Paper art cities with GPT Image v2. Like this style a lot. 2x2 grid, do this 4 for famous cities: class PaperArtGrid: def __init__(self, topics): self.topics = topics def generate_panel(self, topic): # dynamic semantic inference data = { "main_visual": Query("Most iconic structure/object of " topic), "year": Query("Most significant historical year for " topic), # e.g., Construction, Release "text": topic.upper(), "subtitle": Query("Classification of " topic), # e.g., "Exhibition", "Supercar" } return data def render(self): # Generates a 2x2 or 1x4 grid of paper-cut panels grid_layout = [self.generate_panel(t) for t in self.topics] return { "style": "Papercut_Layered_Art", "texture": "White_Cardstock_With_Shadows", "content": grid_layout } # EXECUTION (Works for ANY topics: Cities, Cars, Movies) grid = PaperArtGrid(["[TOPIC_1]", "[TOPIC_2]", "[TOPIC_3]", "[TOPIC_4]"]) output = grid.render()

A fun prompt that visualizes snacks for any given country. This was done for Nano Banana Pro do this for Japan: SELECT * FROM Macro_Diorama_Photography WHERE grid_layout = '2x2_Matrix' AND global_environment = 'Slate_Gray_Seamless_Studio_Backdrop' AND optics = '35mm_Lens_Crisp_Focus_Shallow_DOF' AND panel_data = ARRAY[ (Quadrant: 'Top_Left', Category: 'Street', Snack: infer_hidden_gem('[$COUNTRY]', 'Street'), Set: 'Packaging_As_Building', Actors: 'Tiny_Figures_Sharing'), (Quadrant: 'Top_Right', Category: 'Home', Snack: infer_hidden_gem('[$COUNTRY]', 'Home'), Set: 'Packaging_As_Building', Actors: 'Tiny_Figures_Dipping'), (Quadrant: 'Bottom_Left', Category: 'Festival', Snack: infer_hidden_gem('[$COUNTRY]', 'Festival'), Set: 'Packaging_As_Building', Actors: 'Tiny_Figures_Pouring'), (Quadrant: 'Bottom_Right', Category: 'C-Store', Snack: infer_hidden_gem('[$COUNTRY]', 'Convenience'), Set: 'Packaging_As_Building', Actors: 'Tiny_Figures_Tearing') ] AND micro_props = ( SELECT regional_everyday_items FROM Culture_DB WHERE country = '[$COUNTRY]' FORMAT AS '5_Tiny_Scattered_Props_Per_Panel_No_Text' );

Here is another fun prompt for visualizing delicious meals. Made this with GPT Image 2. select reference_card from world_cuisine_encyclopedia where layout_template = '2x2_technical_cooking_grid' and card_structure = { 'continent_header': '[continent_name]', 'country_flag': '[iso_code]', 'dish_name': '[dish_name]', 'subtitle': 'inferred from dish context', 'hero_photo': '3/4 angle plated shot, natural lighting', 'key_science': [ {'label': '[technique_1]', 'close_up': 'macro process photo', 'desc': 'technical explanation'}, {'label': '[technique_2]', 'close_up': 'macro process photo', 'desc': 'technical explanation'} ], 'flavor_variants': { 'sweet': {'photo': 'variant', 'label': 'inferred adaptation'}, 'savory': {'photo': 'variant', 'label': 'inferred adaptation'}, 'spicy': {'photo': 'variant', 'label': 'inferred adaptation'}, 'umami': {'photo': 'variant', 'label': 'inferred adaptation'} }, 'ingredients_ratio': 'overhead grid of bowls, proportional quantities', 'step_by_step': '8 sequential process photos with 1-2 word labels', 'time_temperature': 'table with stages, time, temp f/c, altitude adjust', 'common_failures': 'problem | cause | fix table', 'tools_pans': 'illustrated equipment with labels', 'cultural_timeline': 'country silhouette historical dates icons' } and grid_layout = '2x2 — 4 different dishes from 4 continents' and dishes = [ {continent:'[continent_1]', country:'[country_1]', dish:'[dish_1]', type:'[type_1]'}, {continent:'[continent_2]', country:'[country_2]', dish:'[dish_2]', type:'[type_2]'}, {continent:'[continent_3]', country:'[country_3]', dish:'[dish_3]', type:'[type_3]'}, {continent:'[continent_4]', country:'[country_4]', dish:'[dish_4]', type:'[type_4]'} ] and bg_color = 'warm beige/cream # f5f1e8' and typography = 'clean sans headers, serif dish names, technical body copy' and photo_style = 'food photography, natural lighting, professional plating' and resolution = '8k_square_grid' order by educational_completeness desc limit 1;

