Prompt library · 15 min read ·

AI Image Prompts: 30 Copy-Paste Examples by Category

A good AI image prompt is one or two plain sentences naming the subject and what it is doing, the place, the light, the framing and one style, for example "A fruit seller arranging oranges at a morning market, soft overcast light, eye-level close-up, photo." Put words that must appear in the image in double quotes, set the shape in the tool's size control, and change one detail per try.

These 30 AI image prompts are written for Image Generator in Toybox AI, and every block opens that tool with the prompt already typed in. Paste them into ChatGPT, Gemini or Midjourney too, as starting points: each model reads a prompt its own way, so expect different results. None of them needs a photo. The list runs through six categories, from realistic photos and watercolor art to anime panels, original characters, food and interiors, and designs with words. Under each prompt, a "Change:" line names the one or two details to swap for your own idea.

How a good AI image prompt is built

An AI image prompt is the text description an image generator turns into a picture. The ones that work describe a single picture in full sentences, not a pile of tags: Google's prompting guide for its Gemini image models says a list of keywords won't do and asks you to describe the scene as a narrative. The 30 AI image prompt examples below all use the same six parts, in this order:

  1. Subject and action. One figure or object doing one thing: "a fruit seller arranging oranges", not "woman, market".
  2. Setting. The place and what sits around the subject: "wooden crates and handwritten price cards".
  3. Light. One source and the time of day: overcast daylight, a single lamp, blue hour.
  4. Framing. Distance and angle, plus a lens for a photo look: "eye-level medium close-up, 85mm lens".
  5. Style or medium. One only: photo, watercolor, linocut, pixel art, 3D render.
  6. Exact words. Any text in the picture, inside double quotes.

Leave the shape out of the words and pick it in the tool's size setting instead. To write your own, fill in the brackets of this template:

A [style or medium] of [subject] [doing what] in [setting]. [Light source and time of day]. [Framing and camera angle]. [Colors or mood]. The only text in the image reads "[exact words]".

Photorealistic photo prompts

Each photo prompt names a lens or a camera angle and one light source, the way a photographer would describe the shot. Before you post one, check the fingers, the lettering on signs and the direction of every shadow, using the checklist in how to make AI images look real.

Morning market portrait

A close portrait of an invented person at work, with real skin texture.

A photorealistic portrait of a fruit seller in her sixties arranging oranges at an outdoor morning market, laughing at something off camera. Soft overcast daylight. Wooden crates and handwritten price cards blurred behind her. Eye-level medium close-up with an 85mm lens and a softly blurred background. Visible wrinkles and real skin texture, no retouching.

Change: the job and the produce, such as a fishmonger with ice trays.

Rainy street at night

A wide city scene where the wet ground does most of the work.

A photorealistic wide shot of a narrow city street after rain at night. A lone cyclist in a yellow raincoat rides away from the camera over wet cobblestones that reflect neon shop signs. Light mist, warm light from shop windows on both sides. Low camera angle, 24mm lens.

Change: the raincoat color and the vehicle, such as a delivery scooter.

Dog shaking off water

An action shot that needs a fast shutter to look real.

A candid photo of a scruffy terrier mix shaking water off its coat on a pebble beach, droplets frozen in the air around it. Late afternoon sun from the side. Telephoto lens, fast shutter speed, gray sea and sky softly blurred behind the dog.

Change: the breed and the setting, such as a backyard sprinkler.

Lake cabin at blue hour

A quiet landscape with one warm light against cool tones.

A photorealistic landscape of a small wooden cabin on a lake shore at blue hour. One window glows orange, thin fog drifts over the still water, and pine-covered hills rise behind. The cabin is mirrored in the lake. Wide-angle lens, steady tripod shot with a long exposure look.

Change: the season, with snow on the roof or red maples on the hills.

Hands shaping bread

A tight detail shot, useful for bakeries, crafts and blog headers.

A close-up photo of a baker's flour-dusted hands shaping a round loaf of dough on a worn wooden counter. Early morning window light from the left, a cup of coffee steaming in the soft background. Macro detail on the dough and flour. Muted warm colors.

Change: the craft, such as clay on a potter's wheel or yarn on knitting needles.

Illustration and AI art prompts

These AI art prompts name one medium and the marks it leaves: soft washes, carved lines, gouache texture, square pixels. OpenAI's prompting guide suggests naming the medium and adding a texture cue, such as textured brushstrokes, when the look matters.

Watercolor heron

A loose painting with the paper left showing.

A loose watercolor painting of a blue heron standing in shallow water among reeds. Soft washes of teal and gray, visible paper texture, white paper left bare around the edges. Very little detail in the background.

Change: the bird and the palette, such as a red fox in autumn orange.

Two-color linocut lighthouse

A print look with bold carved lines and uneven ink.

