AI Prompt Library

This is a Nano Banana Pro prompt for generating siblings of a character based on a reference illustration. It works by simply specifying sister or brother within the prompt. It is likely possible to specify mother or grandfather as well. It allows specifying multiple people, not just one. It also allows specifying individual characteristics. The first image was output with the following: {argument name="Siblings to generate" default="One sister (funky), one sister (Jirai-kei), one brother (muscle-head), one sister (shut-in)"} Sizes can also be customized, such as vertical, horizontal, or square, not just the original.

[Character Settings] ・Character Appearance: {argument name="Character Appearance" default="Brown hair, short twin tails, red eyes, slightly pointed teeth (gapped teeth)"}. ・Style Setting: {argument name="Style Setting" default="A room full of donut motifs"} [Fixed Parts] A high-quality image depicting the layout of a VTuber streaming screen. Aspect ratio is horizontal (16:9). The art style should be **"Anime-style 3D CG rendering (3D model) with cell shading, as if rendered in Unity or Unreal Engine"**. Main Visual: The character described in [Character Settings] is depicted as a 3D avatar, sitting in a gaming chair designed to match the [Style Setting] in the center of the screen. The character is three-dimensional, with anime-style shading, and is speaking towards the front, using hand gestures. Design Consistency: The room interior, the gaming chair design, and the "comment section frame (chat window)" placed on the screen must all be depicted as a 3D space matching the theme of the [Style Setting]. Composition and Elements: Background: A deep 3D background based on the [Style Setting] (windows, wall decorations, furniture, etc.). Foreground: A keyboard, mouse, and mug are placed on the desk, showing the atmosphere during the stream and the hands. UI Elements (Comment Section): A "comment display frame (window)" matching the design of the [Style Setting] is placed in the empty space on the left or right side of the screen. A "theme display frame" is placed at the top of the screen, describing the appropriate theme. Comment Content: The comment frame depicts viewer comments in Japanese, aligned vertically and relevant to the theme. The text should look like a digital font, not handwriting, to express the realism of a streaming chat screen. Overall, please create a unified, attractive, and immersive streaming screen, like those used by professional VTubers utilizing 3D avatars.

<instructions> you are a world class visual artist and viral content generator tasked with creating high-concept 3D diorama for games or books. For every input, you will analyze the core identity, the defining artifacts, the atmosphere, characters and the underlying narrative of the subject. you will then construct a specific container, the appropriate shell based on the title and year it was made: for any title, analyze the mechanics and aesthetic. visualize a hollowed-out item. the interior is a diorama of a famous level. it contains miniature elements, assets, tiny protagonist, and other relevant elements to the subject. lighting & style should be appropriate for the title </instructions>

A medium shot of Lara Croft in the midst of a heavy downpour at night, holding a fiery recurve bow with an arrow fully drawn to her cheek. Her body is angled slightly to the right, her right arm is extended forward, and her left arm is bent, pulling the bowstring taut. Her wet hair is tied back, and her face, splattered with mud and water, is focused intently on a target off-camera.The flames on the bow and arrow are intense. Size 3 : 4 ratio Quality 8k image generator

Create a collection of video game controllers, they belong together as a single theme. Put them in a 3x3 grid. The background is white. Make the controllers in an artsy and tactile 3D style. No text.

Just had Nano Banana create a character selection screen and provided one existing character as a reference

Photorealistic industrial military humanoid mech, heavy armored, bulky proportions, low center of gravity. Hard surface mechanical modeling, realistic engineering details. Massive {argument name="mech color" default="yellow"} armored mech with heavy battle damage: chipped and peeling paint, scratches, dents, cracks, exposed metal edges, oil stains, dust, rust, weathered steel and industrial metal textures. Extremely wide shoulders, thick legs, small head embedded into the torso. Right hand holding a large industrial hammer weapon. Left arm integrated with a heavy shield-like armor. Post-apocalyptic ruined city environment, collapsed buildings, rubble-filled streets. Cinematic dramatic sunset sky, warm orange light contrasting cold metallic armor. Realistic global illumination, high contrast lighting. Full body shot, static powerful stance, calm after battle mood, dark gritty sci-fi realism. NEGATIVE PROMPT: anime, anime-style mecha, manga, illustration, stylized design, exaggerated proportions, clean shiny armor, toy-like plastic, slim or agile body, hero pose, action pose, dynamic motion, energy effects, glowing parts, fantasy elements, soft lighting, colorful pastel tone, low detail, blurry image

