Keyword-stuffed alt text: find and rewrite it

What it means

One or more <img> alt values look keyword-stuffed: a single word repeated three or more times, or a long, low-variety keyword list. This is distinct from images.alt_too_long, which fires on raw length alone — this rule targets the repetition/variety pattern.

Why it matters

Alt text serves accessibility and image-search ranking. Google's image documentation asks for short, descriptive alt text and warns explicitly against keyword stuffing in alt attributes, which it treats as a spam signal that can harm the page's overall quality assessment. A concise, accurate description does more for both ranking and screen-reader users than a list of repeated keywords.

How to fix it

1. Rewrite the alt to describe what is actually in the image, as you would to someone who cannot see it.
2. Aim for one or two clauses — roughly 8–15 words.
3. Include a keyword only if it accurately describes what is shown, and don't repeat it across images on the same page.
4. For decorative images use alt="" (empty, not omitted) so screen readers skip them.

Example

<img src="/shoe.jpg" alt="running shoes cheap running shoes best running shoes buy running shoes 2026 running shoes">

<img src="/shoe.jpg" alt="Acme Trail 200 running shoe in red">

Alt text describes the image; it isn't a keyword slot. Repeated terms read as spam to Google and are useless to screen-reader users.

When it's not a problem

Long alt text is sometimes legitimate for complex diagrams, infographics or charts. Because this rule keys on word repetition rather than length alone, a long but non-repetitive description will not trip it. If a flagged alt genuinely describes complex visual content without repetition, review before editing.

How Crawlinx detects it

We analyze alt text for repeated keyword tokens and unnatural density. Alt attributes that read as keyword lists rather than descriptions are flagged as stuffed.

How common is it?

1304 audited sites in our corpus currently show this issue. The breakdowns below show which platforms, gatekeepers, verticals and countries are most exposed.