Turn any picture into words. This AI describes your image in detail and writes a prompt for Midjourney, Flux and Stable Diffusion, plus a short caption for alt text. It runs in your browser, so your images are never uploaded.
Describe an image now →- Free
- Runs on your device
- No sign-up
Natural suits ChatGPT, Gemini and Flux. Midjourney adds the aspect ratio. Keywords suits Stable Diffusion.
How to turn an image into a prompt
Three steps, no account.
Add your image
Drop it in, click to choose, or paste it with Ctrl+V.
Press Describe image
The AI studies the picture and writes a short caption and a detailed description. The first run downloads the model once.
Pick a style and copy
Choose Natural, Midjourney or Keywords, edit the text if you like, and copy it.
Why use this image describer
Most image-to-prompt sites upload your pictures and meter free use. This one runs on your own device.
Private by design
Client work, personal photos and unreleased designs are analysed on your device and never uploaded.
Free and unlimited
No credits, no sign-up, no daily limit.
Three results at once
A detailed description, a short caption for alt text, and a prompt in the style you pick.
Edit before you copy
Every result is an editable text box: fix details and add your own style words.
Honest about limits
The AI describes what it sees and can be wrong. You stay in control of the final text.
Useful for learning
See how a picture is put into words, then change one detail and generate a new one.
What people use it for
A few ideas.
| Use | Which result | Tip |
|---|---|---|
| Recreating the look of a picture in an AI image generator | AI image prompt | Add the medium and lighting: “oil painting, golden hour” |
| Learning how to write prompts | Detailed description | Compare the description with the image, then change one detail |
| Alt text for websites and social posts | Short caption | Say what matters for your page, in about 125 characters |
| Product listings and catalogue notes | Detailed description | Check colours and counts before you publish |
| Organising a photo library | Short caption | Use it as a file name or a search note |
Frequently asked questions
What does this image to prompt tool do?
It looks at your image and writes three things: a detailed description, a short caption you can use as alt text, and a ready-to-edit prompt for AI image generators such as Midjourney, Flux, Stable Diffusion, ChatGPT and Gemini.
Is it free?
Yes. There are no credits, no sign-up and no daily limit. The page is supported by ads.
Are my images uploaded?
No. The vision model is downloaded to your browser once and runs on your own device. Your images and the text written about them never leave it.
How accurate is the description?
It is good at the main subject, setting, colours, clothing and mood. It can miscount objects, misread small text, miss the artistic style, and it does not recognise specific people or brands reliably. Always read the text against the image before you use it.
Will the prompt recreate my image exactly?
No. A prompt describes what is in the picture; an image generator then makes a new picture from that description. Expect a similar scene, not a copy. Add style words such as the medium, lighting and camera lens to get closer.
Which prompt style should I choose?
Natural is a plain-English description and suits ChatGPT, Gemini and Flux. Midjourney is the same text with the aspect ratio of your image added. Keywords turns the description into comma-separated phrases, which older Stable Diffusion models prefer.
Why does the first image take longer?
The vision model (about 215 MB) is downloaded the first time and saved by your browser. After that an image takes about a minute on an ordinary computer and longer on phones. Computers with a strong graphics chip can tick “Use my graphics chip”.
Can I use it to write alt text?
Yes. The short caption is written as one plain sentence, which is what screen readers need. Edit it so it says what matters in the context of your page, and keep it to about 125 characters.
Does it read text in images?
It may mention large, clear text. To copy text accurately, use our Image to Text converter.
More free AI tools
All of them run in your browser, with no sign-up.