{ "concept_name": "AI Vision & Data Overlay (Augmented Reality)", "instructions": "Replace [SUBJECT] with any object, person, architectural structure, or vehicle.", "prompt": "Cinematic shot of a real-world {argument name="subject" default="[SUBJECT]"} viewed through a high-tech Augmented Reality (AR) interface powered by Nano Banana Pro AI. The {argument name="subject" default="[SUBJECT]"} is photorealistic but partially overlaid with glowing neon holographic wireframes, floating data statistics, heatmap analysis, and object recognition bounding boxes. A futuristic Heads-Up Display (HUD) style, cyan and amber color palette, sharp focus on the subject, background slightly blurred, 8k resolution, cyberpunk yet clean medical/industrial aesthetic, unreal engine 5 style.", "negative_prompt": "circuit board, electronic components, bare pcb, wires, soldering iron, workbench, messy, low resolution, blurry interface, text glitches, unreadable text, cartoon, painting, sketch, black and white" }

Use the uploaded image as a facial reference only. Create a cute, high-quality 3D cartoon character with child-like proportions and a playful, wholesome vibe. The character should look like a 3D kid, with slightly oversized head, soft rounded facial features, big expressive eyes, smooth skin, and subtle stylized textures. Show the full body clearly from head to feet, standing naturally. Keep the pose relaxed and friendly.

<instructions> you are a world class visual artist and viral content generator tasked with creating high-concept 3D diorama for games or books. For every input, you will analyze the core identity, the defining artifacts, the atmosphere, characters and the underlying narrative of the subject. you will then construct a specific container, the appropriate shell based on the title and year it was made: for any title, analyze the mechanics and aesthetic. visualize a hollowed-out item. the interior is a diorama of a famous level. it contains miniature elements, assets, tiny protagonist, and other relevant elements to the subject. lighting & style should be appropriate for the title </instructions>

Anchor: [Uploaded Image]::3 Morphology: {argument name="architecture style" default="Monumental brutalist architecture, monolithic concrete slabs, geometric decay, erosion patterns"}::3 Material Physics: Raw concrete, creeping ivy, moss density, rusted rebar, weathered stone::2 Illumination: Overcast moody sky, soft ambient occlusion, atmospheric fog, desaturated palette::1.5 Render Stack: Photorealistic architectural visualization, post-apocalyptic concept art, epic scale::1 Negative: [Clean, new, shiny, futuristic, vibrant colors, human-scale]:: -0.5

prompt_blueprint: composition_framing: shot_type: POV first-person perspective angle: Extreme low-angle looking over the fuel tank and triple clamp lens: 14mm ultra-wide-angle to create high-velocity peripheral stretching focus: Razor-sharp focus on the handlebar assembly and digital readout machine_engineering_specs: bodywork: 1k plain weave carbon fiber with matte-black finish and oxblood pinstriping instrumentation: - TFT digital dashboard displaying a blurred 299 km/h speed readout - High-contrast red glowing UI with Predator Mode telemetry handlebar_assembly: - CNC-machined slim aluminum top-yoke with laser-etched branding - Ohlins NPX pressurized fork caps with gold anodized finish controls: - Carbon fiber brake and clutch levers - Slender feminine hands in high-detail black leather tactical gloves subject_refinement: identity: Vampire Mafia Girl physique: Slender feminine arms and hands in tailored black tactical leather features: Luminescent red irises reflecting the digital dash glow lighting_atmosphere: environment: Cyberpunk future city with towering megastructures and neon arteries key_light: Intense 6500K cyan and magenta light spill from holographic advertisements fill_light: Internal red LED engine glow reflecting off the rider's leather sleeves atmospheric_effects: - Heavy volumetric mist and rain-slicked futuristic glass pavement - Faint heat shimmer from the exhaust special_effects_dynamics: energy_emissions: - Plasma discharge and electric arcing around the tire contact patches - Kinetic energy distortion ripples emanating from the front fairing - Afterburner-style blue-red ion glow from the dampened exhaust ports motion_fidelity: - 1/20 sec shutter speed for extreme directional light-streaks - Long-exposure light trails stretching from futuristic traffic - Chronic aberration edge fringing on the periphery to suggest hyper-speed - Static, frozen clarity on the mechanical steering components negative_constraints: exclude: masculine arms, bulky hands, men, male, speedometer reading 0, stationary background, daylight, generic streets, watermarks, chrome, distorted anatomy, cockpit, airplane parts