A two-color linocut print of a lighthouse on a rocky point with waves breaking below. Bold carved lines, deep navy ink on cream paper, slightly uneven ink coverage and rough edges like a hand-pulled print.

Change: the landmark and the ink color, such as a barn in rust red.

Picture book hedgehog

A soft children's book scene with one glowing light source.

A warm children's picture book illustration of a small hedgehog carrying a paper lantern through a snowy forest at night. Rounded shapes, gouache texture, a soft glow around the lantern, gentle blues and warm yellows.

Change: the animal and what it carries, such as a mouse with a basket of berries.

Isometric bakery cutaway

A tidy miniature building, good for presentations and websites.

An isometric 3D illustration of a tiny corner bakery, cut away to show the oven, the counter, shelves of bread and a baker in an apron. Pastel colors, soft shadows, clean edges, plain light gray background.

Change: the building, such as a bike repair shop or a flower stall.

Pixel art campfire

A retro game scene with a strict palette.

A pixel art scene of a campfire beside a small tent under a starry sky, in a 32-bit game style. Limited palette of dark blue, purple and orange, crisp square pixels, no blur or smoothing.

Change: the scene, such as a fishing pier at sunset.

Anime and cartoon style prompts

These prompts name the technique instead of a studio or a show: clean line art, cel shading, screentone, speed lines, visible thumbprints in clay. Describing the technique, rather than naming a studio, show or character, keeps each prompt about an original drawing. For the soft, hand-painted Ghibli-inspired look, the Ghibli-style prompts go further.

Anime train platform at dusk

A quiet slice-of-life scene with a big evening sky.

An anime-style illustration of a young woman in a yellow raincoat waiting on a small countryside train platform at dusk, a paper bag of oranges at her feet. Power lines and rice fields stretch behind her under a huge pink and orange sky. Clean line art, cel shading, soft painted clouds.

Change: the season and the weather, such as falling snow with the platform lamps lit.

Manga panel with one speech bubble

A black and white comic panel with a short line of dialogue.

A black and white manga-style comic panel of an original cat detective in a long trench coat peeking around a brick street corner at night. Speed lines behind him, screentone shading on the wall. One speech bubble reads "Too quiet" in bold hand lettering.

Change: the detective's animal and the line, such as an owl saying "Found it".

Four-panel cactus comic

A wordless comic strip, easy to share in a group chat.

A four-panel comic strip in a clean, flat cartoon style. Panel one: a small potted cactus smiling on a sunny windowsill. Panel two: a gray cloud covers the sun. Panel three: the cactus looks shocked. Panel four: the sun is back and the cactus wears tiny sunglasses. No text anywhere.

Change: the plant and the problem, such as a fern that gets watered too much.

Claymation picnic

A stop-motion look with visible handmade texture.

A claymation-style scene of two original round clay monsters, one blue and one orange, sharing sandwiches on a checkered picnic blanket. Visible thumbprints in the clay, felt grass, a cardboard sun on a wire. Soft studio light and shallow depth of field, like a photo of a stop-motion set.

Change: the activity, such as the monsters flying a paper kite.

Anime ramen bowl

A close food shot with the glossy, exaggerated style of animated food scenes.

An anime-style close-up of a steaming bowl of ramen on a wooden counter: a soft-boiled egg cut in half, two slices of pork, green onions and a sheet of nori. Glossy broth, curls of steam, warm light from a paper lantern above, rich saturated colors.

Change: the dish, such as a stack of fluffy pancakes with melting butter.

Character and fantasy prompts

Every character here is described by its own features (a red scarf, dented copper, braided silver hair) rather than by a film, game or show it resembles. To help one character look the same across several images, make the character sheet first, download it, add it as a reference photo in each later run, and compare every result with the sheet.

Elf archer in a birch forest

A painted fantasy portrait with detail on the costume.

A painted fantasy portrait of a young elf archer with braided silver hair and a moss-green hooded cloak, resting a longbow on her shoulder in a misty birch forest. Soft light from behind that outlines her hair, fine stitching on the leather armor, digital painting style.

Change: the class and the setting, such as a desert healer in a sandstone temple.

Copper robot gardener

A friendly character with clay-like materials.

A friendly, rounded robot made of dented copper, watering potted tomato plants on a sunny apartment balcony with city rooftops behind it. Soft 3D render with clay-like materials, gentle morning light, cheerful colors.

Change: the chore, such as the robot hanging laundry or walking a small dog.

Research station on a red plateau

A wide sci-fi landscape with a clear sense of scale.

A wide cinematic shot of a small research station of white modular cabins on a red rocky plateau, with a huge ringed planet low in a pale pink sky. A six-wheeled rover leaves tracks toward the cabins and dust hangs in the air. Muted warm palette, tiny human figures in orange suits for scale.

