Broken link checker
Enter a page and we check every link, image, script and stylesheet on it, then show you the exact HTTP status of anything broken. Great for catching 404s before your visitors (or Google) do.
How it works
- We fetch the page you enter and pull out every link (
<a href>), image, script and stylesheet URL, then remove duplicates so each unique URL is only checked once. - Up to 50 URLs are checked per run, around 10 at a time, each with a 7-second limit. Servers that reject the initial lightweight HEAD request get a follow-up GET, so servers that mishandle HEAD are not wrongly reported as broken.
- Results are classified honestly: a real 4xx/5xx or unreachable host is broken (red); a 401, 403 or 429 usually means bot protection, and a timeout could be a slow but healthy server, so both show amber as could not verify and never inflate the broken count.
- Everything runs on demand and nothing is stored. The default view lists only issues; use the toggle to see every checked URL.
What to do with the results
- Fix internal links first. A broken link to your own page wastes crawl budget and strands any authority the linking page was passing. Correct the URL, restore the target page, or add a 301 redirect to its closest replacement.
- Update or remove dead outbound links. Find the current location of the content and relink, or drop the reference. Leaving 404s in place erodes reader trust and makes content look unmaintained.
- Spot-check the amber ones in a normal browser. Most amber results are bot-defended sites that load fine for a human. Only treat them as broken if they fail for you too.
- Re-run after fixing. A clean pass confirms the repairs, and running the check on your highest-traffic pages first gets the most value from each run.
Fixing today's 404s is a one-off win, but links rot continuously as other sites restructure. Ongoing website protection catches the next breakage automatically instead of waiting for a visitor to report it.
Frequently asked questions
It fetches your page, extracts every link, image, script and stylesheet URL, removes duplicates, then requests each one. A response of 400 or above (excluding 401, 403 and 429) or an unreachable host counts as broken; anything under 400 is fine. Bot-blocked responses and timeouts are reported separately as amber "could not verify" results, so the broken count never includes links that are probably fine but refused an automated check.
Amber means the target refused or stalled the automated request: it returned 401, 403 or 429 (usually bot protection or rate limiting) or did not answer within 7 seconds. Sites behind Cloudflare and similar services often do this even though the page loads normally in a browser. Spot-check amber links by opening them yourself; only red links returned a genuine error or were unreachable.
Up to 50 unique URLs per run, checked around 10 at a time so the whole thing finishes within the request time limit. If your page contains more, the tool tells you how many it found and that the list was cut off. It also retries politely: servers that reject the initial HEAD request get a follow-up GET, which catches servers that mishandle HEAD and would otherwise look broken.
Broken internal links waste crawl budget, stop authority flowing between your pages, and can leave content effectively orphaned; broken outbound links are mostly a user-experience and credibility problem. There is no direct ranking penalty for a 404, but the crawl and trust effects are real, which is why the usual priority is fixing internal broken links first, then updating or removing dead outbound references.
This tool checks a single page on demand, so it tells you the state of the links at that moment. Link rot is continuous: pages you link to get moved, merged or deleted without warning. SiteGauge can monitor your key pages on a schedule and alert you when their content changes or they start returning errors, so you hear about breakage before your visitors or Google's crawler find it.