Prompt Library · Updated April 2026
Real prompts that work in Midjourney, DALL-E, and Stable Diffusion — with explanations of why each one works
Every prompt below is copy-paste ready. Click “Copy Prompt” under any card. For Midjourney: paste after /imagine. For DALL-E / ChatGPT: paste directly into the chat. For Stable Diffusion: paste into the positive prompt field. Parameters like --ar 16:9 are Midjourney-specific — remove them for other tools.
Why some prompts are labeled for specific tools
Midjourney responds best to short punchy phrases with parameters. DALL-E understands natural sentences. Stable Diffusion uses comma-separated keywords with weights. Where a prompt works best on one tool, it’s tagged — but most prompts in this library work across all three with minor adjustments.
Category 01
Portrait Prompts
Portraits are the most-generated category in AI art. The difference between a flat result and a striking one almost always comes down to lighting and specific facial detail. These prompts lead with both.
Cinematic Film Portrait
close-up portrait of a woman with freckles, golden hour sunlight streaming through window blinds creating striped light patterns across her face, shallow depth of field, film grain, Kodak Portra 400 color science, natural expression, photorealistic, 85mm lens –ar 4:5 –v 7
Why it works: Naming a specific film stock (Kodak Portra 400) gives the model a precise color palette to target — warm skin tones, lifted shadows. The window blind detail creates directional light without vague terms like “dramatic.”
Classical Oil Painting Portrait
oil painting portrait of an elderly man with deep wrinkles and kind eyes, Rembrandt lighting, dark warm background, visible brushstrokes, warm umber and sienna tones, classical realism, museum quality –ar 3:4 –v 7 –s 200
Why it works: “Rembrandt lighting” is a term the model understands precisely — one strong side light, triangle shadow under the eye. Combined with specific pigment names (umber, sienna), this anchors the palette firmly.
Fashion Editorial
high fashion editorial portrait, model wearing structured geometric clothing, stark white studio background, dramatic directional lighting, sharp focus, Vogue magazine aesthetic, clean and minimal, medium format camera, 70mm lens –ar 3:4 –v 7 –s 300
Why it works: “Vogue magazine aesthetic” and “medium format camera” are specific enough that the model knows exactly what visual language you’re targeting — without you needing to describe every detail of that style.
Watercolor Character Study
watercolor portrait of a young child laughing, loose wet-on-wet technique, colors bleeding naturally at edges, minimal background wash, light pencil sketch visible underneath, warm and joyful palette, spontaneous expressive brushwork
Why it works: “Wet-on-wet technique” and “colors bleeding naturally” are specific watercolor terms that push toward a painterly, loose result rather than a digital approximation of watercolor, which tends to look overworked.
Victorian Wet Plate Photography
formal Victorian portrait in wet collodion process style, 1860s, distinguished woman in her 50s, high-collared dark silk dress, center-parted silver-streaked hair in chignon, dignified expression, ornate carved chair, painted romantic landscape backdrop, natural north-window studio light, warm sepia and brown albumen print tones –ar 3:4 –v 7
Why it works: Historical photographic processes have very distinctive visual signatures the model knows well. “Wet collodion,” “albumen print,” and “north-window studio light” all pull toward authentic period aesthetics rather than generic sepia filter.
Double Exposure Silhouette
woman’s profile silhouette filled with a twilight cityscape — car light trails, illuminated skyscraper windows, neon signs within her outline, hair dissolving into wisps of cloud, deep navy to peach gradient background, double exposure photography style, digital art blend, ultra-clean edges –ar 2:3 –v 7 –s 400
Why it works: Double exposure is a technique the model understands as a photographic concept. Specifying what goes inside the silhouette (the city) and what the background does (gradient) prevents the AI from inventing random elements.
Category 02
Landscape Prompts
The gap between a flat landscape and a breathtaking one is almost always time of day and atmospheric detail. Every prompt here specifies both. Skip those and you get generic midday blue-sky scenery.
Alpine Sunrise
panoramic alpine landscape at sunrise, jagged snow-capped peaks piercing a sea of clouds, golden light on highest summits while valleys remain in cool blue shadow, glacial lake in foreground reflecting orange-pink sky with mirror stillness, dark boulders breaking the water surface, 14mm ultra-wide lens, high dynamic range photography –ar 21:9 –v 7
Why it works: The warm-top / cool-bottom light contrast (golden peaks, blue valley shadow) is specific and real. It prevents the AI from defaulting to uniform lighting. The glacial lake mirror adds visual depth without adding complexity.