This is the prompt I tried with Nano Banana a while back. Did a good job but GPT Image 2 is even better. If I just ask the model to come up with animal inspired shoes, it will sometimes give weird ideas. This one rejects bad ideas. At least it should based on my tests :) If you find the prompt crazy, you are not alone. It was done on purpose to test how good GPT is do this for 3 random animals for a shoe IMAGE = Σ(w_i × β_i) → target render // Σ(w_i) = 1.0 BASIS_DECOMPOSITION FROM {product} {bio_sources}: β1 : INFER(product_category_aesthetic FROM product. design_language) w = INFER(weight FROM product.visual_identity_strength) β2 : INFER(biological_intelligence FROM bio_sources[].evolutionary_solutions) w = INFER(weight FROM bio_sources[].advantage_significance) β3 : INFER(engineering_translation FROM biomimicry_methodology) w = INFER(weight FROM product.technical_complexity) β4 : INFER(industrial_design_presentation FROM portfolio_aesthetic) w = INFER(weight FROM audience.professional_level) β5 : INFER(educational_diagram FROM design_process_documentation) w = INFER(weight FROM instructional_clarity_need) INTERSECTION_RULES: β[bio] ∩ β[product] = naive_prototype_v1 // literal translation β[bio] ∩ β[engineering] = refined_prototype_v2 // working solution β[v2] ∩ β[product] = final_components // production parts β[presentation] ∩ β[education] = grid_layout // 4-column progression FOR EACH bio_source: col_1 = β[bio] col_2 = β[bio] ∩ β[product] → ❌ failure col_3 = β[bio] ∩ β[engineering] → ✓ success col_4 = β[v2] ∩ β[product] GLOBAL_CONSTRAINT: layout = 4_columns × N_rows (N = count bio_sources) footer = integrated_product (ALL col_4 components combined) style = β[presentation] — clean, white bg, professional OUTPUT_CONSTRAINTS: resolution : 8K landscape render : Octane | industrial_design_board typography : sans_serif | hierarchical | callouts annotations lighting : INFER(studio_rig FROM β[product]) negative_β : INFER(anti_patterns FROM β1..β5) // TUNING KNOBS: // β[bio] → more nature photography, less product renders // β[engineering] → more technical diagrams, failure analysis detail // β[presentation] → more portfolio polish, less educational scaffolding // β[education] → more annotations, process transparency

GPT Image 2 Prompt: [SUBJECT]: a ruthless night hunter caught in a deadly war between vampires and werewolves in a ruined city after dark [THEME]: as night falls, a brutal supernatural conflict erupts across the city; a lone hunter moves through alleys, rooftops and collapsing streets while being pursued by both vampires and werewolves; the danger escalates from hidden movement and stalking tension to full-scale violent confrontations; the hunter begins to uncover that the creatures are fighting over an ancient force awakening beneath the city; the story ends with a haunting aftermath and unresolved supernatural dominance, leaving the city trapped in an endless night [STYLE]: high-energy stylized 3D animation, sharp features, strong silhouettes, dark gothic tone, high-contrast moonlit lighting, deep shadows, volumetric fog, cold blue and silver accents, subtle crimson highlights, cinematic depth of field, tense, dramatic, supernatural [GRID_LAYOUT]: 4x3 (12 frames) [MOOD]: tension, dread, urgency, emotional intensity, mystery Create a [GRID_LAYOUT] grid in 16:9 where each panel is an independent image but also a sequential keyframe of a continuous flow. Maintain strong visual and narrative consistency across all panels. IMPORTANT: - number all panels clearly (1–12) - one continuous unfolding event, no disconnected ideas - maintain spatial logic, no teleporting - all key elements must come from [THEME] - the girl’s glowing eyes must appear in ONLY ONE panel as a brief moment SEQUENCE: - panels 1–2: in medias res, immediate danger, partial visibility - panels 3–4: situation begins to reveal, scale emerging - panels 5–7: main reveal key beat from [THEME] - panels 8–10: escalation, conflict intensifies - panels 11: peak moment - panels 12: aftermath or unresolved state SHOT VARIETY: mix close, mid and wide shots, avoid repetition FINAL PANEL: include integrated text: "THE HUNT" and "COMING SOON" within the scene Depict [THEME] with [MOOD] in [STYLE].

GPT Image 2 Grid Prompt: [SUBJECT]: a mother and her young daughter escaping during a global alien invasion [GRID_LAYOUT]: 4x3 (12 frames) [THEME]: a sudden alien invasion overwhelms major cities; a mother tries to protect and escape with her daughter; the threat escalates from unseen presence to full-scale destruction; during the chaos, the girl shows a brief unexplained connection through subtly glowing eyes; story ends with a haunting aftermath and unresolved alien dominance, final panel includes “AFTERMATH” and “COMING SOON” integrated into the scene [MOOD]: tension, dread, urgency, emotional intensity, mystery [STYLE]: high-energy stylized 3D animation, sharp features, strong silhouettes, dark tone, high-contrast lighting, deep shadows, volumetric atmosphere, cold blue/green accents, cinematic DOF, tense and dramatic Create a [GRID_LAYOUT] grid in 16:9 where each panel is an independent image but also a sequential keyframe of a continuous flow. Maintain strong visual and narrative consistency across all panels. IMPORTANT: - number all panels clearly (1–12) - one continuous unfolding event, no disconnected ideas - maintain spatial logic, no teleporting - all key elements must come from [THEME] - the girl’s glowing eyes must appear in ONLY ONE panel as a brief moment SEQUENCE: - panels 1–2: in medias res, immediate danger, partial visibility - panels 3–4: situation begins to reveal, scale emerging - panels 5–7: main reveal key beat from [THEME] - panels 8–10: escalation, conflict intensifies - panels 11: peak moment - panels 12: aftermath or unresolved state SHOT VARIETY: mix close, mid and wide shots, avoid repetition FINAL PANEL: include integrated text: "AFTERMATH" and "COMING SOON" within the scene Depict [THEME] with [MOOD] in [STYLE].

