# Image to Prompt

> Upload a picture and get a text-to-image prompt that recreates it, for ChatGPT, Midjourney, Stable Diffusion or Flux.

- Type: free browser tool
- Category: AI & LLM
- Page: https://www.sandbase.ai/tools/ai/image-to-prompt

## About

Upload a photo, artwork or screenshot and get a prompt that describes it: subject, setting, composition, light, colors and style. Pick natural language, Midjourney or Stable Diffusion. Free runs each day, no sign-up needed.

## How it works

1. **Step 1** Drop a picture (JPG, PNG or WebP).
2. **Step 2** Pick the prompt style (natural language, Midjourney or Stable Diffusion) and short or detailed.
3. **Step 3** Select Get prompt, then copy it into your image generator and edit what you want to change.

## FAQ

### Which style should I pick?

Natural language for ChatGPT, Gemini, Flux and most new image generators. Midjourney for comma-separated phrases with --ar set to your picture’s shape. Stable Diffusion for tags plus a negative prompt.

### Will the new image look exactly the same?

No. A prompt describes the picture, so an image generator makes a new picture in the same spirit: similar subject, setting, light and style. Add or change details to get closer.

### Does it name people in the picture?

No. People are described by what is visible, such as hair, clothes and pose, never by who they are.

### Which pictures are refused?

Pictures with nudity, sexual content or graphic violence. A refused picture does not use a run.

### Is my picture stored?

The picture is shrunk to 1024 pixels in your browser and sent to the vision model to write the prompt; SandBase does not save it. The same picture and settings within a day are answered from a copy, without using a run.
