bokeh, no artificial illustration effects.

Constraints: Include exactly the 10 panels listed above; preserve the visible English/Swedish text on signs and objects; avoid extra captions, logos, people looking at camera, distorted architecture, duplicated landmarks, or messy collage borders.
使用此提示词 ↗
Social Media Post - Golden Phoenix Crown Guofeng Portrait — output

Social Media Post - Golden Phoenix Crown Guofeng Portrait

biuzoneGPT Image 2
Create a vertical 9:16 ultra-realistic Chinese guofeng bridal portrait of {argument name="character name" default="a young noblewoman"} looking back over her shoulder toward the viewer, with her body turned away and her face in a soft three-quarter profile. She has porcelain skin, delicate makeup, glossy dark eyes, soft red lips, and {argument name="hair color" default="black"} hair styled in a low traditional updo with loose wispy strands framing her face. She wears an extremely ornate golden phoenix crown: exactly one large sculpted phoenix with raised wings on the upper right side, one dense filigree crown base covered with pearls and red gemstones, three prominent long vertical tassel clusters hanging on the right, two shorter gold tassels near the temple, and one visible red-and-gold drop earring. Her robe is a luxurious embroidered Chinese wedding garment in {argument name="robe color" default="champagne gold with deep red and muted blue floral embroidery"}, featuring metallic gold thread, large peony blossoms, phoenix-like swirling embroidery on the back, red trim, and heavy brocade texture. Use warm candlelit palace interior lighting with soft amber bokeh in the background, shallow depth of field, cinematic realism, intricate jewelry detail, luminous silk sheen, elegant melancholic expression, high-end fashion editorial composition, close portrait crop from head to upper torso, no text, no watermark.
使用此提示词 ↗
Comic / Storyboard - Anime Café Terrace Friends — output

Comic / Storyboard - Anime Café Terrace Friends

AIossansanGPT Image 2
Create a polished anime-style single-panel illustration of two young women sitting together at an outdoor café terrace in warm late-afternoon sunlight. On the left is {argument name="left character hair color" default="short pale blonde bob hair"}, amber eyes, a cheerful open-mouth smile, leaning forward while holding one tan paper sheet with both hands; she wears an airy cream blouse with puffed sleeves, red ribbon straps/suspenders, a small red hair ribbon with dangling ornament near one side, and a delicate gold necklace. On the right is {argument name="right character hair color" default="long straight black hair"}, violet eyes, a soft reserved smile, looking toward the blonde girl; she wears a dark navy button-up blouse and a subtle pendant necklace, with her arms resting on the wooden table. Include exactly two visible people, exactly one paper sheet, exactly one white teacup on a saucer in front of the right girl, and exactly one small flower arrangement at the lower left edge with white daisies and red flowers. The setting is a cozy European-style terrace with wicker café chairs, a polished wooden tabletop, leafy greenery framing the sides, and a softly blurred townscape in the background with pale buildings and one church-like spire. Use cinematic golden-hour backlighting from the upper left, soft bokeh, gentle rim light on hair, detailed expressive faces, delicate linework, painterly anime rendering, warm inviting atmosphere, and a horizontal 3:2 composition. No text, no watermark, no extra characters, no modern signage.
使用此提示词 ↗
Profile / Avatar - Chibi Pixel Art Fantasy Mage — output