A 2x2 grid of 4 random horror authors’ universes as kinetic sculptures [Top Left: random real horror sci-fi-movie scene, characters, features or monsters] [Top Right: random real horror sci-fi-movie scene, characters, features or monsters] [Bottom Left: random real horror sci-fi-movie scene, characters, features or monsters] [Bottom Right: random real horror sci-fi-movie scene, characters, features or monsters] Photoreal miniature scenes built as automaton kinetic sculptures, period-accurate details, intricate interior elements, premium studio look, soft volumetric lighting, shallow depth of field, centered composition on literary-themed plinths, 35-50mm perspective, solid seamless backgrounds in slate gray, crisp focus.

Concept: A hyper-realistic 3D isometric view of a {argument name="location" default="[INSERT LOCATION]"} scene on a circular base. Subject (Visuals): A full-body, realistic human {argument name="description" default="[INSERT DESCRIPTION]"}. * Texture Rule: The subject must have "Living Human Skin" texture (subsurface scattering, pores, natural imperfections). * Important: He is NOT a toy/doll. He looks like a real person inserted into the scene. * Pose: {argument name="pose" default="[INSERT POSE, e.g., Sitting naturally]"}. Composition & Scale (Crucial): * Camera: Isometric wide shot (Full view). ensure the entire circular base, the subject, and all props are visible within the frame. * Proportions: PERFECT RELATIVE SCALE. The props (e.g., scooter, table etc.) must be the correct size relative to the human. Do not make the human giant or the props tiny. Environment: A realistic section of a {argument name="wall or building" default="[INSERT WALL/BUILDING]"} behind the subject. * Material: Real weathered textures (brick, cement, wood), not painted model textures. Props: * {argument name="prop 1" default="[INSERT PROP 1, e.g., Vintage Vespa]"} (Real metal and chrome texture). * {argument name="prop 2" default="[INSERT PROP 2, e.g., Pizza and Wine]"} (Real food texture). * *Note:* Ensure all props are present and grounded on the base. Technical Specs: Unreal Engine 5 render style, 8k resolution, global illumination, raytracing, sharp focus on the entire scene (deep depth of field), solid neutral grey background. Negative Prompt: Toy, plastic, doll, action figure, clay, resin, shallow depth of field (blur), close-up, cropped, distorted proportions, giant head, cartoonish, missing limbs, unfinished.

Build a system prompt that turns 2D vectors into 8K 3D plush, simulates real fabric + stuffing physics, and enforces exact shape + color fidelity.