Stormy Coastline
stormy coastline with towering sea cliffs, massive waves crashing against dark basalt rocks, ocean spray and mist in the air, overcast sky with one dramatic break of golden light on the horizon, long exposure smoothing the water, 10-stop ND filter aesthetic, cinematic and powerful –ar 16:9 –v 7 –s 300
Why it works: “10-stop ND filter aesthetic” tells the model to smooth the water into silky streaks — a specific photographic look. “One dramatic break of light” prevents the flat grey overcast that makes stormy landscapes look dull.
Northern Lights Aurora
northern lights filling the sky with sweeping curtains of electric green and deep violet above a frozen Arctic lake, aurora reflected with crystalline clarity creating a perfect mirror effect, dark snow-laden pine silhouettes on shore, Milky Way peeking through gaps in the aurora, pure darkness and total stillness, ultra-wide shot –ar 16:9 –v 7 –s 500
Why it works: The frozen lake mirror doubles the visual impact without making the composition complicated. Specifying both the aurora and the Milky Way tells the model these should coexist without competing — which it handles well.
Japanese Autumn Garden
traditional Japanese garden in peak autumn, curved red wooden bridge over a calm koi pond, brilliant red and orange maples reflected in dark mirror-like water, white and gold koi beneath the surface, moss-covered stepping stones, thatched tea house among bamboo, soft overcast light saturating autumn colors, low bridge-level perspective, 16:9 –ar 16:9 –v 7
Why it works: “Overcast light saturating colors” is the key insight — overcast sky acts as a giant softbox, making autumn colors more vivid than harsh sun. “Low bridge-level perspective” creates an intimate composition versus the typical overhead tourist shot.
Desert Minimalism
minimalist desert landscape, single sand dune with perfect curved ridge, deep shadows on one side warm golden sand on the other, vast empty space, clean composition, fine art landscape photography, no other elements, 4:5 ratio –ar 4:5 –v 7 –s 150 –style raw
Why it works: The constraint “no other elements” is doing the heaviest lifting here. Most AI landscape prompts add clutter — clouds, rocks, plants. Explicitly forbidding them creates the clean minimalist result.
--style raw reduces Midjourney’s tendency to stylize.
Foggy English Forest
dense woodland at dawn, ancient oak trees with twisted roots, heavy morning fog filtering pale grey-gold light through the canopy, narrow muddy path disappearing into mist, dew on ferns and moss-covered stones, eerie and mysterious atmosphere, cinematic photography, rule of thirds composition –ar 16:9 –v 7 –s 250
Why it works: Fog is one of the most effective atmospheric elements in landscape prompts because it naturally creates depth, mystery, and soft light in a single word. The specific tree detail (twisted roots) prevents a generic neat forest.
Category 03
Fantasy Prompts
Fantasy prompts work best when you treat them like world-building documents, not wish lists. Specify scale, light source, and one concrete unusual element. The model fills the rest with appropriate genre detail.
Enchanted Ancient Forest
ancient enchanted forest with massive trees hundreds of feet tall, shafts of ethereal green light filtering through the canopy, glowing mushrooms and fireflies at ground level, a narrow mossy path winding into deep shadow, mystical atmosphere, fantasy concept art, detailed digital painting –ar 9:16 –v 7 –s 400
Why it works: Establishing scale (“hundreds of feet tall”) is the key decision here. It transforms an ordinary forest into something monumental. The light shaft detail creates natural focal depth without requiring complex composition instructions.
Dragon’s Lair
massive dragon sleeping on a mountain of gold coins and gemstones inside a vast underground cavern, bioluminescent crystals on cave walls casting blue and purple light, smoke curling from the dragon’s nostrils, treasure glinting in the crystal glow, epic fantasy illustration, detailed scales and textures, wide establishing shot –ar 21:9 –v 7 –s 600
Why it works: The bioluminescent crystal light source is the standout choice — it’s more interesting than “torchlight” or “glowing lava” and creates an unusual blue-purple palette inside a traditionally warm scene. Always specify your light source in dark interior scenes.