Profile / Avatar - Chibi Pixel Art Fantasy Mage

ち~ぽん✨🤔GPT Image 2
Create a centered chibi fantasy character in crisp 16-bit pixel art on a plain white background. The subject is a cute small mage/priestess named {argument name="character name" default="Astra"}, shown full body, facing forward, with a neutral slightly serious expression, large dark red eyes, pale skin, and short blond hair with straight side locks. She wears an ornate golden headpiece resembling goggles or a circlet across the forehead, with a small red jewel centered above the brow, plus dark ear-covering side pieces. Her outfit consists of exactly 6 main visible clothing/accessory pieces: a dark blue scarf/capelet wrapping around the neck and flowing to the right, a white tunic dress, a gold waist belt with a small rectangular buckle, dark brown gloves or arm guards with gold cuffs, dark boots, and a deep blue cloak or backpack-like mantle behind her sides. In her left hand she holds one tall wooden staff positioned vertically along the left side of the image; the staff has a crooked dark-brown shaft, a round top frame holding a bright cyan crystal, and exactly one small red hanging charm dangling from a short cord below the crystal. Use chunky square pixels, limited color palette, visible black/dark-brown outlines, simple cel shading, no anti-aliasing, no gradients beyond pixel clusters, no text, no watermark, and keep the character occupying most of a square canvas. Customize the hair as {argument name="hair color" default="golden blond"}, the eye color as {argument name="eye color" default="dark red"}, the scarf color as {argument name="scarf color" default="deep navy blue"}, and the staff crystal color as {argument name="crystal color" default="bright cyan"}.
使用此提示词 ↗
Profile / Avatar - Golden Hour Ocean Splash Portrait — output
Infographic / Edu Visual - Four Anime School Reference Panels — output

Infographic / Edu Visual - Four Anime School Reference Panels

さんかくてんGPT Image 2
Goal: Create a 2x2 comparison collage showing four generated anime-style reference images for a short video concept, on a dark charcoal background with large white labels.

Canvas: Landscape 3:2 composition, approximately 1200x900 px. Use a matte black/dark gray background. Arrange exactly 4 rectangular image panels in a clean grid with generous margins and gutters. Each panel has a large bold white label above it: exactly 4 labels, “@image1” at top left, “@image2” at top right, “@image3” at bottom left, and “@image4” at bottom right.

Overall visual style: Soft hand-drawn anime illustration, warm muted colors, gentle line art, subtle watercolor-like shading, nostalgic Japanese school atmosphere, clean character design, no photorealism. The recurring subject is {argument name="character description" default="a cute middle-school Japanese girl with shoulder-length dark brown wavy hair, simple dot eyes, a small smile, and a sailor-style school uniform"}. Her outfit is consistent across panels: cream short-sleeve sailor blouse with navy trim, orange neckerchief, dark plaid pleated skirt with orange and teal lines, navy knee socks with orange stripes, and beige sneakers.

Panel @image1: A single illustration on a warm beige background. Show exactly 1 girl sitting at a classic school desk and chair in side three-quarter view, facing right. She is relaxed with both hands on the desktop as if writing or reading. Include exactly 1 desk and exactly 1 chair, with thin black metal legs and a wooden seat/backrest. Her legs extend forward slightly; sneakers are visible. Keep the background plain and minimal.

Panel @image2: A character turnaround sheet on a clean white background. Show exactly 3 full-body views of the same girl, evenly spaced: 1 front view on the left, 1 side profile view in the center facing left, and 1 back view on the right. Keep the proportions chibi-like but not super-deformed, with a large head, small body, and simple facial features. No extra props or text inside this panel.

Panel @image3: A bird’s-eye view illustration of {argument name="school setting" default="a small suburban Japanese school"}. Show exactly 1 main school building, rectangular and two stories tall, with a dark teal roof and long rows of windows. Include exactly 1 sandy schoolyard with faint white court markings, exactly 1 perimeter wall, and surrounding residential neighborhood houses, trees, narrow streets, utility poles, and warm afternoon lighting. The perspective should be high aerial/isometric, like a reference image for a 3D scene.

Panel @image4: View into a classroom through a large window frame, as if the viewer is outside looking in. In the foreground, show exactly 1 main girl seated at a desk near the window, facing slightly right, with an open book or notebook on her desk. Behind her, show exactly 6 background students seated at separate desks, softly blurred and simplified, facing forward or downward as if studying. Include warm sunlight, window reflections, classroom wall panels, and rows of desks. The window frame should visibly border the panel and create a cinematic shot.