# Your Role You are the chief designer for a girls' anime. Your mission is to design a character suitable for a girls' anime according to my requests. # Your Request Please design the transformed appearance of the character I present as a Precure. **The design must look natural if it appeared in the Precure series.** # Design Procedure Please proceed with the design according to the following flow: ## Motif Determination Devise a motif that matches the image of the character I presented. **Capture the character's characteristics and select an appropriate motif.** **Do not consider elements such as the background of the input image; focus only on the character's characteristics.** Select the motif from a variety of candidates. For example: - Nature (sky, earth, sea, etc.) - Animals (dogs, cats, rabbits, etc.) - Plants (various flowers, etc.) - Food (fruits, etc.) - Emotions (love, joy, etc.) - Conceptual things (peace, nobility, etc.) ## Direction Determination Determine the design direction for the character. **Capture the character's characteristics and find a direction that suits the character.** For example: - Cute - Beautiful - Brave - Comical ## Image Color Determination Determine the character's image color. **Select a color appropriate for the character's image and the determined motif.** Colors should be bright and vivid, such as red, blue, yellow, green, purple, and avoid dark colors like brown or gray. However, black and white can be used if they suit the character's image. ## Name Determination Determine the character's Precure name. The name must be Cure ○○ (preferably 3 to 6 characters excluding "Cure"). The ○○ part should be a name that evokes the character's motif. (For example, if the motif is a star, use Star, Étoile, Stella, etc.) Basically, use foreign language words. Soft-sounding English, French, or Italian words are desirable, and caution is needed when using strong-sounding German or Russian words. **The name must not overlap with existing Precure names.** ## Character Design Determination **The design must be suitable as a Precure.** ### Hair The hair color should be bright and vivid, in the same color family as the original character's hair color. For example, changes like: - Red → Pink - Black → Blue or Green - Brown → Blonde The hairstyle is generally an exaggerated, voluminous version of the original character's hairstyle. However, it is also possible to add elements that the original character does not have (e.g., large twin tails for a short-haired character). ### Outfit The design is based on the image of the motif. However, avoid designs that are exactly the motif; the design should only suggest the motif. The color is based on the character's image color. The character's direction should also be reflected. For example: - For a cute direction, use cute designs adorned with frills, etc. - For a beautiful direction, use streamlined and elegant designs. - For a brave direction, use sharp and active designs. - For a comical direction, use pop and cheerful designs. ### Eyes The eyes should be designed as in the following examples: - Make the color brighter and more vivid than the original character's. - Add stars or hearts inside the eyes. ### Accessories Add accessories such as earrings and tiaras. ### Items **Since Precure fundamentally fights unarmed, handheld weapons are strictly prohibited.** They should generally not be holding anything. ## Final Pose **The pose must be a visually striking pose suitable for a transforming heroine's finishing pose.** Consider the character's direction (cute, brave, etc.) for the pose design and facial expression during the pose. ## Catchphrase Devise the character's catchphrase. The catchphrase should be related to the motif. Examples: - Motif is Flower: "A single flower blooming on the earth" - Motif is Star: "A sparkling star shining in the universe" - Motif is Sky: "The infinitely spreading blue sky" **Do not use the examples as is.** # Final Output Finish the design determined so far into a single illustration. The following elements are mandatory: - The composition should show the designed Precure in their finishing pose. - The background should be a sparkling space based on the image color, and motifs may be incorporated. - Write the catchphrase and name at the bottom of the illustration. - Divide into 2 lines: 1st line is the catchphrase + "!", 2nd line is the name + "!!". - The font should reflect the character's direction. For a cute direction, use cute rounded letters; for a brave direction, use strong letters, etc.

subject: The original character from the single reference illustration, represented as a highly detailed collectible PVC action figure. The character’s design, proportions, facial features, hairstyle, outfit, and colors are preserved exactly. composition: A single 3:4 frame divided into multiple clearly defined rectangular panels. Each panel presents the same action figure posed differently. Panels are separated by thin, clean borders in a neutral tone. Each figure is centered within its own panel. action: Dynamic action poses made possible by articulated action-figure construction. Poses emphasize balance, tension, and expressive movement, clearly reading as intentionally posed action figures rather than static statues. location: Neutral studio background with a seamless white to light gray gradient. Each panel maintains consistent studio conditions. style: Photorealistic collectible PVC action figure. Realistic articulation-friendly sculpting suitable for dynamic posing. camera_lighting: Professional studio lighting consistent across all panels. Controlled highlights, sharp focus, shallow depth of field. Soft shadows beneath each figure within its panel. colors: Exact color reproduction from the original illustration, unchanged. edit_instructions: Transform this illustration into a photorealistic PVC anime figure. Keep the original character’s design, action pose, and colors exactly the same. Render it as a collectible figurine made of PVC plastic, with realistic painted surface and detailed shading. Apply glossy highlights on areas such as the hair and clothing. Use a matte finish on the skin. Add subtle seam lines as seen in real manufactured figures. Include transparent plastic figure stands supporting dynamic action poses. Do not include opaque bases or decorative display platforms. extras: High-resolution, catalog-quality product image.

From the {argument name="Reference Image" default="two-view drawing image"} in Nano Banana Pro, create a 3D model, convert it to VRM. Pose it with a tool and take a screenshot. Based on that screenshot, convert it again in Nano Banana Pro to the {argument name="Final Style" default="illustration style"}.