Floating City
grand floating city suspended above the clouds by enormous chains, waterfalls pouring off the city edges into the void below, airships docked at crystal spires, sunset casting long golden shadows across the stone architecture, fantasy worldbuilding illustration, epic scale, matte painting style –ar 21:9 –v 7 –s 700
Why it works: “Matte painting style” signals to the model that this should feel like a hand-crafted film backdrop — a specific aesthetic with soft edges, atmospheric haze, and deliberate composition. The chains are a concrete anchor detail that prevents generic floating-island clichés.
Elf Warrior Character
a mystical elf warrior wielding a sapphire-tipped enchanted spear, standing beneath an ancient gnarled tree in a magical mist-filled forest, backlit by ethereal light, intricate silver armor with nature motifs, high fantasy character concept art, detailed, 8K –ar 9:16 –v 7 –s 250
Why it works: Backlighting a character against a bright background creates instant visual drama — the character silhouettes against the misty glow. “Nature motifs” on the armor is a small specific detail that the model handles surprisingly well and makes the character feel internally consistent.
Sorceress Casting Spell
a wise sorceress with long curled brown hair, narrow green eyes, lace mask covering her lower face, flowing emerald cloak, casting a powerful spell with an enchanted wand, illuminated by the glow of a full moon through storm clouds, dark fantasy illustration, dramatic upward angle –ar 9:16 –v 7 –s 300
Why it works: The “dramatic upward angle” (low camera looking up) makes the character feel powerful and monumental. The lace mask is an unusual specific detail that distinguishes this from generic fantasy mage art — give your characters one distinctive thing.
Underground Library
underground library carved into living rock, bioluminescent plants providing warm amber light, massive stone bookshelves reaching 30 feet high, a tiny human figure in the foreground for scale, leather-bound books of every color, spiral stone staircases, detailed fantasy concept art, digital painting –ar 16:9 –v 7 –s 700
Why it works: The tiny human figure for scale is the critical element — it makes “30 feet high” actually feel 30 feet high. Without it, the model generates bookshelves that look normal sized. Always include a scale reference when you need vast spaces to feel vast.
Category 04
Cyberpunk & Sci-Fi Prompts
This genre lives and dies on reflections and light sources. Wet asphalt, neon signs, holographic surfaces — these are the vocabulary. Every prompt here anchors the visual with at least one specific source of colored light.
Rain-Soaked Neon Street
rain-soaked cyberpunk street at night, towering holographic advertisements in Japanese and English, neon reflections on wet cobblestones, a lone figure with an umbrella walking past steaming food stalls, atmospheric fog at street level, dense urban crowd in background, Blade Runner visual language, volumetric fog, cinematic wide shot –ar 21:9 –v 7 –s 500
Why it works: “Blade Runner visual language” is an extremely well-established reference the model understands precisely — dark, rain-slicked, neon-heavy, morally complex. Using cultural touchstones as style anchors is one of the most efficient prompt techniques available.
Space Station Observation Deck
interior of a massive space station observation deck, floor-to-ceiling windows revealing a gas giant planet filling the entire view, bioluminescent plants in hydroponic planters, clean futuristic furniture, soft ambient lighting reflected off polished surfaces, a solitary astronaut silhouetted against the planet, hard sci-fi aesthetic, cinematic composition –ar 16:9 –v 7 –s 400
Why it works: “Hard sci-fi aesthetic” signals scientifically grounded design — no fantasy spaceships, no excessive chrome. The gas giant filling the entire window creates an immediately striking visual contrast between interior warmth and cosmic scale.
Alien World Bioluminescent
alien planet landscape with bioluminescent flora, three moons visible in a purple-green sky, crystal formations growing from the ground, an exploration rover parked near a cliff edge for scale, strange atmospheric haze, NASA concept art style, scientifically grounded alien world, detailed terrain –ar 16:9 –v 7 –s 600
Why it works: “NASA concept art style” anchors this in a specific aesthetic tradition — the kind of scientifically informed illustration you’d see in a planetary science publication. It prevents the common AI failure of making alien worlds look like generic fantasy landscapes with different colors.
