We Scanned Random Websites for Missing Alt Text. Even Wikipedia Failed.

Alt text isn't optional. Under WCAG 2.1 success criterion 1.1.1 ("Non-text Content"), every
meaningful image needs a text alternative — and it's one of the most commonly cited failures in
automated accessibility audits (WebAIM's Million report has flagged missing alt text as the
single most frequent error on the web for years running).

Finding every offending image by hand is the boring part, so we built a small crawler that walks
a site's public pages and flags every <img> with a missing or empty alt attribute. To sanity
check it, we pointed it at a handful of real, well-known domains instead of a synthetic test
page.

What it found

One of the sites we scanned was wikipedia.org. A single-page scan of the portal landing page —
the one that lets you pick a language — turned up an image with no alt attribute at all: the
Wikipedia wordmark logo itself (Wikipedia-logo-v2.png). One of the most-visited, most-audited
sites on the internet, and its own front door ships a logo image a screen reader has no way to
describe.

That's not a dig at Wikipedia specifically — it's the point. If a site with that much scrutiny can
ship an unlabeled image on its landing page, it happens everywhere, silently, on every site that
doesn't check for it systematically.

Why this keeps happening

A few patterns we see repeatedly when auditing real sites:

• Logos and icons get skipped. Teams treat "obviously branding" images as exempt from needing alt text, but a screen reader user still needs to know what's there (or that it's purely decorative and should have alt="", not a missing attribute).
• Alt text is added once, at launch, and never re-checked. New pages, new CMS templates, new marketing images ship without it because nothing in the pipeline catches the omission.
• "Missing" and "empty" get conflated. A missing alt attribute is a hard WCAG failure. An empty alt="" on a genuinely decorative image is correct. Telling the two apart at scale is exactly the kind of repetitive checking that's easy to skip by hand and easy to automate.

Checking your own site

If you want to see this on your own domain rather than take our word for it, we put a free,
no-signup version of the same scanner online: it crawls up to 10 public pages, lists every image
with missing or empty alt text, and drafts a ready-to-paste alt text for each one using a vision
model.

Try the free scanner →

No login, no email required for the sample scan — paste a domain and see what it finds. The paid
version (CSV + report, up to 300 images) is there if you want the full site covered end to end,
but the free scan alone is enough to tell you whether it's worth doing.