Turns the image into graphite lines and shading on paper, often with loose, unfinished edges.
When to use it
For concept ideas, storyboards and artistic drafts.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/sketch' as the word 'sketch', so '/sketch' and 'pencil sketch' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Also asking for 'vibrant colors'.
Expecting clean vector lines; use /lineart for that.
Example
A cat curled up on a windowsill, pencil sketch on paper, graphite lines, hand-drawn hatching and shading, loose unfinished edges, no color
/watercolor
Also written: watercolor, watercolour, watercolor painting
A painting made with transparent, flowing water-based paint.
What the AI does
Uses soft washes, color bleeding at the edges, visible paper texture and light, airy tones.
When to use it
For gentle illustrations, cards, children's books and soft backgrounds.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/watercolor' as the word 'watercolor', so '/watercolor' and 'watercolor painting' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Adding 'photorealistic', which fights the painted look.
Expecting sharp, precise detail.
Example
A small fishing village at sunrise, watercolor painting, soft translucent washes, colors bleeding at the edges, visible paper texture, light airy palette
/comic
Also written: comic, comic book, comic style
A comic-book illustration look.
What the AI does
Adds bold ink outlines, flat bright colors, halftone dots, dramatic angles and sometimes panels.
When to use it
For posters, characters, story panels and fun social posts.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/comic' as the word 'comic', so '/comic' and 'comic book style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Expecting speech-bubble text to be spelled right; add text later in an editor.
Naming real comic characters or publishers.
Example
A superhero cat leaping between rooftops, comic book illustration, bold black ink outlines, flat bright colors, halftone dot shading, dynamic dramatic angle
/anime
Also written: anime, anime style, manga style
The Japanese animation look.
What the AI does
Draws clean outlines, cel-shaded flat colors, expressive large eyes and simple, painted backgrounds.
When to use it
For characters, fan-art style scenes, avatars and stylized posters.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/anime' as the word 'anime', so '/anime' and 'anime style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Naming a specific studio or living artist; describe the look instead.
Mixing it with 'photorealistic'.
Example
A girl with a red umbrella waiting at a rainy bus stop, Japanese anime style, clean line art, cel shading, vibrant flat colors, expressive large eyes, painted background
/sticker
Also written: sticker, die-cut sticker, sticker style
A cute or bold sticker design with a white border.
What the AI does
Makes one simple subject with bold outlines, flat colors and a thick white die-cut border on a plain background.
When to use it
For stickers, emojis, chat packs and merch ideas.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/sticker' as the word 'sticker', so '/sticker' and 'sticker design' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Expecting a transparent background; you get a plain one to remove later.
Putting several subjects in one sticker.
Example
A happy avocado holding a coffee cup, die-cut sticker design, single subject, bold outline, flat vibrant colors, thick white border, plain background
/oilpainting
Also written: oilpainting, oil painting, classical painting
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/oilpainting' as the word 'oilpainting', so '/oilpainting' and 'oil painting' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Naming a living painter's style.
Asking for tiny sharp details that brushwork hides.
Example
A sailing ship on a stormy sea, oil painting on canvas, thick visible brushstrokes, rich blended colors, canvas texture, classical lighting
/pixelart
Also written: pixelart, pixel art, 8-bit, retro game art
Retro video-game art made of visible square pixels.
What the AI does
Uses a low resolution look, a limited color palette and blocky shapes like old game sprites.
When to use it
For game ideas, retro posters and fun avatars.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/pixelart' as the word 'pixelart', so '/pixelart' and 'pixel art' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Expecting true pixel-perfect sprites you can use in a game; results only imitate the look.
Example
A knight standing in front of a castle gate, pixel art, visible square pixels, low resolution retro game look, limited color palette, crisp blocky shapes
/3drender
Also written: 3drender, 3D render, CGI, 3d style
A clean computer-generated 3D look.
What the AI does
Makes smooth surfaces, soft studio light, realistic materials and gentle shadows, like animation or product CGI.
When to use it
For product concepts, characters, icons and playful scenes.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/3drender' as the word '3drender', so '/3drender' and '3D render' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Expecting an actual 3D model file; you get a flat image.
Naming a specific animation studio.
Example
A cute robot watering a plant, polished 3D render, smooth surfaces, soft studio lighting, realistic materials, gentle shadows, clean background
/claymation
Also written: claymation, clay style, plasticine
A look made of modeling clay, like stop-motion animation.
What the AI does
Gives soft rounded shapes, visible fingerprints and tool marks, and warm miniature-set lighting.