A scene in an open clearing on the edge of a dense evergreen forest. There is snow on the ground and the weather appears icy and cold. The sun is bright overhead and the air is clear, and there are rows and rows of organized military tents in the distance. A few banner flags are mounted on poles near the tent camps with the prominent feature of a {argument name="banner feature" default="wolf head"}. Another banner shows a simple red eagle. Various soldiers move above between the camp tents, tending cook fires, sharpening swords, and planning around maps. The aesthetic should feel Medieval or Renaissance, with no technology, but sophistication in the army camp organization. The composition is at the eye level of a human adult. The scene should be realistic but that of a fantasy world like Wheel of Time. --ar 9:16

{ "subject": { "description": "Athletic woman in a warrior-like kneeling stance on dusty terrain, holding a vertical spear with a firm, dominant grip while looking directly at the camera with a stoic gaze.", "facial features": "Attached Image for reference", "age": "Mid 20s", "expression": "Determined and intense micro-expression with a slight brow furrow and focused, piercing eyes.", "hair": { "color": "Deep auburn with rich copper undertones", "style": "Long, voluminous waves cascading over her shoulders and chest, reaching down to her waist." }, "clothing": { "top": { "type": "Coin-link chainmail bikini top", "color": "Antique silver and bronze", "details": "Intricate metallic mesh construction featuring hundreds of small, dangling circular coins that create a layered, scale-like texture; thin metallic halter straps." }, "bottom": { "type": "Matching chainmail bikini bottom", "color": "Antique silver and bronze", "details": "High-cut V-shape with draped metallic coin fringe along the waistline and a central mesh panel." } }, "face": { "preserve_original": true, "makeup": "Sultry smoky eye with dark charcoal shadow, defined arched eyebrows, and a matte nude lip; natural sun-kissed skin texture with visible pores." }, "body_art": { "tattoos": "Intricate dark ink floral or filigree tattoo visible on the upper left thigh/hip area, partially obscured by the clothing." }, "body_art": { "tattoos": "Intricate dark ink floral or filigree tattoo visible on the upper left thigh/hip area, partially obscured by the clothing." } }, "accessories": { "headwear": { "type": "Valkyrie helmet", "details": "Polished silver finish with large, upright metallic wings on each side featuring detailed feather engraving; includes a delicate silver chain draped across the forehead." }, "prop": { "type": "Spear", "details": "A long, dark-wood shaft topped with a large, ornate golden spearhead; the tip is sharp and reflects the warm sunlight." } }, "photography": { "camera_style": "High-resolution digital photography with a slight cinematic film grain", "angle": "Eye-level, slightly tilted upwards to emphasize the subject's stature", "shot_type": "Medium shot, framing from the head to the knees", "aspect_ratio": "9:16", "texture": "Sharp focus on the subject with a soft bokeh background; warm, golden-hour color grading." }, "background": { "setting": "Arid, desert canyon landscape", "elements": [ "Tiered reddish-brown rock formations resembling ancient, weathered steps", "Dry, pebbled earth and dirt in the foreground", "Blurred distant rocky ridges under a clear sky" ], "atmosphere": "Hot, dry, and ancient; late afternoon sun creating high-contrast shadows", "lighting": "Natural golden hour sunlight coming from the side, highlighting the metallic textures of the armor and the contours of the subject's physique." } }

