Prompt Generator

Image to Prompt Generator

Upload a picture and get the prompt that recreates it on Midjourney, Nano Banana, GPT Image, FLUX and more. Free, no sign-up.

Already have a prompt? Generate an image

More AI tools

View more
Image Prompt GeneratorPrompts in the syntax Midjourney, Nano Banana, GPT Image, FLUX, Seedream and Stable Diffusion follow.
AI Prompt GeneratorTurn a rough idea into a full prompt for ChatGPT, Claude, Gemini, DeepSeek and Grok.
Video Prompt GeneratorShot, motion and camera prompts for Veo, Seedance, Sora, Kling and Runway.
ChatGPT Prompt GeneratorRole, task, context and a clear output format.
Claude Prompt GeneratorXML-tagged context, instructions and output.
Gemini Prompt GeneratorTask first, inputs, steps and format.
AI Image GeneratorDescribe a picture, choose Nano Banana, GPT Image 2, FLUX.2, Seedream or another model, and get the image. The credit cost shows before you generate.
AI Logo GeneratorType your brand name, what you do and a style. Recraft V4.1 Vector draws the logo as an SVG, so it stays sharp on a browser tab and on a delivery van.
Image to ImageUpload a picture and describe what to make from it. Keep its layout, its subject or only its style, and get a new image on Nano Banana, FLUX.2 or Seedream.
Image UpscalerUpscale an image to twice its width and height with the Clarity upscaler, which adds texture instead of blur. For AI art, small photos and prints.
Background RemoverUpload a photo and get the subject on a transparent background as a PNG. Works on people, products, pets and logos, with clean edges around hair and fur.
Object RemoverName what should go, such as a passer-by, a sign or a cable, and the model takes it out and rebuilds the background behind it. You describe it in words instead of painting a mask.

Prompt library

Prompts with the image each one made, ready to copy.

Browse all
A square clear glass perfume bottle with amber liquid, a black cap and a cream NOIR 07 label, centred on a flat white background with a faint shadow under its base.Nano Banana ProCut a product out onto a clean white backgroundThe small living room widened to 16:9: the lit Christmas tree, grey sofa and patterned rug stay in place, with new bookshelves and a mantel added on the left and a radiator and framed print on the right.nano-banana-2Extend a photo beyond its edgesThe bearded man from the balcony photo as a passport-style ID photo: head and shoulders, facing the camera with a neutral expression, light blue shirt, plain white background and even light.nano-banana-2Passport-style ID photo from a selfieThe beagle photo shown as a smartphone lock screen, with the time 9:41, a date, status icons and flashlight and camera buttons, the phone standing on a tan background.nano-banana-2Preview a photo as your phone lock screenThe woman from the window selfie relit with one soft key light from the left: her far cheek and the wall behind her fall into soft shadow while her face, hair and white t-shirt stay the same.nano-banana-2Relight a portrait with a classic studio lighting setupThe bearded man on his balcony reshot with a shallow depth of field: his face stays sharp while the plants, railings and apartment blocks behind him fall into a soft blur.nano-banana-2Reshoot a phone photo with real camera settingsA crisp, clean photo of a smiling bearded man in a light blue cotton shirt taking a selfie on a city apartment balcony with potted plants and apartment blocks behind him.Nano Banana ProSharpen a blurry or low-resolution photoThe woman from the window selfie with a soft everyday makeup look: a thin brown liner, defined brows, warm blush and a nude rose lip, with her hair, white t-shirt and room unchanged.nano-banana-2Try a makeup look on your own photoA 3x3 grid of the same woman in a white t-shirt on a light grey background, each tile with a different haircut: a short pixie, a chin-length bob, long layers, loose waves, a sleek ponytail, long hair with curtain bangs, a top bun, a straight shoulder-length cut and tight curls.nano-banana-2Try nine hairstyles on your own faceThe tricolour beagle turned into a soft plush toy with floppy felt ears, embroidered eyes, a dark collar and a small fabric tag, sitting on the same oak floor by the window.nano-banana-2Turn your pet into a plush toy

Three images, three prompts

Two prompts recreate the whole picture, in sentences for Nano Banana and in weighted tags for Stable Diffusion. One keeps only the style for a new subject.