Cyborg Character Portrait
close-up portrait of a female cyborg, half-face organic half-face exposed mechanical components, glowing cyan circuitry lines beneath translucent skin, neon-lit rain on the mechanical half, intense expression, dramatic rim lighting, cyberpunk character concept art, detailed, ultra-realistic –ar 2:3 –v 7 –s 350
Why it works: The 50/50 split (half organic, half mechanical) is a classic but effective compositional rule for cyberpunk character design. Specifying cyan for the circuitry lines creates contrast against the warm skin tones. Rim lighting separates the subject from dark backgrounds.
Cyberpunk Street Market
cyberpunk street market at night, rain-slicked cobblestones reflecting neon signs in Japanese and English, street food vendor stalls with steam rising into cold air, diverse crowd of augmented humans, atmospheric fog at street level, overhead tangled cables and lanterns, cinematic wide shot –ar 21:9 –v 7 –s 400
Why it works: Ground-level steam and overhead cables create visual layers — something to look at in the foreground, middle, and background. Cyberpunk scenes without layering feel flat. “Augmented humans” signals the genre without requiring specific technical descriptions.
Retro-Futurist Space Scene
retro-futurist 1960s space illustration, astronaut in a bulky white suit on a lunar surface, Earth visible in a star-filled black sky, streamlined rocket on the horizon, vintage NASA recruitment poster aesthetic, bold flat graphic style, limited color palette of red white and blue, clean lines –ar 4:5 –v 7 –s 200
Why it works: “Vintage NASA recruitment poster aesthetic” gives the model a very precise visual target — bold, optimistic, graphically clean. The limited palette instruction (“red white and blue”) prevents the model from defaulting to full-color photo realism in a style that should be graphically restrained.
Category 05
Anime & Illustration Prompts
Anime prompts work best when you name the studio or director whose aesthetic you want. “Studio Ghibli” and “Makoto Shinkai” pull toward completely different visual results — both beautiful, both distinct. General terms like “anime style” produce generic output.
Studio Ghibli Countryside
lush green rolling countryside on a windy summer afternoon, a young girl in a white dress running through tall grass, small farmhouse in the distance, fluffy clouds casting moving shadows across the hills, warm golden afternoon light, Studio Ghibli animation style, hand-painted background quality –ar 16:9 –niji 6
Why it works: For Midjourney,
--niji 6 switches to a specialized anime model that produces far better illustrated results than the standard model with “anime style” added. For other tools, “Studio Ghibli animation style, hand-painted background quality” is the key phrase pair.
Makoto Shinkai Twilight
a young woman standing on a train platform at twilight, warm amber light from station lamps, telephone wires against a gradient sky of deep blue and orange, light rain beginning to fall, solitary and contemplative mood, Makoto Shinkai cinematic anime style, highly detailed background, lens flare –ar 16:9 –niji 6 –s 300
Why it works: Makoto Shinkai’s visual signature includes hyper-detailed backgrounds, atmospheric light, telephone wires, and lens flare. All four are in this prompt. The contrast between the warm lamp light and the cool twilight gradient is specifically his palette.
Cherry Blossom Scene
a young woman in a flowing white summer dress beneath a massive cherry blossom tree in full bloom, one hand reaching to touch a low branch, petals drifting in a gentle spiral, long dark hair with pink petals caught in the strands, a small white cat at her feet, peaceful Japanese riverside park, soft diffused sunlight, anime illustration –ar 3:4 –niji 6
Why it works: The small white cat at her feet is a classic Ghibli compositional element — small companion animals create warmth and scale simultaneously. The petal-in-hair detail is visual poetry that AI handles surprisingly well when stated literally.
Botanical Watercolor Illustration
botanical illustration of tropical flowers and leaves, detailed watercolor technique, scientific illustration style, crisp white paper background, precise botanical accuracy, hand-painted quality, natural history museum plate aesthetic –ar 2:3 –v 7 –s 350
Why it works: “Natural history museum plate aesthetic” is a precise enough reference that the model understands the white background, the isolated specimen approach, and the hand-painted-but-precise quality. “Scientific illustration style” reinforces accuracy over artistic interpretation.
Storybook Children’s Illustration
charming children’s book illustration, a small fox wearing a striped scarf exploring a snowy village at dusk, warm light glowing from cottage windows, footprints in the fresh snow, soft pastel color palette, whimsical and cozy mood, flat gouache illustration style with visible paper texture
Why it works: “Visible paper texture” is the detail that separates a convincing gouache illustration from a digital approximation. “Flat” is the critical style instruction — it prevents the model from adding dimensional shading that would make it look like a 3D render instead of a storybook page.