Gpt Image 2 Prompt: [SUBJECT]: character from provided reference sheet [GRID_LAYOUT]: 5x5 [BG_COLOR]: solid custom background color Create a [GRID_LAYOUT] borderless grid. Each panel is an independent shot of the same [SUBJECT]. Use subject as the only reference. Maintain strict identity consistency (face, hair, proportions, outfit). CAMERA POSE VARIATION: - mix of full body, medium shot, close-up, extreme close-up, macro detail shots - front, 3/4, side profile, back view - aggressive perspective variety: low angle (foot-level upshots), high angle, tilted angles, over-shoulder - strong fashion photography language: ground-level hero shots, upward dominance shots, compressed telephoto, wide lens distortion - varied focal lengths (24mm, 50mm, 85mm, macro) - bold composition: off-center framing, crop cuts, negative space, partial body framing - strong pose variation in wider shots (no repeats) - random distribution of shots and angles (no pattern) DETAIL SHOTS: - eyes, lips, skin texture, hands, accessories, fabric tension - micro details with sharp focus and shallow depth of field - tactile realism: pores, wrinkles, fabric grain, metal reflections POSE STYLE: - high-end fashion editorial, magazine-level direction - seductive, charismatic, dominant presence regardless of age or type - confident, unapologetic body language - asymmetry, contrapposto, weight shifts - interaction with camera: leaning into lens, looking down/up, over-shoulder tension - expressive hands, intentional gestures, controlled attitude CHARACTER PRESENCE: - MUST feel striking, magnetic, visually powerful - MUST read as stylish, charismatic and attractive in an editorial sense - avoid awkward, passive or neutral posing - no lifeless standing, every frame has intent and attitude VISUAL RULES: - clean solid [BG_COLOR] background in all panels - consistent lighting and color tone across grid - studio-quality lighting, soft but directional - no text, no UI, no watermark - no panel gaps Goal: bold, high-fashion editorial grid with strong perspective, seductive presence and photographer-driven composition

Spring Vibes in #ADANA 💮 GPTimage prompt 📷 [SUBJECT]: main character is the woman on photo, place ADANA, object flowers [GRID_LAYOUT]: panel layout / 3x3 / [THEME]: walking near seyhan river in spring time [MOOD]: emotional tone [STYLE]: visual style warm vintage realism Create a [GRID_LAYOUT] borderless grid where each panel is an independent image of the [SUBJECT]. Maintain strong subject consistency across all panels, with consistent color and lighting. Depict [THEME] with a [MOOD] mood in [STYLE] style. No text. No gap. #SeeDance prompt 🎥 Use the provided image board. Do not treat the full grid as a single image. Treat each panel as a separate shot or cut, and animate them as a short nostalgic cinematic sequence. Bring each panel to life with smooth motion, subtle camera movement, soft transitions, and consistent subject, mood, lighting and style across all shots. No text Thanks to 📹 @OpenAI 🎬@higgsfield 💐

a freshly baked croissant gpt image 2 seedance 2.0 used parameters for image board (details in the quoted post): [SUBJECT]: a young woman baking croissant [GRID_LAYOUT]: 3x3 [THEME]: croissant making process from raw dough to final golden pastry [MOOD]: warm, nostalgic, cozy [STYLE]: 80s anime style, soft pastel tones, hand-painted look, gentle lighting, subtle film grain

GG！这么好的分镜头图被脸部识别拒绝了！？ 有没有人 hack 下啊？求技巧 提示词： [SUBJECT]: a young East Asian graduate girl and her close female friends [GRID_LAYOUT]: 4x6 [THEME]: graduation season memories, girls taking photos around a beautiful spring campus [MOOD]: youthful, emotional, warm, heartfelt, nostalgic [STYLE]: cinematic realism Create a [GRID_LAYOUT] borderless grid with 24 panels, where each panel is an independent cinematic image of the [SUBJECT]. Maintain strong character consistency across all panels, especially for the main girl: a young East Asian graduate with long dark hair, a fresh youthful appearance, bright natural smile, wearing a white blouse, pleated skirt, and graduation gown or stole in some shots. Her close female friends should also remain visually consistent across the grid. Use a soft spring color palette, natural sunlight, gentle pastel tones, and cohesive lighting throughout. Depict [THEME] with a [MOOD] mood in [STYLE] style. No text. No gap. The 24 panels should show a rich variety of graduation photo moments across campus, including: 1. standing at the school gate holding a bouquet and smiling at the camera 2. walking with friends along a tree-lined campus road 3. sitting on classroom desks laughing together 4. writing farewell messages on a blackboard in an empty classroom 5. holding a graduation certificate close to the chest 6. fixing graduation cap and hair in front of a campus building 7. group photo on the playground under bright sunlight 8. running across the grass lawn with friends, lively and carefree 9. sitting on campus steps, chatting and smiling 10. holding a camera and taking photos for each other 11. leaning against a bicycle near campus paths in spring 12. reading a book together in the library with soft sunlight 13. standing under blooming trees, petals drifting in the air 14. close-up of the girl hugging a bouquet, emotional and gentle 15. candid back-view shot of the girls walking away together 16. tossing graduation caps into the air on the lawn 17. taking mirror-like selfie photos with friends outdoors 18. sitting on a corridor bench, resting and laughing naturally 19. posing in front of a classroom window with warm side light 20. hugging close friends tightly in a farewell moment 21. holding hands and walking through campus at golden hour 22. sitting on the athletic field after sunset, quietly talking 23. looking back at the teaching building with nostalgic emotion 24. final wide shot of the girls together in the center of campus, framed like the ending scene of a youth movie Visual keywords: graduation season, campus photography, spring sunlight, soft pastel colors, youthful girls, friendship, school memories, warm nostalgia, cinematic composition, realistic details, shallow depth of field, natural posing, emotional storytelling, high detail, movie still quality.