Change: the world, such as an ice shelf under a green aurora.

Fox courier character sheet

A reference sheet that fixes one character's look from three sides.

A character reference sheet of an original cartoon fox courier wearing a red scarf and a brown messenger bag, shown from the front, the side and the back on a plain white background. Same proportions and colors in all three views. Clean line art with flat colors.

Change: the animal and the outfit, such as an owl librarian in a green cardigan.

Sleepy library dragon

A cozy scene built around a small creature.

A small sleepy dragon curled around a stack of old books on a library windowsill, rain on the glass behind it. Warm lamp light, detailed green scales, a storybook oil painting style.

Change: the creature's color and perch, such as a blue dragon on a teapot.

Food, object and interior prompts

These name the material of every surface, such as linen, cast iron, pale oak or hand-thrown ceramic, instead of a plain "table" or "vase". For listings and shop photos of your own products, the product photography prompts go further.

Weekend brunch from above

An overhead table with a human hand for scale.

An overhead photo of a weekend brunch table: shakshuka in a cast-iron pan, a basket of torn bread, a bowl of olives and two glasses of mint tea on a creased linen cloth. Natural window light, and a hand reaching in from the lower corner.

Change: the meal, such as pancakes with berries and a coffee pot.

Vase still life

A calm object shot with long shadows.

A still life photo of a hand-thrown ceramic vase with three dried poppy stems, standing on a wooden stool against a textured plaster wall. Soft side light, long gentle shadows, muted earthy colors, the vase centered with space on both sides.

Change: the object, such as a stack of cookbooks or a brass kettle.

Small living room

An interior shot with straight vertical lines, the way interior photographers shoot rooms.

A photo of a small living room with a pale oak floor, a linen sofa, a wool rug, a tall potted plant and a reading lamp. Late afternoon sun through sheer curtains. Wide-angle interior photography with straight vertical lines.

Change: the style, such as dark green walls with a leather armchair.

Camping gear flat lay

A knolling layout where every item sits at a right angle.

A knolling flat lay of camping gear arranged at right angles on a dark green canvas background: tent stakes, a folding knife, a tin mug, a paper map, a headlamp and a coil of rope. Even top-down lighting, sharp focus on every item.

Change: the kit, such as art supplies or a baking set.

Lemon layer cake

A bakery photo with a slice removed to show the inside.

A tall three-layer lemon cake with one slice removed, showing pale yellow sponge and cream between the layers, on a white cake stand. A few lemon slices and a cake server beside it. Soft daylight, shallow depth of field.

Change: the flavor, such as chocolate with raspberries.

Prompts for designs with words

Each design below puts its words in double quotes and keeps them to five words or fewer, so there is less to misspell and less to proofread. In Image Generator, open "More options" and switch on "Enhanced quality", which the tool labels "Sharper text and logos" (currently 45 extra credits). A flyer with a date, an address and prices carries far more text than these; Flyer Maker is the Toybox tool for that.

Birthday card for a child

A card front with a name and an original animal.

A birthday card design with the words "Happy 7th Birthday, Maya" in big rounded hand lettering across the top. Below it, an original cartoon giraffe wearing a party hat. Pastel confetti, lots of white space around the words.

Change: the name, the age and the animal.

Typographic poster

A minimal poster where the words are the design.

A minimalist typographic poster that reads "Slow Mornings" in a large serif typeface across the top third. Below it, a single cup of coffee on a saucer drawn in thin black lines. Cream background with plenty of empty space.

Change: the phrase and the drawing, such as "Garden Days" with a watering can.

Vegetable sticker sheet

Six characters with die-cut borders, laid out like a sticker sheet.

A sticker sheet of six original cute vegetable characters: a carrot, a tomato, a broccoli floret, a pea pod, a radish and an eggplant, each with a happy face and a thick white die-cut border. Bold outlines, flat bright colors, light gray background. The only text is "Eat Your Greens" at the bottom.

Change: the theme, such as six fruits or six houseplants.

Painted shop sign

A photo of a sign, where the words sit on a real surface.

A photo of a hand-painted wooden sign hanging outside a small bookshop. The sign reads "Open Late Fridays" in gold letters on dark green paint. Red brick wall behind it, warm evening street light, slight wear on the paint.

Change: the shop and the message, such as a bakery sign reading "Fresh Bread Daily".

Phone wallpaper with a short phrase

A tall wallpaper that leaves room for the clock and icons.

A vertical phone wallpaper of a calm ocean horizon at dawn in soft pinks and blues. The words "Breathe in" in small white letters centered in the lower third. Lots of empty sky in the upper half.

Change: the scene and the words, such as a pine forest with "Stay curious". Pick the Story 9:16 size so it fills a phone screen.