Line Art Character Sheet
clean line art character design sheet, a young female adventurer with a long coat, leather satchel, and goggles pushed up on her head, three views (front, side, back) on white background, professional animation concept art style, clean consistent lines, no color, no shading –ar 16:9 –v 7 –s 200 –no color, shading, background
Why it works: “Three views” instruction rarely produces perfect results but pushes the model toward turnaround-style reference sheets. The
--no color, shading negative prompt is critical — without it, the model almost always adds color to line art prompts.
Category 06
Abstract & Surreal Prompts
Abstract prompts need physical anchors. “Abstract art” alone produces noise. Specifying a real movement (Dalí, Kandinsky, op-art), a color palette, and a physical object or texture gives the model enough structure to produce something intentional rather than random.
Dalí-Inspired Surreal Desert
surreal desert landscape beneath a sky split vertically — deep starry night on one side, blazing noon sun on the other, a melting grand piano with dripping keys, a massive teacup filled with a tiny ocean, a grandfather clock whose face has rolled away, golden sand, Dalí-inspired surrealism, oil painting technique, warm gold and deep blue palette –ar 16:9 –v 7 –s 500
Why it works: Each unusual object is concrete and specific — not “surreal objects” but a piano, a teacup, a clock with specific modifications. The split sky is a simple visual concept the model handles well. Physical specificity is what separates good surreal prompts from gibberish.
Fractal Space Abstract
infinite fractal landscape, deep space black background, electric blue recursive geometric patterns, warm gold energy threads running through the fractal edges, purple and magenta nebula glow in the depth, crystalline white at finest detail level, first-person perspective floating between mirrored fractals, 1:1 square, abstract mathematical art –ar 1:1 –v 7 –s 700
Why it works: “First-person perspective floating between” is an unusual and effective framing for fractal art — it creates depth and immersion instead of the flat overhead fractal view. The color palette (blue, gold, magenta) prevents the default rainbow gradient that AI applies to fractals without instruction.
Fluid Color Pour Abstract
abstract fluid pour painting, turquoise and copper liquid marble patterns, tiny gold leaf fragments floating on the surface, deep cell formation, silicone oil technique, macro photography of paint surface, soft ambient studio lighting, ultra-detailed, 1:1 –ar 1:1 –v 7 –s 300 –style raw
Why it works: Specifying an actual pour painting technique (“silicone oil technique” creates the distinctive cell formations in acrylic pours) gives the model a real physical process to reference. “Macro photography” frames it as a close-up surface study rather than a full painting.
Geometric Op-Art
op-art geometric illusion, concentric circles creating a pulsing tunnel effect, pure black and white, crisp clean edges, Bridget Riley style, optical illusion of motion and depth, print quality vector aesthetic, high contrast, centered composition, 1:1 –ar 1:1 –v 7 –s 100 –style raw
Why it works: Naming Bridget Riley gives the model an extremely precise art historical reference — black and white geometric op-art that creates perceptual movement.
--style raw with low stylization (--s 100) prevents Midjourney from adding color or texture the style doesn’t want.
Textured Abstract Expressionism
large-scale abstract expressionist painting, heavy palette knife textures, burnt sienna and cadmium yellow against deep prussian blue, gestural slashing brushstrokes, impasto technique with visible ridges and peaks, emotional and raw, museum canvas quality, side lighting revealing paint texture –ar 4:3 –v 7 –s 400
Why it works: “Side lighting revealing paint texture” is a rarely used but highly effective instruction — it tells the model to render the physical peaks and ridges of impasto paint catching directional light, which gives the image a convincingly tactile quality.
Ink and Water Photography
macro photograph of black ink dissolving in clear water, captured at the exact moment of dispersion, wispy tendrils and cloudy plumes, pure white background, perfect symmetry, slow-motion suspended quality, fine art photography, high contrast, sharp focus on ink edges fading to soft blur –ar 4:5 –v 7 –style raw
Why it works: “Captured at the exact moment of dispersion” is a time-anchoring phrase that the model interprets as capturing peak turbulence — the most visually interesting frame. Without it, ink-in-water prompts produce either fully dispersed or too-solid results.
