How it works

Describe the image in a sentence or two: what it shows, what it's for, and anything it must include or avoid.
Pick your image model, a style and the shape you need, then click Write my prompts.
Paste one of the three prompts into Midjourney, ChatGPT, Stable Diffusion, Flux or Gemini. Generate a few images from each and keep the one that fits.

Examples

Sample results for the inputs shown, so you can see the format and quality before you try it.

Example 1: input

What should the image show? Header image for a LinkedIn post about remote teams missing the small talk of the office

Image model: Midjourney

Style: Editorial illustration

Shape: Landscape 16:9

Example 1: output

Prompt 1 (closest to your idea): editorial illustration of a video call grid with six remote coworkers in different home rooms, each looking a little bored, one empty tile showing an office kitchen with a coffee machine and two mugs, medium wide view, flat shapes with subtle paper grain, soft even light, muted ochre, slate blue and cream palette, quiet wistful mood, clean empty space on the right --ar 16:9 --no text, logos Prompt 2 (bolder composition): editorial illustration, a giant laptop screen seen from above like a floor plan, tiny coworkers standing alone in separate video tiles, a single speech bubble floating between two tiles, strong diagonal composition, warm orange light from one side against deep navy, bold graphic shapes, gentle humor --ar 16:9 --no text, logos Prompt 3 (simpler, fewer elements): editorial illustration of two coffee mugs on opposite ends of a long empty desk, a thin dotted line connecting them, lots of negative space, pale sand background, one accent of coral, calm minimal style --ar 16:9 --no text

Example 2: input

What should the image show? Our new matte black insulated water bottle, 750 ml, with a bamboo lid, for the product page

Image model: ChatGPT / GPT image

Style: Product shot on plain background

Shape: Square 1:1

Example 2: output

Prompt 1 (closest to your idea): A studio product photo of a matte black insulated water bottle with a light bamboo lid, standing upright in the center of a plain warm-gray paper backdrop. Soft, diffused light from the upper left with a gentle highlight along the bottle's edge and a soft shadow on the right. The matte finish shows no fingerprints or reflections. Square 1:1 image, sharp focus on the whole bottle, clean ecommerce style. Prompt 2 (bolder composition): A dramatic square product photo of a matte black insulated water bottle with a bamboo lid, placed on a slab of pale stone, shot from a low angle so the bottle looks tall. A single hard light from the side creates a crisp shadow line and shows the wood grain of the lid. Deep charcoal background fading to black. Premium, confident mood. Prompt 3 (simpler, fewer elements): A matte black water bottle with a bamboo lid, photographed straight on against a pure white background, soft even lighting, no props, square 1:1 image, ready to cut out for a product listing.

Example 3: input

What should the image show? Warehouse team doing a morning safety huddle, for our careers page

Image model: Stable Diffusion

Style: Photorealistic

Shape: Portrait 4:5

Example 3: output

Prompt 1 (closest to your idea): photorealistic photo, small diverse team of six warehouse workers in hi-vis vests standing in a loose circle for a morning safety huddle, team lead holding a clipboard, tall shelving and pallets in the background, early morning light through high windows, natural expressions, medium shot, 35mm lens, shallow depth of field, warm documentary style [4:5] Negative prompt: blurry, distorted faces, extra fingers, extra limbs, text, logo, watermark, cartoon Prompt 2 (bolder composition): photorealistic photo from a low angle inside a large warehouse, workers in hi-vis vests and hard hats gathered for a morning briefing, one laughing, long rows of shelving leading to a bright loading door, strong backlight with dust in the air, cinematic contrast, 24mm lens [4:5] Negative prompt: blurry, distorted faces, extra fingers, text, logo, watermark, oversaturated Prompt 3 (simpler, fewer elements): photorealistic close-up of three warehouse workers in hi-vis vests listening to a briefing, soft morning light, plain blurred warehouse background, natural skin texture, 85mm lens [4:5] Negative prompt: blurry, extra fingers, text, logo, watermark

Tips for better results

Say where the image will be used. "LinkedIn header with room for a headline on the right" gets a composition you can actually use.
Pick the shape before you generate. A square image cropped to a 16:9 banner loses most of its composition.
Keep words in the image under six and put them in quotes. Longer text usually comes out misspelled; add it later in Canva or Figma.
Describe a look instead of naming an artist or brand. "Bold flat shapes, thick outlines, two colors" is more reliable and avoids copyright trouble.

FAQ

Is this image prompt generator free?

Yes. Enter your email once to use it (you'll also get Something Big, our free weekly AI newsletter, and you can unsubscribe anytime). There's no account and no credit card.

Does this tool make the image itself?

No. It writes the prompts. Paste them into the image model you use, such as Midjourney, ChatGPT, Stable Diffusion, Flux or Gemini, to generate the picture.

Why are the prompts different for each model?

Models read prompts differently. Midjourney works well with short descriptive phrases and parameters such as --ar, ChatGPT and Gemini follow full sentences, and Stable Diffusion benefits from a negative prompt listing what to exclude.

Can I use it for Midjourney parameters?

It adds --ar for the shape you pick and --no for things to exclude. It leaves out version and style parameters so the prompt works with whatever settings your account uses.

More free AI tools

Related prompts

Prefer to use ChatGPT or Claude directly? These free prompts do similar jobs.