Find the broken links on your website

It starts at your home page, reads your sitemap, and follows the links between your own pages from there. On each page it collects every address — links to pages, images, scripts, stylesheets, icons, embeds, video and audio — and tests each unique one once. Email and phone links are checked for obvious mistakes rather than dialled.

A link is reported as broken only when the server clearly said so. A site that blocks automated checks, asks for a login, or limits how fast it answers is reported as “could not verify” and never counted against you — LinkedIn, Instagram and Facebook all behave that way. A slow answer is a warning, not a failure.

Someone who clicks a broken link has already shown you they want something. Most leave rather than hunt for it. Google treats an address that answers 404 or 410 as content that does not exist, so the link leads nowhere for search either.

Links to other websites break on their own, without you touching anything, when those sites move or close. Expired domains are often bought and reused, which can leave an old link pointing your visitors somewhere you would never recommend. A site that was clean last year rarely still is.

Each problem in the report carries its own source — Google Search Central, MDN, web.dev or the W3C — so you can check the reasoning rather than take our word for it.

Where the check could not cover something — a page it could not read, a page longer than it reads, addresses past its limit, or pages your robots.txt closes to automated readers — the report says so, in its own section, before the findings. A clean list of findings over part of a site is not a clean site, and this will not tell you otherwise. If too little could be tested to say anything, you get no score rather than a bad one.

For what your site shows the public beyond its links — HTTPS, redirects, contact path, share preview — or how fast it loads, run the speed test.