Image Description Generator
Write alt text and captions from the picture itself.
This one sends your image to a server to be described. It is held in memory for a second or two and never kept.
Drop an image here
Get alt text for it in about a second. Paste works too.
Looking at the image…
That didn't work.
This tool sends your image to a server to be described. It is held in memory for the second the request takes, never written to disk, and never kept afterwards.
Keyboard shortcuts
- / or ⌘K
- Search every tool
- ⌘↵
- Run this tool, from anywhere on the page
- ⌘S
- Download the result
- ?
- Open and close this list
About the Pathine Image Description Generator
Looks at an image and writes the sentence that should go in its alt attribute — what a reader who cannot see it would need to know. Drop one in and you have it in about a second.
Alt text is the accessibility job most often skipped, and the reason is friction rather than indifference. Writing forty descriptions for a page is dull work and an empty attribute takes no time at all. Removing the friction is most of the fix.
It describes only what is visible. It will not guess who someone is, where a picture was taken, or what a face is feeling — those are the inventions that make alt text worse than useless, because the person relying on it has no way to check.
When you'd reach for it
A page full of images and no alt text
The common case. A first pass that is accurate and needs light editing beats an empty attribute by a distance, and beats the alt text people write when they are bored of it.
Meeting an accessibility standard
WCAG requires a text alternative for non-decorative images. This gets you a real one per image rather than a filename.
Captions for a post or a newsletter
Switch to caption and you get a sentence or two of context in a neutral publishing voice, for the text under the picture rather than inside the tag.
Product images in a catalogue
Descriptions of what is actually shown, written the same way each time — which matters more across two hundred listings than across one.
How to write alt text for an image
-
Drop in the image
Or paste it. JPG, PNG, WebP, HEIC and AVIF, up to 8MB.
-
Choose alt text or a caption
Alt text is one sentence under 125 characters, for the attribute. A caption is a little longer and written to be read alongside the picture.
-
Read it before you use it
It is accurate about what is in the frame and knows nothing about your context. If the image matters because of something off-camera, only you can add that.
-
Copy it into your alt attribute
Straight into the markup, the CMS field or the alt box wherever you are publishing.
Common questions
Is it accurate enough to use as-is?
Usually, and it is worth a glance regardless. It describes what is in the frame well. What it cannot know is why you chose the image — if a photo is there to illustrate a specific point, the alt text may need to say so, and only you know that.
Will it guess who is in the photo?
No, deliberately. It will not name people, place the location, or describe what someone appears to be feeling. Those guesses are confidently wrong often enough to be dangerous in a description someone cannot verify.
What about decorative images?
It will tell you when an image carries no information, and the right answer there is empty alt text — alt="" — so a screen reader skips it entirely rather than reading out a description of a background swirl.
Why is alt text kept so short?
Because it is read aloud, in sequence, without the option to skim. Around 125 characters is the length most screen-reader guidance settles on. If an image needs more than that, it usually needs a caption or a description in the body text instead.
Is my image uploaded?
Yes — this is one of the few tools here that sends anything. Describing a picture needs a model too large to run in a browser tab. It is scaled down on your device first, held in memory for the second the request takes, and never written to disk or kept.
Does it work in other languages?
It writes in English by default. Paste the result into a translator if you publish in another language — a description translated well reads better than one written in a language the prompt did not ask for.