How to Write AI Image Prompts That Work: A Simple Formula With 25 Examples

The short version

A useful image prompt gives the model concrete choices instead of asking it to guess. Describe the subject, action or setting, style or medium, composition, lighting, and mood or color. You do not need every part every time; add the details that matter to this image.

Keep the wording specific, change one thing between attempts, and check the image tool's current terms before using an output commercially. These examples are templates, not guaranteed results: different generators interpret prompts differently.

Why vague prompts miss

A prompt such as “a dog in a park” leaves the breed, pose, season, camera angle, lighting, and visual style open. A more directed prompt might be: “A golden retriever puppy sitting on autumn leaves in a quiet park, soft morning light, shallow depth of field, warm color palette, candid photograph.”

The extra details give the model a clearer target. They do not guarantee exact control, but they make it easier to judge what to adjust after each generation.

A six-part prompt formula

1. Subject: Name what should appear. Prefer “an elderly fisherman with a weathered face” to “a man.”
2. Action or setting: Say what the subject is doing and where, such as “mending a net on a wooden pier at sunrise.”
3. Style or medium: Choose a visual form such as studio photograph, watercolor illustration, flat vector, pencil sketch, or 3D render.
4. Composition and camera: Describe framing: close-up, wide shot, overhead view, eye level, centered, or shallow depth of field.
5. Lighting: Try soft window light, golden hour, dramatic side lighting, neon glow, overcast daylight, or studio softbox.
6. Mood and color: Set the feeling and palette, for example calm with muted earth tones, energetic with bright orange, or nostalgic with warm pastels.

Put together: “[Subject] + [action or setting] + [style or medium] + [composition] + [lighting] + [mood and color].” Add only the parts that help the image you have in mind.

25 prompts to copy and adapt

Product and e-commerce

1. A matte black water bottle standing on a light gray surface, studio product photograph, centered, soft diffused lighting, clean minimal look, subtle shadow.
2. A pair of white running shoes floating above a pastel blue background, commercial photography, low angle, bright even lighting, playful mood.
3. A glass bottle of face serum on a marble shelf beside green leaves, lifestyle photograph, shallow depth of field, soft window light, fresh and calm.
4. A ceramic mug of hot coffee on a wooden table by a rainy window, cozy close-up photograph, warm lamp light, comfortable mood.
5. A minimalist backpack on a concrete ledge in an urban setting, editorial photograph, wide shot, overcast light, muted tones.

People and portraits

1. A smiling baker in a flour-dusted apron holding a fresh loaf, environmental portrait, medium shot, warm morning light from a bakery window.
2. A young musician tuning a guitar backstage, candid photograph, side view, dramatic stage lighting, moody blue and amber tones.
3. An elderly woman laughing in a sunlit garden, natural portrait, 85mm lens, shallow depth of field, soft golden-hour light.
4. A confident business owner standing in front of a small shop, professional headshot style, neutral background, soft studio lighting.
5. A child flying a kite on a windy beach, joyful lifestyle photograph, wide shot, bright afternoon light, saturated colors.

Landscapes and places

1. A misty mountain valley at sunrise with a winding river, landscape photograph, wide-angle view, soft pink and gold light, serene mood.
2. A narrow cobblestone street in an old European town after rain, cinematic photograph, reflections on the stones, warm streetlights, evening blue hour.
3. A desert dune ridge under a starry night sky, long-exposure look, deep blue and silver tones, quiet and vast.
4. A tropical beach with turquoise water and palm trees, aerial drone view, bright midday sun, vivid colors.
5. A cozy mountain cabin in snow with smoke rising from the chimney, digital painting, warm window glow, cool blue shadows.

Illustration, art, and characters

1. A friendly robot gardener watering flowers, children's book illustration, soft watercolor texture, pastel palette, cheerful mood.
2. A fox in a scarf reading a book under a lantern, storybook illustration, warm autumn colors, gentle lighting.
3. A samurai standing on a cliff at dawn, ink and watercolor art style, dramatic composition, muted reds and grays.
4. An isometric illustration of a tiny coworking office with plants and desks, clean vector style, soft gradients, bright pastel colors.
5. A fantasy castle on a floating island among clouds, detailed digital art, epic wide shot, golden sunset light.

Design, food, and interiors

1. A minimalist logo of a mountain and sun inside a circle, flat vector, two colors, plain white background.
2. A flat-lay of a healthy breakfast bowl with berries, nuts, and yogurt on a wooden table, overhead food photograph, bright natural light, fresh colors.
3. A modern living room with a beige sofa, large windows, and indoor plants, interior design photograph, wide shot, soft daylight, Scandinavian style.
4. A social media banner for a fitness studio, bold geometric shapes, energetic orange and dark gray palette, clean modern graphic design, space for text on the left.
5. A vintage travel poster of a seaside village, retro illustration style, limited color palette, bold shapes, sunny and nostalgic mood.

Choose a generator for the task

Different image tools offer different models, editing controls, reference-image workflows, and text rendering. Midjourney, Adobe Firefly, Leonardo AI, Ideogram, Meta AI, and ChatGPT are examples to compare; check each product's current capabilities, account requirements, and plan terms before choosing.

For posters or logos with words, generate a simple layout and add exact copy afterward in a design editor if the model distorts the text. For repeatable characters or product shots, look for reference-image or editing controls that fit your workflow.

Common prompt mistakes

Vague adjectives such as “beautiful” or “amazing” give little visual direction. Replace them with concrete choices about subject, medium, light, and framing.

Combining styles that conflict, such as “photorealistic cartoon watercolor,” can make the target unclear. Pick one main style and add compatible details.

Crowding a prompt with many unrelated subjects can produce a muddled image. Start with one focal subject, then add supporting details.

Exact text may come out misspelled or distorted. Keep generated text short or add it later in an editor. Decide on a square, vertical, or widescreen aspect ratio before generating.

Changing several details at once makes it hard to learn what helped. Adjust one element per attempt and save prompts that work.

Improve a prompt step by step

Start with the subject and style. Generate once, then identify the biggest issue: framing, lighting, palette, or a missing detail. Change that one element, generate again, and keep whatever improved the result.

If you are starting from a rough idea, ask an assistant to turn it into a structured prompt using the six parts. Review the wording yourself, then adapt it to the image tool you are using.

A note on usage rights

An AI-generated image is not automatically cleared for every use. Terms on commercial use, ownership, attribution, and permitted inputs differ by tool and plan, and can change. Check the current terms for your account before using an image in advertising, client work, or a product.

Avoid prompts that copy a living artist's style, a brand's logo, or a real person's likeness without permission. This is general information, not legal advice.

Frequently asked questions

How long should an AI image prompt be? There is no fixed length. One or two clear sentences covering the details that matter are usually a good start.

Do I need negative prompts? Only when your tool supports them. Some generators have a separate negative prompt field; others ignore it.

Should I use the same prompt in every image tool? The basic formula transfers, but some tools respond better to natural sentences and others to short phrases. Test and adjust.

Why do my images look generic? The prompt may leave too many choices open. Add a specific medium, lighting, or composition detail.

Can I use AI-generated images commercially? It depends on the tool and plan. Check the provider's current terms.

What is the easiest way to start? Copy one of the examples, replace the subject with your own, and change one detail at a time.

Keep exploring

There is no magic phrase for a good prompt. Describe what is in the picture, how it should look, and how it is framed and lit. Start with one example, adjust one element at a time, and build a reusable structure that suits your work.

Explore AI image tools, AI assistants, and design tools in the Toolshub directory.