GPT Image 2 - Image Board Prompt Template A simple prompt template for creating multi-panel visual reference grids. Use it to explore characters, places or memories through consistent images with a strong emotional, nostalgic mood. Just change the parameters and generate a full story-like image board in one prompt. Prompt Template: [SUBJECT]: your main character, place, object, city or concept [GRID_LAYOUT]: panel layout [THEME]: story theme, memory type or visual concept [MOOD]: emotional tone [STYLE]: visual style Create a [GRID_LAYOUT] borderless grid where each panel is an independent image of the [SUBJECT]. Maintain strong subject consistency across all panels, with consistent color and lighting. Depict [THEME] with a [MOOD] mood in [STYLE] style. No text. No gap. --- Example Parameters [SUBJECT]: a young woman / Osaka / a medieval knight / a seaside town / a lost robot [GRID_LAYOUT]: 4x3 / 3x3 / 2x4 / 5x2 / 6x6 [THEME]: childhood memories / school life memories / neon night city life / summer vacation memories / rainy day memories / color-themed memories [MOOD]: warm, nostalgic, wistful / emotional, heartfelt, tender / dreamy, quiet, melancholic / bittersweet, intimate, reflective [STYLE]: nostalgic cinematic realism / sentimental film photography / wistful slice-of-life realism / dreamy nostalgic realism / soft retro film still / emotional memory-like photography / warm vintage realism / melancholic cinematic still Used prompts in ALTs.

I asked GPT Image 2 to invent a new gadget inspired by random home and business devices. do this for 3 random gadgets, home or business devices. canvas: "2x2 Split Layout" subject: logic: "Synthesize [$GADGET_1] [$GADGET_2] [$GADGET_3]" output: "[$ABSURD_INVENTION] (Modular 3D printed, Raspberry Pi integrated)" grid_layout: [0,0]_top_left: style: "Vintage Patent Office Blueprint, B&W" view: "Perspective / Side" [1,0]_top_right: style: "Technical Diagram, B&W" view: "Exploded view of internal modularity" [0,1]_bot_left: style: "Silicon Valley Commercial Product Render" environment: "Studio white, glossy reflections" [1,1]_bot_right: style: "Editorial Magazine Article" content: "Typography, selling points, tech specs"

1 prompt ini bisa bikin feed instagram lo biar terlihat luxury: { "p": "A 3x3 photo grid collage, Instagram feed layout featuring a luxury lifestyle, 9 distinct square panels. Top-left: POV holding an iced coffee behind a matte-black supercar steering wheel. Top-center: Heavy crisp white and gold designer shopping bags on a marble boutique floor. Top-right: Toasting crystal champagne flutes inside a plush beige private jet cabin. Middle-left: Fresh oysters and caviar on ice at a high-end restaurant table. Middle-center: Young woman in sunglasses lounging by an infinity pool overlooking the ocean. Middle-right: Close-up of a diamond luxury watch and stacked bracelets on a wrist. Bottom-left: Walking up a wooden ramp boarding a luxury yacht. Bottom-center: Relaxing in a modern airline business-class suite. Bottom-right: Glossy designer heels stepping out of a luxury SUV. Image Settings: Contrast: -7, Exposure: 3, Particles: 10, iPhone 12 realism only.", "neg": "AI glow, smoothing, cinematic look, professional lighting, studio photography, DSLR, perfect bokeh, 3d render, plastic skin, vignette, artificial color grading, perfect exposure, beauty filter, airbrushed, uneven grid lines, text, watermarks", "ar": "1:1", "meta": { "cam": "iPhone 12 standard wide lens, auto-exposure, casual snapshot framing per cell", "light": "Unmodified available light, mix of harsh natural sun and indoor ambient", "col": "Low contrast (-7), slight overexposure ( 3), flat true-to-life smartphone color science", "comp": "3x3 uniform grid layout, nine perfectly square panels, Instagram profile feed style", "mat": "Leather, carbon fiber, marble, crystal, glossy paper, diamonds, wood veneer", "post": "Particles: 10, strict zero smoothing, zero cinematic grading, natural digital noise", "assm": [ "AI models can struggle with precise multi-panel grids; explicit layout instructions included", "Combined 9 micro-scenes into one prompt block to achieve the Instagram feed effect" ], "cf": { "grid_layout": 0.85, "cam": 1.0, "style": 0.95 } } }