Tips for better AI image prompt results

Change one line of a prompt at a time, and trim it to suit the tool you paste it into.

  • Swap only the subject first. Keep the light, framing and style lines and replace the subject using the "Change:" line. If the new picture looks worse, the new subject caused it, not the rest of the prompt.
  • Keep the counts when you swap items. The brunch table names two glasses and the camping flat lay names six items. Replace a number with another number, never with "some", or the count in the picture drifts.
  • Trim to one sentence for Midjourney. Midjourney's Prompt Basics says short, simple prompts usually work best there. Keep the first sentence and the style line, drop the rest, and add the shape as a parameter at the end, such as --ar 4:5.
  • Put "Create an image:" in front for ChatGPT and Gemini. Gemini's help center suggests starting an image request with draw, generate or create, and OpenAI's help center says a plain request for an image in the chat is enough.
  • Change the lens when you change the distance. The close portraits use an 85mm lens, while the street and lake scenes use a 24mm or wide-angle lens. When you turn a close-up into a wide scene, swap the lens words too.
  • Keep designs with words to five words. If you add a line to the birthday card or the sign, add it as a separate quoted phrase with its own position ("at the bottom", "on the door").

For why these six parts work, how each tool wants a prompt shaped, and one idea taken through three rounds, read the guide to writing image prompts. Edits to your own photo and prompts that combine two or three photos are in the Nano Banana prompts.

Fixes for common problems

Problem Add this to your prompt
Watercolor comes out looking like digital painting "Visible paper texture, soft washes, white paper left bare at the edges"
Pixel art looks blurry "Crisp square pixels, no blur, no smoothing, limited palette"
The character changes between images Add your character sheet result as a reference photo, then "Same character as in my photo"
Sticker borders are uneven "Each sticker has the same thick white die-cut border"
Extra text you didn't ask for "No other text or lettering anywhere in the image"
A word is misspelled Cut it to a few words in double quotes, then add "spelled exactly as written"
Wrong number of objects "Exactly two cups, no more"
Head or feet cut off "Full body in frame, with space above the head and below the feet"
Two styles blended together Keep one medium, such as "watercolor on paper" only
Subject looks pasted in "Lit by the same light as the scene, with a soft shadow on the ground"

Run these prompts in Image Generator

Tap "Try this prompt" under any block to open Image Generator with the prompt filled in, pick a size, and tap "Create image". It usually takes about 30 seconds and currently costs 50 credits per image. Read every word on the designs with text before you share them, because a generator can still drop or swap a letter. On a Pro plan, a "Want to change something?" box under the result takes one fix, such as a wrong color or a stray word, so you don't start over (currently 85 credits). Plans and top-ups are on the pricing page.

Frequently asked questions

Can I use these AI image prompts for ChatGPT?

Yes. Paste a prompt into ChatGPT with "Create an image:" in front of it, or select More, then Images, before you paste. As of September 2026, ChatGPT Images is on every plan, and OpenAI lists image generation on the Free plan as limited and slower. Pick the shape with ChatGPT's aspect ratio picker instead of typing it into the prompt.

Is there an AI image prompt generator?

A chatbot can do the job. Paste the fill-in template from this page, ask it to complete the brackets for your idea, and delete any detail you never asked for. Google's Nano Banana prompting guide suggests using Gemini to help write prompts. Before you use it, check that the prompt still describes a single picture with one style.

What are good ChatGPT image prompt ideas?

Start from something you need this week: a birthday card, a shop sign, a blog header or a phone wallpaper. The designs-with-words prompts on this page cover cards, signs and wallpapers, and the brunch and cake prompts adapt to a cafe menu or a bakery post. An event flyer with a date, an address and prices has far more text, so use a flyer tool for it and proofread every line.

How many credits do these prompts use in Toybox?

Image Generator currently costs 50 credits per image, or 95 with "Enhanced quality" switched on. Image Generator Lite currently costs 35 credits for a 1K image. New accounts get 20 free credits, which does not cover an image, and you can currently buy 500 credits for $4.99 without a subscription.

Can I use the images I make from these prompts commercially?

Toybox's Terms of Service say you own the rights to the outputs you generate, subject to the Terms, and you are responsible for how you use them. The US Copyright Office concluded in 2025 that AI output gets copyright protection only where a person determined enough of its expressive elements, and that writing prompts alone isn't enough. Check copyright, trademark and platform rules first. This is general information, not legal advice.

Can I use a photo of myself with these prompts?

Yes. In Image Generator, tap "Add reference photos" and add up to 3 JPG or PNG photos, then name each photo's job in the prompt, such as "the dog from my first photo". The photo guides the AI rather than being copied, so a face can come out different. Toybox's privacy policy says your prompts, uploads and results are not used to train AI models.

Sources

Make it with AI Image Generator

Keep reading