Free tool

Image to prompt, a picture in words.

Get a written description of any picture to reuse as a prompt for an image generator, as alt text, or as a caption. Microsoft's open Florence-2 model looks at the picture in this tab, so it is never uploaded. Pick how much detail you want, edit the words, copy them. No sign-up, no limit.

Drop a picture here, click to choose, or paste one PNG, JPG, WebP or HEIC. It stays on your computer.

This runs in your browser. Your pictures are never uploaded.

How it works

Three steps, nothing to install.

01 / Drop

Choose a picture

A photo, an illustration, a screenshot, or paste one from the clipboard.

02 / Describe

Let the AI write it

Pick a short caption, a detailed one or a very detailed description.

03 / Copy

Edit and copy

Change any word, add a style, then paste it into your image generator.

Questions

Good to know.

Which AI model writes the prompt?

Florence-2 base, an open vision model by Microsoft released under the MIT licence, in the fine-tuned version converted for the browser by the onnx-community project. It was trained to describe pictures, read text in them and find objects; this page uses its three caption tasks.

Is my picture uploaded?

No. The model is downloaded to your browser from Hugging Face, the open model library, and runs there. The first time it is large, about 360 MB with a graphics card (WebGPU) or 275 MB without; the browser keeps it for next time.

Will the prompt recreate the same picture?

No. It describes what is in the picture, the subject, the setting, the colours and the layout, so a generator makes something similar, not a copy. Add words for the style you want, such as "watercolour" or "35 mm photo", for more control.

Can I use it for alt text?

Yes. "Short, one line" gives a caption that suits the alt text of an image on a web page. Read it first: the model can miss what matters in your context or get small details wrong.

Which languages does it write?

English only.