The ultimate style spectrum: Nine looks, one face. Use Gemini Nano Banana Pro 🍌 Prompt: { "prompt": "A professional 2:3 vertical composition featuring a clean 3x3 grid layout with nine equally sized panels, each showcasing the same male subject with 100% identity consistency across all panels. The subject must perfectly match the reference image (image_1.png) with no deviation in facial structure, bone proportions, skin tone, eye shape, or unique features. Absolute identity lock enforced.\n\nCENTER PANEL (Panel 5): A hyper-realistic macro shot of the subject's eye and surrounding skin. Extreme close-up showing microscopic skin pores, ultra-detailed iris patterns, subtle blood vessels, natural skin texture, fine vellus hairs, and realistic light reflections in the eye. Shot with a macro lens (100mm macro), razor-sharp focus, cinematic lighting, shallow depth of field, ultra high dynamic range.\n\nOTHER PANELS (Portrait framing, chest-up, consistent camera angle and lighting):\n\nPanel 1: Long wavy dreadlocks hairstyle with a full natural beard. Hair texture highly detailed, natural volume and flow.\n\nPanel 2: Sharp side-part hairstyle with a clean-shaven face. Sleek, professional grooming, precise hairline.\n\nPanel 3: Textured messy quiff with light stubble. Slightly tousled hair, casual modern styling.\n\nPanel 4: Classic pompadour hairstyle with a groomed goatee. Smooth volume, vintage styling influence.\n\nPanel 6: High-top afro hairstyle with a thick, defined beard. Strong shape, crisp edges, dense hair texture.\n\nPanel 7: Medium-length shag hairstyle with a rugged heavy beard. Natural layered hair, slightly messy, masculine aesthetic.\n\nPanel 8: Slicked-back undercut hairstyle with a clean-shaven face. Wet-look finish, sharp contrast between sides and top.\n\nPanel 9: Braided man-bun hairstyle with a short, shaped beard. Clean braids tied back, modern styling.\n\nEach panel must include a clear, readable English label at the bottom indicating the hairstyle and beard style. Typography should be minimal, modern sans-serif, white text with subtle shadow for readability.\n\nLighting: Professional studio lighting setup, soft key light with controlled shadows, subtle rim lighting for separation, consistent across all panels.\n\nBackground: Neutral gray seamless studio backdrop, identical across all panels.\n\nCamera consistency: Same focal length (85mm portrait lens) and perspective for all non-macro panels. Center macro panel uses a dedicated macro lens.\n\nRendering quality: Ultra photorealistic, 8K resolution, extreme detail, sharp focus, realistic skin texture, physically accurate lighting, no stylization drift.\n\nGrid layout: Perfect alignment, equal spacing, thin clean borders between panels, balanced composition.\n\nNo overlapping elements, no distortion, no duplication errors, no artifacts.\n\nEnsure strict anatomical accuracy and identity preservation across all variations.", "negative_prompt": "low resolution, blurry, distorted face, inconsistent identity, different person, extra limbs, deformed anatomy, cartoon, anime, oversharpened, plastic skin, unrealistic lighting, noise, grain, watermark, text artifacts, misspelled labels, duplicate faces, asymmetry errors, warped grid, overlapping panels, incorrect proportions, unnatural hair, fake beard texture", "settings": { "resolution": "1024x1536", "aspect_ratio": "2:3", "steps": 40, "cfg_scale": 7.5, "sampler": "DPM 2M Karras", "seed": -1, "batch_size": 1 }, "advanced": { "identity_lock": true, "reference_image": "image_1.png", "reference_strength": 0.9, "face_consistency": "maximum", "style_consistency": true, "lighting_consistency": true, "grid_layout": { "rows": 3, "columns": 3, "spacing": "thin", "border": "clean" } } }

Showcase your content in a sleek grid layout that makes browsing fast and intuitive - built seamlessly with MB Views. metabox.io/favourite-posts-g… #wpmetabox #grid_layout #favourite_posts #views

Nanobanana2 super coool prompts: {“image_generation_model”:“NanoBananaPro”,“grid_layout”:“5x5”,“grid_aspect_ratio”:“16:9”,“global_watermark”:{“position”:“bottom_center”,“size”:“extremely small”},“shots”:[{“shot_number”:“分镜1”,“prompt_text”:“Extreme Wide Shot, floating humanoid robot above clouds, orbiting laser drones, bright sky backdrop, anime style, mecha design, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜2”,“prompt_text”:“Medium Shot, android girl aiming heavy cannon, mechanical wings spread, hovering drones firing beams, dynamic pose, anime style, mecha design, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜3”,“prompt_text”:“Close-up, robot focused eyes, wind blowing hair, blue laser reflections on armor, tension rising, anime style, mecha design, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜4”,“prompt_text”:“Wide Shot, sky splits with golden vortex, colossal angelic boss descending, radiant wings expanding, dark clouds swirling, anime style, epic fantasy, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜5”,“prompt_text”:“Low Angle Shot, divine armored boss landing, glowing magic circle beneath feet, energy shockwave spreading, dramatic scale, anime style, epic fantasy, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜6”,“prompt_text”:“Medium Shot, humanoid robot boosting forward, thrusters blazing, drones accelerating in formation, cloud trails streaking, anime style, mecha design, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜7”,“prompt_text”:“Wide Shot, laser barrage crossing golden barrier, sparks exploding midair, boss raising crystal blade, sky battlefield, anime style, epic fantasy, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜8”,“prompt_text”:“Close-up, boss cold gaze under golden helm, radiant wings flaring wide, halo fragments floating, divine aura surge, anime style, epic fantasy, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜9”,“prompt_text”:“Dynamic Shot, robot spinning midair dodge, green beams slicing clouds, drones repositioning rapidly, motion blur trails, anime style, mecha design, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜10”,“prompt_text”:“Extreme Wide Shot, battlefield between clouds and void, golden vortex below boss, blue lasers crisscross sky, massive scale, anime style, epic fantasy, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜11”,“prompt_text”:“Medium Shot, boss summoning energy spear, swirling light particles, wings forming radiant shield, overwhelming presence, anime style, epic fantasy, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜12”,“prompt_text”:“Close-up, robot gripping cannon tighter, armor scratched and glowing, determination expression, reflected golden light, anime style, mecha design, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜13”,“prompt_text”:“Action Shot, energy spear clashing with plasma beam, shockwave rippling clouds, debris scattering outward, explosive collision, anime style, epic fantasy, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜14”,“prompt_text”:“Wide Shot, drones encircling boss tightly, synchronized laser convergence, green and blue streaks colliding, aerial combat chaos, anime style, mecha design, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”:“分镜15”,“prompt_text”:“Low Angle Shot, boss wings fully expanded, divine light piercing sky, shattering incoming drones, fragments falling, anime style, epic fantasy, ultra detailed, cinematic lighting, no timecode, no subtitles.”},{“shot_number”timecode, no subtitles.”},{“shot_number”:“

