Open menu

From visual reference to creative prompt

Image to Prompt Generator

Generate an AI prompt from an image with an editable image-to-prompt generator. Describe the subject, composition, lighting and style in English or Chinese for AI art, image generation and Midjourney workflows.

  • Subject, composition, lighting & style
  • English / Chinese output
  • Four editable formats

Upload an image

JPG / PNG / WebP · 10MB · Up to 12 megapixels

Need an image? Try a sample

Prompt format
General description

A natural-language description for most AI image tools.

Preview locally first. Your image is uploaded only when you sign in and process it.

Checking availability and prices…

Preview the result

See the possibilitiesPre-rendered example

One image. Four ways to describe it.

Explore how each prompt format organizes visual information.

A cream bouclé armchair with wooden legs
Reference image
SAMPLE OUTPUT
A cream bouclé armchair with wooden legs. Centered product close-up with negative space, A clean, warm-gray studio backdrop, Soft natural side lighting with subtle shadows. Cream white and warm gray, Realistic commercial product photography.

These are pre-written format examples, not a live analysis. Load a sample, choose a format and submit when ready.

A simple creative workflow

Understand a reference. Describe your own idea.

Turn visual inspiration into text you can read, edit and reuse.

  1. 01

    Upload a visual reference

    Choose an image to describe its visible subjects, setting, colors and visual style.

  2. 02

    Choose language and format

    Select English or Chinese, and a general, structured, JSON or Midjourney format. Check the price before generating.

  3. 03

    Edit and continue creating

    Adjust the subject, lighting or style, then copy the prompt or choose to open it in the text-to-image workspace.

Made for real creative work

Turn “I like this look” into useful words

Find reusable ideas in product photography, portraits and everyday scenes.

A product still-life exampleOriginal
Example description

A centered product still life in a minimal setting, soft side lighting, tactile materials and a restrained color palette.

Product visuals & creative direction

Unpack the visual language of a product image

Describe placement, background, lighting and colors as a starting point for a new product concept. Keep the atmosphere you like and replace the subject with your own.

  • Describe composition, background and lighting
  • Adapt the description to your own product and brand
Generate prompt
A portrait photography exampleOriginal
Example description

A natural photographic portrait, the person as the main subject, soft light, an uncluttered background and realistic skin and clothing textures.

Portrait style & lighting references

Find words for the mood of a portrait

Describe a portrait through its viewpoint, framing, surroundings and light. Explore another setting or color mood with editable text as your starting point.

  • Describe visible elements and stylistic qualities
  • Adjust framing, setting and color mood
Generate prompt
A pet photography exampleOriginal
Example description

A pet as the main subject with an expressive gaze, soft natural light, visible fur texture and a relaxed, warm photographic mood.

Visual learning & everyday creation

Discover a new direction in an everyday photo

Describe the subject, action, setting and style of a pet or everyday photo. Keep its story, then explore an illustration, a photographic treatment or a different environment.

  • Start with natural language that is easy to read
  • Edit and copy into your creative workflow
Generate prompt

These images illustrate use cases. Your result depends on the source image and selected options.

General / Structured / JSON / Midjourney

Four formats for the way you create

Read a reference in natural language or organize it into a structure you can edit and reuse. These are illustrative format examples.

General description

A natural-language description for most AI image tools.

An orange cat sits by a window, its fur lit by soft morning light against pale curtains. Close-up composition in a natural photographic style.
Structured

Separates the subject, composition, lighting and style for easy editing.

Subject: An orange cat
Composition: Centered close-up
Setting: A window with pale curtains
Colors: Warm orange and cream
Lighting: Soft morning light
Style: Natural photography
JSON

Uses fixed fields for tools and workflows that accept JSON.

{
  "subject": "An orange cat",
  "composition": "Centered close-up",
  "setting": "A window with pale curtains",
  "colors": "Warm orange and cream",
  "lighting": "Soft morning light",
  "style": "Natural photography",
  "prompt": "An orange cat by a window in soft morning light, centered close-up, natural photography."
}
Midjourney

A prompt for Midjourney, with an aspect ratio when appropriate.

An orange cat by a window, pale curtains, soft morning light, centered close-up, natural photography --ar 1:1

Input & output specifications

Input formats
JPG / PNG / WebP
Input limit
10 MB · 12 megapixels
Output languages
English / Chinese
Output formats
General · structured · JSON · Midjourney

Choose the right approach

Describe the image or create with a reference?

Get a written description to understand the image, or use image-to-image when you want a visual reference.

Describe the image or create with a reference?
ConsiderationImage to promptImage-to-image creation
InputA reference image and output formatA reference image and a new creative instruction
OutputAn editable written descriptionA new image based on the reference and instruction
Best fitLearn visual language, describe style and prepare promptsChange a scene, explore a direction or continue editing
Next stepCopy, edit or send to text-to-imageCompare the new visual and download it

FAQ

Frequently asked questions

A few useful details about your files, results and next steps.

Can it recover the exact original prompt?

No. It creates a new description from the visible image, not hidden generation settings. It cannot guarantee the original camera, model, identity or wording.

Does analyzing an image generate another image?

No. Edit or copy the prompt first, then choose to use it in text-to-image or continue with the reference. Image generation is a separate action with a separate price.

Can I generate prompts in English and Chinese?

Yes. Select the output language before submitting, then edit the generated text directly. Choose the language that suits your image model and project.

Which of the four formats should I choose?

General is easy to read and copy. Structured separates subjects, composition, lighting and other elements. JSON can be read by tools that support it. Midjourney is written for that creative workflow and may include an aspect-ratio parameter.

Can it read all the text and generation settings in an image?

The tool describes visible content. It is not a dedicated OCR tool and cannot reliably read hidden generation settings. Verify any exact wording, brand details and technical parameters yourself.

How is image-to-prompt priced?

The workspace displays the actual credit price before you submit. Analysis produces a text description. Using that prompt to generate an image is a separate submission with its own price.

Prepare your assets, understand a reference and take the next creative step.

Explore all tools
1× enhancement

Image Upscaler

Improve image quality with a free online AI image upscaler and enhancer. Upload a photo, product visual or AI image, upscale it to high resolution at 1×, 2× or 4×, and download the result at its original aspect ratio.

Open tool
Automatic subject isolation

Remove Background

Remove the background from a photo online with a free AI background remover. Create a transparent PNG from portraits, product photos and other images, inspect the cutout edges, and download the result for your design.

Open tool

Let one reference spark your next idea

Upload a visual reference and turn what you see into words you can make your own.

Upload an image to get started