🔧 Universal Modifiers — Add These to Any Prompt
–ar 16:9
Widescreen. Landscapes, establishing shots
–ar 9:16
Portrait / mobile. Character art, phone wallpapers
–ar 1:1
Square. Social media posts, icons, abstract art
–style raw
Reduces Midjourney’s own artistic interpretation. Better for photorealism
–niji 6
Switches to Midjourney’s dedicated anime model. All anime prompts
–s 100–700
Stylize. Low = literal, high = artistic interpretation. Start at 200–300
–no X, Y, Z
Exclude elements. Watermark, blurry, text, extra limbs, flat lighting
–v 7
Current Midjourney model. Default as of mid-2025. Always include
| Swap This In | Effect on Output | Best Used In |
|---|---|---|
| Kodak Portra 400 | Warm skin tones, lifted shadows, analogue film character | Portrait, lifestyle photography |
| Fuji Velvia 50 | Punchy saturated colors, especially greens and blues | Landscape, nature photography |
| Ilford HP5 (black and white) | High-contrast monochrome, classic photojournalism feel | Street photography, portraiture |
| matte painting style | Hand-crafted film backdrop quality — soft, deliberate, epic | Fantasy environments, sci-fi worlds |
| Blade Runner aesthetic | Rain, neon, dark, morally complex urban future | Cyberpunk scenes, dystopian settings |
| Wes Anderson palette | Symmetrical composition, pastel color blocking, flat irony | Architecture, interior scenes, character portraits |
| NASA concept art style | Scientifically grounded, optimistic, technically detailed | Space scenes, alien worlds, retro-futurism |
| Bridget Riley style | Black and white geometric op-art, perceptual movement | Abstract graphic art |
Film stock names, director names, and art movement references are among the most efficient tokens in any prompt — they carry years of visual vocabulary in two or three words.
Sources and prompt provenance: Prompts in this library are synthesized from and adapted across multiple tested collections: ZSky AI Prompt Library (March 2026), Seedance 50 Image Prompts (February 2026), AI Tool Discovery Midjourney Guide (March 2026), and Aiarty Fantasy Prompts. Parameter information verified against official Midjourney documentation. Midjourney V7 became the default model in mid-2025. Pricing verified April 2026. No affiliate relationships with any tools referenced.
From the BestPrompt.Art Community
These prompts are tested and copy-paste ready. The forum threads below show what happens when the community stress-tests similar patterns — and where the real experimentation happens:
Prompt Swap: Share a Prompt and See How Others Interpret It The “Swap Table” in this post — Kodak Portra 400 vs. Fuji Velvia 50 vs. Ilford HP5 — is exactly what this thread does live. Post any prompt from this library, watch five people run it with their own modifiers, and see how film stock references, aspect ratios, and stylize values change the output. Fastest way to learn what each token actually does.
How Do You Describe Your Art Style in a Prompt? The modifier vocabulary in this post — “matte painting style,” “NASA concept art style,” “Bridget Riley style” — is the starting point, not the endpoint. This thread documents which style references survive across model updates and which ones produce inconsistent results. Critical if you’re building a reusable prompt library.
Weirdest AI Art Results: Share Your Funniest Fails! Every prompt in this library has a failure mode. The “Enchanted Ancient Forest” prompt with --s 400 can oversaturate. The “Cyborg Character Portrait” can produce anatomically impossible neck joints. This thread catalogs the specific failure patterns for prompts in the same categories — and the negative prompt additions that fix them.
Best Prompts for Surreal and Dreamlike Art The Abstract & Surreal section in this post leans on physical anchors (melting piano, teacup ocean) to prevent noise. This thread pushes further — testing what happens when you remove physical anchors entirely, or combine incompatible ones, or let the model fill gaps with its own associations. The boundary between controlled surrealism and visual gibberish is mapped here.
Top Tools and Resources for AI Artists If you’re running these prompts at volume, the tooling layer matters. This community-curated list includes batch generation tools, seed-locking utilities, and comparison frameworks that let you run the same prompt across Midjourney, DALL-E, and Stable Diffusion without manual copy-paste. Useful for the iterative testing these prompts are designed for.