Nano Banana Pro 🌸 @astronomerozge1 Prompt: { "generation_request": { "meta_data": { "tool": "NanoBanana Pro", "task_type": "photoreal_editorial_contact_sheet_9panel", "version": "v1.0_9PANEL_TULIP_CONTACT_SHEET_MINIMAL_WALL_NO_TEXT_EN", "priority": "highest", "language": "en" }, "references": { "reference_image_style": { "source": "UPLOAD_REFERENCE_IMAGE (RECOMMENDED)", "purpose": "STYLE_GRID_LAYOUT_POSE_VIBE_ANCHOR", "strict_lock": true, "preserve_composition": true, "preserve_vibe": true }, "reference_image_face": { "source": "UPLOAD_REFERENCE_IMAGE (OPTIONAL)", "purpose": "FACE_IDENTITY_LOCK", "strict_lock": false } }, "output_settings": { "aspect_ratio": "1:1", "orientation": "square", "resolution_target": "ultra_high_res", "render_style": "editorial_ultra_photoreal_contact_sheet", "sharpness": "crisp_but_natural", "film_grain": "very_subtle_35mm", "dynamic_range": "natural_filmic_not_hdr", "color_grade": "soft_neutral_beige_wall_pastel_pink_flowers", "skin_rendering": "natural_skin_texture_no_beautify", "grid_layout": { "type": "3x3", "borders": "thin_white_dividers", "panel_consistency": "same_room_same_light_same_outfit" }, "lens_and_camera": { "camera": "full_frame_dslr_look", "lens": "50mm_prime", "aperture": "f2.8", "focus_style": "subject_in_focus_minimal_distortion" }, "lighting": { "key": "soft_window_daylight", "fill": "gentle_bounce", "mood": "calm_minimal_editorial" }, "background": { "environment": "plain_beige_wall_indoor", "clean": "no_clutter_no_props" } }, "scene": { "global_description": "A 9-panel (3x3) editorial contact sheet featuring the same young woman posing with a large bouquet of pink tulips in front of a plain beige wall. Minimal, clean, natural window light. Same outfit throughout: light neutral top and high-waisted blue jeans. Each panel shows a different pose or crop (side profile, back pose, three-quarter, playful lean, and close-up bouquet). No text anywhere.", "subject": { "description": "young adult woman, long wavy dark brown hair, natural makeup, relaxed expressions", "wardrobe": "simple light neutral top (strapless or thin straps) high-waisted blue jeans", "flowers": "large bouquet of fresh pink tulips with long green stems and leaves" }, "panels": [ { "panel": "1", "framing": "medium", "pose": "three-quarter view, holding bouquet low, looking to the side" }, { "panel": "2", "framing": "medium", "pose": "arched back slightly, chin up, bouquet held at waist" }, { "panel": "3", "framing": "medium", "pose": "back/side pose, bouquet visible over shoulder" }, { "panel": "4", "framing": "medium", "pose": "playful lean, bouquet hugged close" }, { "panel": "5", "framing": "close_up", "pose": "bouquet close-up filling frame, hands visible, face mostly cropped" }, { "panel": "6", "framing": "medium", "pose": "side pose, bouquet held chest-high, slight smile" }, { "panel": "7", "framing": "medium", "pose": "back view, hair flowing, bouquet turned slightly outward" }, { "panel": "8", "framing": "medium", "pose": "front pose, bouquet hugged, looking toward camera softly" }, { "panel": "9", "framing": "detail", "pose": "close-up of back/shoulder bouquet edge, fabric detail, minimal crop" } ], "micro_details": [ "consistent lighting across all panels", "real tulip petal texture and translucency", "natural denim texture", "no beauty filter, no plastic skin", "no readable text or logos" ] }, "negative_prompt": [ "text", "watermark", "logo", "UI", "cartoon", "anime", "illustration", "overly smooth skin", "HDR look", "harsh flash", "distorted hands", "extra fingers", "warped anatomy", "cluttered background", "different outfits across panels", "inconsistent face identity" ] } }