  1. Low shot of a skateboarder in an orange hoodie rolling toward a concrete ledge at golden hourWhole image
    Prompt for Nano Banana
    A skateboarder in a faded orange hoodie and dark jeans rolls toward a long concrete ledge across an empty city plaza at golden hour. The camera sits low beside the board, so the wheels and the textured paving fill the foreground while pigeons lift off behind the rider. Warm, low sunlight from behind creates a soft flare and long shadows. Cinematic, slightly desaturated colour, shallow depth of field, wide 16:9 frame.
  2. A sage-green ceramic mug on a travertine block beside a linen napkin in soft morning lightStyle only
    Prompt for FLUX
    [SUBJECT] on a pale travertine block with a folded linen napkin beside it, photographed in the calm, warm style of a premium homeware catalogue. Soft morning light from a window on the left casts a long, gentle shadow to the right. Muted sage, cream and stone palette, matte textures, eye-level camera with an 85mm lens and shallow depth of field, lots of quiet negative space, square 1:1 frame.
  3. A woman in a dark trench coat holds a clear umbrella on a rainy city street at night, with pink and teal neon reflected on the wet pavementWhole image
    Prompt for Stable Diffusion
    (young woman with long dark wavy hair:1.2), gentle smile, navy trench coat, holding a clear umbrella, (rainy city street at night:1.2), (pink and teal neon reflections on wet asphalt:1.1), blurred car headlights, falling rain, bokeh, cinematic colour grade, 50mm, shallow depth of field, vertical 9:16, photo, high detail Negative prompt: blurry face, extra fingers, deformed hands, warped umbrella, text, watermark, overexposed, low contrast

What is Image to Prompt?

It looks at a picture and writes the text prompt that would make an AI image model produce something like it. A vision model reads the subject, setting, composition, light and style, and the prompt is written in the syntax your chosen model follows.

Whole image
Recreates the picture as it is: subject, setting, composition, light, lens and style. Use it when you want a close copy on another model, or a starting point you then edit.
Style only
Keeps the look and drops the content. The subject becomes [SUBJECT], so you can reuse a lighting setup, colour grade or illustration style on something new.
Subject only
Describes the main subject closely (shape, materials, clothing, pose) and leaves the scene as [SETTING]. Good for putting the same product or character somewhere else.
Add a change
Type one change, such as night instead of day or a teapot instead of the mug, and the prompt is written with it.

Why use Image to Prompt?

Most image to prompt tools only copy the picture. Choosing what to capture gives you a prompt you can reuse.

01
Written for the model you use
Midjourney prompts end with --ar and --stylize, Stable Diffusion prompts get weighted tags and a negative prompt, and Nano Banana, GPT Image and FLUX get full sentences.
02
A prompt you can reuse
Style only and Subject only leave [SUBJECT] or [SETTING] in the prompt, so one picture gives you a look or a product you can place anywhere.
03
Free every day
10 prompts a day on each device without an account, 20 a day with a free account.
04
Your image is not kept
Your browser shrinks the image and removes its metadata. It is sent once to write the prompt and is not saved on our servers.

The same image, written for each model

Midjourney
Comma phrases, then --ar and style parameters
Nano Banana
Full sentences: subject, scene, light, camera
GPT Image
Plain description, exact text in quotes, layout
FLUX
Subject first, rich natural language, no negatives
Seedream
Concise scene, style and text in quotes
Stable Diffusion
Weighted tags plus a negative prompt
Any model
Plain description that works in any image tool

How to use Image to Prompt

Three steps from a picture to a prompt you can paste.

  1. 11. Add an imageDrop it in, paste it from your clipboard, or choose a file. Photos, renders, illustrations and screenshots all work.
  2. 22. Pick the model and what to captureChoose where you will paste the prompt and whether you want the whole image, its style, or just its subject. Add a change if you want one.
  3. 33. Copy or run itCopy the prompt into your tool, or run it in the studio on Nano Banana, GPT Image or FLUX to compare the result with the original.

Frequently asked questions

What people ask before they upload.

1What is an image to prompt generator?
It looks at a picture and writes the text prompt that would make an AI image model produce something like it. A vision model reads the subject, setting, composition, light and style, and the prompt is written in the syntax your chosen model follows.
2Is it free?
Yes. You get 10 free prompts a day on each device without an account, and 20 a day with a free account. The image tools share that allowance with the prompt generator.
3Do you keep the images I upload?
No. Your browser shrinks the image and removes its metadata, then it is sent once to our AI providers to check it and write the prompt. It is not saved on our servers or shown to anyone.
4Which models can it write prompts for?
Midjourney, Nano Banana, GPT Image, FLUX, Seedream, Stable Diffusion, Any model. Midjourney prompts end with --ar and --stylize, Stable Diffusion prompts get weighted tags and a negative prompt, and Nano Banana, GPT Image and FLUX get full sentences.
5Will the new image look exactly like mine?
Close, not identical. A prompt describes the picture in words, so the new image keeps the subject, layout, light and style but not every pixel. To keep a face or product exactly, use an image editing model with your photo as the reference, and use the prompt to describe the change.
6How is this different from Midjourney's /describe?
Describe only writes Midjourney prompts and gives you four alternatives to choose from. This tool writes one considered prompt for the model you pick, can capture just the style or just the subject, and can apply a change you ask for.
7Can I use it on photos of people?
Yes. People are described by what is visible, such as age range, hair, clothing, pose and expression. It never names or identifies anyone, and it will not write prompts for sexual images or images of children in suggestive situations.

Turn any image into a prompt

Free, 10 a day, no sign-up. Pick the model and what to capture, then copy the result.

Add an image