When to use it
For playful characters, kids' content and friendly ads.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/claymation' as the word 'claymation', so '/claymation' and 'claymation style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Expecting a moving animation from an image prompt; use video for motion.
Example
A family of snails having a picnic, claymation style, handmade modeling clay figures, soft rounded shapes, visible fingerprint texture, miniature set, warm lighting
/lineart
Also written: lineart, line art, outline drawing
A drawing made only of clean lines, usually black on white.
What the AI does
Removes shading and color and keeps clear outlines.
When to use it
For coloring pages, tattoos, logos and minimal art.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/lineart' as the word 'lineart', so '/lineart' and 'line art' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Adding color or texture words, which cancel the clean look.
Example
A hummingbird drinking from a flower, clean line art, black outlines on white background, no shading, no color, even line weight
/vector
Also written: vector, flat design, flat illustration
A flat, clean graphic illustration made of simple shapes.
What the AI does
Uses solid flat colors, simple geometric shapes, no texture and crisp edges.
When to use it
For icons, web graphics, explainers and presentations.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/vector' as the word 'vector', so '/vector' and 'flat vector illustration' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Expecting an editable SVG file; you get a picture that looks like vector art.
Example
A person working on a laptop with charts floating around, flat vector illustration, simple geometric shapes, solid flat colors, crisp edges, no texture, clean background
/isometric
Also written: isometric, isometric view, isometric 3d
A 3D-looking scene drawn at a fixed diagonal angle without perspective.
What the AI does
Shows objects from above at an angle, like a tiny diorama, with even lines and no vanishing point.
When to use it
For tech illustrations, maps, rooms and game-like scenes.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/isometric' as the word 'isometric', so '/isometric' and 'isometric illustration' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Mixing it with 'wide-angle lens', which adds perspective.
Example
A tiny city block with a café and a park, isometric illustration, 30-degree diagonal top view, no perspective distortion, miniature diorama, clean shapes
/lowpoly
Also written: lowpoly, low poly, polygon art
A 3D style made of few, flat triangles.
What the AI does
Builds shapes from visible geometric facets with flat shading and simple colors.
When to use it
For modern minimal graphics, games and backgrounds.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/lowpoly' as the word 'lowpoly', so '/lowpoly' and 'low poly' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Asking for fine details, which the style removes on purpose.
Example
A fox sitting on a hill at sunset, low poly 3D style, visible triangular facets, flat shading, simple color palette, geometric shapes
/cartoon
Also written: cartoon, cartoon style, toon
A fun, simplified drawing style with exaggerated features.
What the AI does
Uses bold outlines, bright flat colors, big expressions and simple shapes.
When to use it
For kids' content, mascots, avatars and light-hearted posts.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/cartoon' as the word 'cartoon', so '/cartoon' and 'cartoon style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Naming a famous cartoon character or studio.
Example
A grumpy cloud raining on a happy umbrella, cartoon style, bold outlines, bright flat colors, exaggerated friendly expressions, simple shapes
/chibi
Also written: chibi, chibi style, super deformed
A super-cute anime style with tiny bodies and big heads.
What the AI does
Draws characters with oversized heads, tiny bodies, huge eyes and simple, adorable poses.
When to use it
For cute avatars, stickers and mascots.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/chibi' as the word 'chibi', so '/chibi' and 'chibi style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Also written: cyberpunk, cyberpunk style, neon future
A gritty, high-tech future city look.
What the AI does
Adds neon signs, rain, dark streets, holograms, cables and futuristic gadgets in pink and blue light.
When to use it
For sci-fi scenes, game art and edgy posters.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/cyberpunk' as the word 'cyberpunk', so '/cyberpunk' and 'cyberpunk style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Expecting readable signs; neon text is often garbled.
Example
A street food stall in a futuristic megacity, cyberpunk style, neon pink and blue lights, rainy dark city streets, holograms, futuristic tech, gritty atmosphere
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/fantasy' as the word 'fantasy', so '/fantasy' and 'fantasy art' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Naming characters from real books or films.
Example
A dragon resting on a mountain above a misty valley, epic fantasy art, magical glowing elements, dramatic painterly lighting, grand landscape, mythical atmosphere
/steampunk
Also written: steampunk, steampunk style
A Victorian-era style mixed with brass machines and steam power.
What the AI does
Adds gears, brass and copper pipes, leather, goggles, clockwork and steam.
When to use it
For inventive characters, machines and story art.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/steampunk' as the word 'steampunk', so '/steampunk' and 'steampunk style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Mixing it with cyberpunk neon, which pulls in another era.