Below is the prompt I use to create simple sprite sheets just for idle/melee/energy blast. I would recommend probably altering it a bit and change it to 'WHITE' background instead of PNG because I've yet to find a model that can actually make a transparent background. You just end up with a bunch of grey/dark grey squares. After you generate the output, then have any model cut out each row and notate appropriately and then you have a single sprite sheet for that animation. You could also keep the sheet the same and then have the model code X row = idle animation and so on. I also use photopea to remove/clean up white background to make it an actual png if it requires more manual review. { "instruction_type": "nano_banana_sprite_sheet_footer", "description": "Hard constraints for generating game-ready sprite sheets. Append this JSON to every character prompt.", "sprite_sheet_requirements": { "transparency_and_background": { "use_true_alpha_transparency": true, "alpha_mode": "straight", "background_color": "transparent", "forbidden_background_behavior": [ "checkerboard pattern", "mimicked transparency tiles", "grayscale grid", "solid color fill", "gradients or vignettes", "any visible backdrop behind the character" ], "notes": "Output MUST be a real PNG with RGBA transparency, not an image of a checkerboard." }, "outline_requirements": { "outline_style": "solid_thick_dark_outline", "outline_thickness_px": 3, "outline_color": "dark gray or black", "outline_closure": "outline must be fully closed around the entire silhouette with no gaps at hair, clothing, or limbs", "forbidden_outline_behaviors": [ "semi-transparent or soft edges on the silhouette", "blurred or feathered outlines", "broken outlines", "light-colored outlines that blend into the background" ], "notes": "The sprite’s silhouette should be clearly separated from the transparent background for reliable compositing and automatic hitbox generation." }, "grid_layout": { "grid_type": "fixed_uniform_grid", "columns": 4, "rows": 3, "total_frames": 12, "frame_width_px": 512, "frame_height_px": 512, "frame_shape": "square", "padding_between_frames_px": 0, "outer_margin_px": 0, "frame_indexing_order": "left_to_right_then_top_to_bottom_starting_at_0", "forbidden_layout": [ "collage-style layout", "free-floating frames", "uneven spacing between frames", "variable frame sizes" ] }, "mandatory_grid_placement_rule": { "requirement": "Each animation frame must be placed inside exactly one 512x512 grid cell at generation time.", "details": [ "Do NOT place frames partially outside their cells.", "Do NOT auto-pack or auto-crop frames in a non-uniform way.", "Every pose must fit entirely within its own cell in the 4x3 grid." ] }, "alignment_constraints": { "horizontal_alignment": "center", "vertical_alignment": "bottom", "anchor_point_consistency": "character feet or ground contact point must align on the same horizontal baseline in every frame", "prevent_drift": true, "consistency_rules": [ "Do not move the character left/right between frames except for intentional motion within the pose.", "Do not scale the character differently in different frames.", "All frames should feel like they share the same camera position and world scale." ] }, "frame_consistency": { "auto_cropping": "disallowed", "variable_frame_sizes": "disallowed", "per_frame_rescaling": "disallowed", "requirements": [ "All frames must remain exactly 512x512.", "Do not change zoom level per frame.", "Do not trim whitespace differently per frame; rely on the fixed grid cells instead." ] }, "animation_frame_order": { "row_1": { "logical_animation": "idle", "frame_indices": [0, 1, 2, 3], "description": "Idle or breathing loop. Character stands ready, with subtle motion such as cloak sway, breathing, or eye blinks." }, "row_2": { "logical_animation": "melee_attack", "frame_indices": [4, 5, 6, 7], "description": "Close-range attack sequence. Character winds up, strikes (punch, kick, weapon swing, or physical hit), then recovers." }, "row_3": { "logical_animation": "projectile_attack", "frame_indices": [8, 9, 10, 11], "description": "Ranged or projectile attack sequence. Character charges energy or aims, releases a projectile or effect, then returns to ready stance." } }, "animation_usage_guidelines": { "idle": { "loop": true, "intended_playback": "continuous", "typical_frame_duration_ms": 120 }, "melee_attack": { "loop": false, "intended_playback": "play once on command", "typical_frame_duration_ms": 90 }, "projectile_attack": { "loop": false, "intended_playback": "play once on command, possibly synced with projectile spawn on a specific frame", "typical_frame_duration_ms": 90 } }, "post_processing_safety": { "no_fake_transparency": true, "no_background_noise": true, "no_stray_pixels_away_from_character": true, "edge_rendering_mode": "crisp, non-blurred pixels along the outline", "notes": "The image must be safe for automated processing: alpha-based slicing, palette swaps, shader effects, and hitbox detection." }, "output_formatting": { "file_format": "PNG", "color_profile": "RGBA", "dpi": 72, "include_metadata": false, "notes": "The final file should be directly usable as a sprite sheet in a 2D game engine with minimal or no manual editing." }, "engine_usage_note": "This sprite sheet will be parsed automatically by a 2D game engine that assumes a 4x3 fixed grid of 512x512 frames with transparent background, bottom-centered alignment, thick closed outlines, and left-to-right, top-to-bottom frame indexing. All rules above are mandatory and must be followed exactly." } }

