alt text audit

Which images say
nothing, and which chose to.

An empty alt is a decision — this picture carries no information, skip it — and a missing one is an omission. This reads every img, image button and image-map link on a page and reports the two apart, so the count you act on is the count that is actually wrong.

It never judges what an alt says. alt="image" is poor alt text and this report calls it present, because whether there is a description is a fact about your markup and whether it is a good one is a judgement about a picture this tool never fetches. It reads img, input type="image" and area — not inline SVG, not CSS backgrounds — and it reads tags rather than nesting, so it says so when a page has a noscript or template block it cannot see into.

This tool works differently from the generator. The generator never uploads your image — it runs in your browser. The checker cannot: it fetches the public address you type, from our server, and reads that page’s HTML in full, not only its head — the body is where images live, so the whole of it is fetched and read here, then turned into findings and dropped: it is never stored, and never sent back. The pictures themselves are never fetched at all; only the markup around them is read. Only the address you enter is sent. The addresses you check are then kept in this browser, so it can tell you what changed since last time; they go nowhere, and you can clear them.

Questions

Does this read my images, or only the markup?

Only the markup. The page is fetched and the img elements are read for what they declare; the image files themselves are never requested. So it can tell you an alt attribute is missing or empty, and it cannot tell you whether the words in one that is present describe the picture.

What does my site see when this runs?

One request for the page, from our server rather than from your browser. The report names the user-agent that was sent, so the visit can be matched against your own logs.