AI Image Captioning

Quality:

Upload Image

Click or drag image file (jpg, png)

Checking your device…

Output

Generated caption will appear here...

Processing speed depends on your device. The first run may take longer while the model is downloaded. Higher quality takes longer and uses more memory.

About this tool

Upload a photo and this tool writes a sentence describing what is in it. Use it to draft alt text for images or to quickly get a sense of what a picture shows.

How to use

  1. Upload a jpg, png or webp photo in the upload area.
  2. If you like, pick a model quality. Higher quality downloads a larger model.
  3. Press the caption button.
  4. Read the description shown below.

Good to know

  • It only works in browsers that support graphics acceleration (Chrome or Edge on a PC, recent phones).
  • Descriptions are written in English.
  • The smaller model is about 310MB and the larger about 830MB. It is downloaded once and kept in your browser.
  • Large photos are scaled down to fit within 1024 pixels before analysis.
  • Small objects, text and busy scenes may be described inaccurately.

Frequently asked questions

Does it work on a phone?
Yes, on recent phones. On a device without graphics acceleration the run button is greyed out and a notice is shown.
Is my photo uploaded to a server?
No. The photo is analyzed inside your browser. Only the model file is downloaded once from an external storage.
Why is the first run slow?
The first run downloads a model of several hundred MB. After that the saved copy is used.
Can it describe photos in Korean?
Not at the moment. Descriptions come out in English only.