Example
An inventor's airship workshop, steampunk style, Victorian era, brass and copper gears, clockwork machinery, leather and goggles, drifting steam
/papercut
Also written: papercut, paper cut, paper craft, layered paper
Art that looks made of layered cut paper.
What the AI does
Builds the scene from stacked paper layers with soft shadows between them and clean cut edges.
When to use it
For cards, kids' content and decorative backgrounds.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/papercut' as the word 'papercut', so '/papercut' and 'paper cut art' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Asking for photorealism, which removes the craft look.
Example
An underwater scene with fish and seaweed, layered paper cut art, stacked colored paper layers, clean cut edges, soft shadows between layers, handmade craft look
/ink
Also written: ink, ink drawing, pen and ink
A drawing made with pen and ink.
What the AI does
Uses black ink lines, cross-hatching and bold contrast, sometimes with ink washes.
When to use it
For book illustrations, tattoos and dramatic black-and-white art.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/ink' as the word 'ink', so '/ink' and 'ink illustration' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Adding lots of color words.
Example
An old tree on a hill under the moon, pen and ink illustration, crisp black ink lines, cross-hatching, bold contrast, occasional ink wash
/popart
Also written: popart, pop art
A bold, colorful style from 1960s-era advertising and comics.
What the AI does
Uses bright flat colors, thick outlines, halftone dots and repeated or high-contrast images.
When to use it
For eye-catching posters and fun portraits.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/popart' as the word 'popart', so '/popart' and 'pop art' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Naming famous artists or their exact works.
Example
A banana on a bright yellow and pink background, pop art style, bold saturated flat colors, thick black outlines, halftone dots, high contrast graphic look
/surreal
Also written: surreal, surrealism, dreamlike
Dream-like scenes where impossible things happen.
What the AI does
Mixes realistic detail with impossible situations, odd scale and strange combinations.
When to use it
For creative posters, album covers and thought-provoking art.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/surreal' as the word 'surreal', so '/surreal' and 'surreal art' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Adding so many strange elements that nothing reads clearly.
Example
A whale floating above a quiet desert road, surreal dreamlike scene, impossible elements, unusual scale, realistic detail, strange but calm atmosphere
/conceptart
Also written: conceptart, concept art
A painted design image used to plan films and games.
What the AI does
Gives painterly, atmospheric scenes focused on mood, scale and design ideas rather than fine finish.
When to use it
For world-building, game ideas and pitches.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/conceptart' as the word 'conceptart', so '/conceptart' and 'concept art' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Expecting a final, polished product image.
Example
An abandoned space station overgrown with plants, concept art, painterly digital illustration, atmospheric lighting, strong sense of scale, cinematic composition
/storybook
Also written: storybook, children's book, picture book
A warm illustration style like a children's picture book.
What the AI does
Uses soft shapes, friendly characters, warm colors and gentle textures.
When to use it
For kids' stories, bedtime content and friendly educational posts.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/storybook' as the word 'storybook', so '/storybook' and 'children's book illustration' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Naming a famous children's illustrator.
Example
A little bear learning to fly a kite, children's picture book illustration, soft rounded shapes, friendly characters, warm gentle colors, light paper texture
/minimal
Also written: minimal, minimalist, minimalism, simple
A very simple image with few elements and lots of empty space.
What the AI does
Removes clutter, keeps one subject, uses a small color palette and plenty of negative space.
When to use it
For clean posters, wallpapers, logos and modern brand images.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/minimal' as the word 'minimal', so '/minimal' and 'minimalist style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Adding many details, which cancels it.
Forgetting to say which colors to keep.
Example
A single red chair in a white room, minimalist composition, single subject, generous empty space, limited color palette, clean simple shapes
/vintage
Also written: vintage, retro, old-fashioned
An old-fashioned, retro look from past decades.
What the AI does
Adds faded colors, worn textures, old printing styles and period details.
When to use it
For nostalgic posters, branding and themed designs.
Where it works
Not an official command in ChatGPT, Gemini or Midjourney: the model reads '/vintage' as the word 'vintage', so '/vintage' and 'vintage retro style' give the same result. Works as plain words in GPT image, Gemini/Imagen, Midjourney, Stable Diffusion and AI Compass's generator.
Common mistakes
Not naming a decade, so the era is vague; say e.g. 'a 1950s feel'.
Example
A travel poster of a seaside town, vintage retro style, faded muted colors, worn paper texture, old print look, period details
Every guide and dictionary card is also in the AI Compass app, with saved items, search and “Try it” buttons that open chat, image or video with the example filled in. Reading is free; creating uses credits, and the price is shown before you start.