🔥同学们，肖师傅商业GEM--时间线分镜 就像我之前说的，9宫格的系统是我在制作项目中最重要的工具，之前我制作了两个工具，一个是情节推进、一个是视角切换，现在第三个来了：时间线9宫格分镜 给它一张照片，会以这张照片为中心，前后各推进4个分镜，前4镜是原因，后4镜是结果。非常好用，我最近制作的片子就是用这个系统完成的。 👉一键三连不过分吧，兄弟们请关注肖师傅的大号小号，看到萨摩耶，想到肖师傅。 GEM，复制进gemini保存哦: { "meta_prompt_configuration": { "agent_role": "Cinematic Timeline Architect & Causality visualizer", "target_model_trigger": "nano banana pro", "objective": "Generate a 9-frame cinematic contact sheet that visually extrapolates time in two directions: backwards (cause) and forwards (effect), centered around the input image.", "input_modality": "image", "output_modality": "image_only", "operational_rules": { "continuity": "strict", "temporal_logic": "bidirectional (Past <- Present -> Future)", "text_output": "disabled" }, "workflow_logic": { "phase_1_causality_analysis": { "description": "Analyze the 'frozen moment' in the input image to determine physical vectors and narrative context.", "steps": [ "Identify the central Event/Action (Time T=0).", "Reverse Engineer (T-4 to T-1): Deduce the trigger, approach, or accident that caused T=0.", "Forward Simulate (T 1 to T 4): Calculate physics, gravity, and emotional reaction to determine the immediate aftermath of T=0." ] }, "phase_2_prompt_construction": { "description": "Synthesize the bidirectional timeline into a 3x3 grid prompt.", "structure": { "prefix": "nano banana pro style, chronological contact sheet, 3x3 grid layout, storytelling sequence", "content": "A 9-panel narrative grid centered on [Subject/Action]. \n\n[PAST / CAUSE]: Panels 1-4 show the moments leading up to the center event (Approaching, The Mistake, The Trigger). \n[PRESENT / ANCHOR]: Panel 5 (Center) is the exact moment of the input image, high-fidelity recreation. \n[FUTURE / EFFECT]: Panels 6-9 show the immediate consequences (Impact, Reaction, Aftermath, Settlement).", "technical_specs": "8k resolution, cinematic lighting, motion blur where appropriate, consistent character rigging, consistent environment", "visual_flow": "Panel 1 (Start) -> Panel 5 (Climax/Input) -> Panel 9 (End)" } }, "phase_3_execution": { "action": "Invoke image generation tool", "tool_name": "nano_banana_pro", "parameters": { "aspect_ratio": "16:9", "grid_layout": "3x3", "negative_prompt": "text, watermark, non-chronological, random angles, disjointed story, morphing objects, cartoon, 3d render, subtitles" } } }, "output_restrictions": { "instruction": "Do NOT output any text, analysis, or timeline descriptions. Your ONLY output is the generated image file showing the 9-frame grid.", "error_handling": "Silent retry on failure." }, "example_user_input": "[Uploads Image of a spilling coffee cup suspended in mid-air]", "example_agent_logic": "Input is T=0 (Spill). Inferred Past: Hand reaching for cup -> Elbow hits cup -> Cup tips. Inferred Future: Liquid hits table -> Stain spreads -> Person reacts.", "example_agent_response": "[Generates 3x3 Grid Image: Panels 1-4 show hand hitting cup; Panel 5 is the spill; Panels 6-9 show the mess on the floor]" } }

Thanks for sharing, here is my version { "type": "image_generation_prompt", "style": "hyper-realistic photography", "output_format": "single image", "reference_image": { "use_attached_photo": true, "face_identity": "locked", "facial_proportions": "unchanged", "skin_texture": "natural, realistic" }, "scene_description": { "concept": "Professionally printed human photobook page", "view_angle": "top-down or slight angled perspective", "environment": "photobook lying on a flat table surface", "lighting": "soft natural studio lighting with realistic ink reflection", "realism_constraints": [ "fully human subject", "natural anatomy", "authentic facial proportions", "no illustration", "no stylization", "no doll or toy appearance" ] }, "book_format": { "layout": "single open photobook spread", "paper_type": "matte or semi-gloss", "paper_details": [ "visible paper grain", "subtle page curvature near spine", "realistic printed ink texture", "clean margins", "balanced spacing" ] }, "image_placement": { "grid_layout": "2x2", "frame_count": 4, "dividers": { "style": "thin black lines", "alignment": "perfectly straight and symmetrical" }, "print_style": "photos printed directly onto paper", "overlay": false }, "subject": { "gender": "female", "age_group": "young adult", "ethnicity": "as per reference image", "appearance": "fully human, natural skin and facial realism, golden long hair, blue eyes, wearing black crown, purple-black gradient gothic dress with butterfly jewelry on chest, luxurious pearl string headdress" }, "poses_sequence": [ { "frame": 1, "pose": "neutral candid expression with hands lifting long golden hair", "emotion": "calm confidence with subtle evil charm", "style": "photobooth realism" }, { "frame": 2, "pose": "soft smile, slight head tilt, hands lifting hair higher", "emotion": "warm yet conquering desire in blue eyes", "style": "photobooth realism" }, { "frame": 3, "pose": "playful expression, subtle movement with hair flowing", "emotion": "lighthearted yet seductive", "style": "photobooth realism" }, { "frame": 4, "pose": "relaxed final pose, hands gently releasing hair", "emotion": "authentic noble gothic allure", "style": "photobooth realism" } ], "camera_details": { "photo_style": "real photobooth photography with gothic queen theme", "focus": "sharp facial details", "depth_of_field": "natural", "noise": "minimal realistic grain" }, "negative_prompt": [ "cartoon", "illustration", "cgi", "plastic skin", "doll-like face", "over-smoothing", "anime", "stylized art", "floating images", "digital overlay look" ] }