Constraints: Keep all 4 panels in the same illustration style and color palette. Do not add extra panels, extra labels, watermarks, captions, or logos beyond the four @image labels. The collage should look like a social media post showing four prompt-generated reference images.
使用此提示词 ↗
Comic / Storyboard - Fluffy Hamster Holding Japanese Sign — output

Comic / Storyboard - Fluffy Hamster Holding Japanese Sign

🐹🐹GPT Image 2
Create a warm, softly lit storybook watercolor illustration of a very fluffy golden Syrian hamster sitting centered on a pale wooden floor in a quiet artist studio. The hamster is round and oversized, with tiny rounded ears, closed sleepy eyes, a small pink nose, delicate whiskers, tiny front paws, and two little pink feet visible at the bottom. It holds one rectangular cream-colored placard hanging from two small metal rings, with hand-painted Japanese text reading {argument name="sign text" default="私はにじじゃーにーで\n既に200枚を\n使い切りました"}. Surround the hamster with exactly 10 blank picture frames or canvases: 6 on the left side consisting of a tan standing frame, a white standing frame, a tall gray leaning frame, a pale rectangular board behind them, a small square board in front, and one flat sheet on the floor; 4 on the right side consisting of a tan upright frame, a white leaning frame, a small tan rear frame, and one flat white board lying on the floor in the foreground. The background is a minimal pale plaster wall with subtle seams and texture, and a tall window at the upper right casts golden late-afternoon sunlight and soft rectangular shadows across the wall and floor. Use delicate ink outlines, soft watercolor washes, cream and honey tones, airy negative space, gentle melancholic cuteness, and a hand-drawn children’s book aesthetic. Keep the composition square, centered, uncluttered, no extra characters, no watermark, and no additional readable text.
使用此提示词 ↗
Infographic / Edu Visual - Seated Portrait Pose Estimation Overlay — output

Infographic / Edu Visual - Seated Portrait Pose Estimation Overlay

じゅげむGPT Image 2
Goal: Create a photorealistic portrait overlaid with a pose-estimation / 3D block-model visualization, as if analyzing a seated human body with cuboid volumes aligned to a floor grid.

Canvas: Vertical 4:5 image, soft natural daylight, realistic camera photo look. Indoor minimalist living room with pale wall, light wood floor, and a cream rug in the foreground. In the left background, include one wooden shelving unit with white books and woven baskets, plus one leafy green houseplant above it.

Main subject: A young East Asian woman, about early 20s, sitting on the floor facing the camera in a compact curled pose. She has {argument name="hair color" default="dark brown"} long straight hair with wispy bangs, natural makeup, warm eyes, and a gentle slight smile. She wears a loose {argument name="top color" default="cream white"} long-sleeve sweater, blue jeans, white socks, and no shoes. Her knees are pulled up close to her chest, arms wrapped around her shins, feet together near the bottom center, body leaning slightly to one side while maintaining eye contact with the camera.

Overlay: Add a semi-transparent technical 3D pose-estimation diagram over the photograph. Use thin brown-gray wireframe cuboids and translucent teal-blue rectangular prisms to approximate the body volumes. Count and show exactly 13 major cuboid body blocks: 1 head block, 1 neck block, 1 torso block, 1 pelvis/hip block, 2 upper-arm blocks, 2 forearm blocks, 2 thigh blocks, 2 lower-leg blocks, and 2 foot blocks. The blocks should be imperfectly aligned but plausibly matched to the seated pose, with visible perspective lines, overlapping transparent faces, and wireframe edges. Include a faint perspective floor grid in pale blue-purple lines, matching the room floor and receding toward the background.

Label: At the top left, place one rounded white translucent label bar with Japanese text exactly: {argument name="overlay label" default="直方体モデルの重ね合わせ|グリッドと同じ1単位"}. The label should use a small clean sans-serif font, dark gray text, and sit slightly inset from the image edge.

Visual style: realistic high-resolution lifestyle photography blended with analytical AR overlay, soft shadows, shallow depth of field, calm domestic atmosphere, clean composition, no watermark, no extra labels or diagrams beyond the single top-left label and the cuboid/grid overlay.
使用此提示词 ↗
按模型、风格或主题浏览

寻找下一个方向

ASTORIE CANVAS

让灵感,继续生长。

在同一个创作空间,连接提示词、参考素材和生成结果。

打开创作画布 ↗