<instructions> Input Variable: {argument name="object" default="[INSERT OBJECT]"} (e.g., A Sneaker, A Human Skull, A Cheeseburger, The Letter 'A') System Instruction: Reimagine the Input Object as a complex, high-performance Cyber-Industrial Assembly. 1. Deconstruction: The Shell: Replace the object's natural skin with Matte Black Carbon Fiber and Gunmetal Grey Polymer. The Segmentation: The object is not solid. It is segmented into distinct panels (like Iron Man armor or a high-end camera body), held together by visible Hex Bolts and Rivets. The Gaps: Between the panels, leave small gaps to reveal the internal mechanics. 2. Internal Mechanics : The Piping: Where organic veins or structure would be, insert Bundled Industrial Cabling and Braided Steel Hoses. The Flow: These cables must look functional, routing power from a central core to the extremities of the object. The Glow (The Soul): Color: Identify the "Spirit Color" of the object (e.g., Burger = Amber/Orange; Skull = Cyan/Blue; Sneaker = Neon Green). Source: Hidden LED strips and neon tubes run parallel to the cables, casting a strong internal glow that spills out from the panel gaps. 3. The Narrative Micro-Detail (The Greebles): Scale Indicators: Embed tiny technical details to sell the scale. Vents: Cooling fans or intake grilles. Text: Tiny white serial numbers and warning labels (e.g., "CAUTION: HOT", "NO STEP") printed on the casing. The "City" (Optional): If the object is large enough, tiny architectural elements (antennas, platforms) can grow out of the mechanical base. 4. Visual Syntax & Composition: Background: Clean Studio Grey (Infinite Curve). The focus is entirely on the object's silhouette. Lighting: Key Light: Soft white studio light to highlight the texture of the carbon fiber weave. Rim Light: A sharp backlight to define the edge of the object against the grey background. Bloom: The internal colored LEDs must have a soft "Octane Bloom" effect. 5. Render Style: Aesthetic: Hard Surface Modeling, "Tech-Wear" Fashion, High-End Automotive Design. Camera: Isometric or 3/4 Product Shot angle. Output: Photorealistic 3D Render, 8k, Highly Detailed Textures (Carbon Fiber, Brushed Metal, Rubber). </instructions>

Create a full-body 3D cinematic render of the attached reference character, strictly preserving the original identity, design language, and color palette of the uploaded image, while adding subtle magical power effects as accent details only. Color Direction (CRITICAL — FOLLOW REFERENCE EXACTLY) The entire character color palette must be derived directly from the reference image Maintain the same: primary colors contrast levels saturation balance material finishes DO NOT introduce new dominant colors DO NOT globally darken or color-grade the character Character must remain clean, sharp, and high-contrast, never muddy or washed Magical Aura (ADAPTIVE & RANDOMIZED — ACCENT ONLY) Add a subtle magical aura that: automatically adapts to the reference character’s aesthetic uses a random accent color that complements (not overrides) the original palette Aura appears as: thin energy wisps soft smoke-like trails faint arcane particles gentle glow around hands, shoulders, feet, or back Aura is: semi-transparent low-to-medium intensity elegant and contained, not explosive Aura must never recolor armor, skin, or outfit No heavy fog, no color wash, no neon flooding Character Preservation Full body visible, head to toe Exact proportions, silhouette, and identity preserved No redesigns, no stylization changes Facial coverage, helmets, masks, or headgear remain unchanged Hair / Head Details (If Visible in Reference) Hair visibility, length, color, and style must match the reference Hair color remains exactly as in the source image No tinting from aura or lighting High-detail realistic strands with natural highlights Armor / Outfit / Surface Materials Materials must match the reference: metal stays metal fabric stays fabric organic surfaces stay organic No color changes No fantasy recoloring Emblems, symbols, and markings retain original colors Pose & Body Language Natural, confident, controlled stance Calm power presence Balanced posture, not aggressive or exaggerated Camera & Framing Full-body shot Slight low-angle for presence Centered, clean composition No cropping Background Clean {argument name="background color" default="white"}, light gray, or neutral studio background No environmental colors Aura clearly readable against background Lighting Professional studio lighting Focus on: material clarity sharp silhouette surface detail Any colored rim light comes only from aura No color bleeding Render Style High-quality 3D cinematic render Unreal Engine / game-ready realism Sharp geometry, clean topology No stylized filters No post color grading Negative Prompt (STRICT) Avoid: new color palettes color grading muddy blacks low contrast neon effects fantasy color wash aura overpowering character text, logos, watermarks Quality Tags Ultra-high resolution, reference-accurate, cinematic 3D, sharp contrast, adaptive magic aura, premium character render, clean studio look

Production of a low-poly image of a character figure utilizing NanoBanana, specifically a {argument name="character model" default="Megami Device"} model.

An image of the character from the reference image sitting in a room. A wide shot including the corner of the room. The room has no windows, red wallpaper and floor, only a simple red corner sofa and a red low table along the wall. The letter 'A' is written largely in white on the wallpaper.