Find Broken Links on Your Website Before Visitors Do

What counts as a broken link

A broken link is any link that does not lead to the content it promises. The most familiar case is a 404 Not Found response, where the server is reachable but the page does not exist at that address. A 410 Gone response is a stronger version of the same thing, telling clients the page was removed on purpose. Server errors in the 500 range, expired certificates, DNS failures and timeouts also break the journey, even though the page may come back later.

Links to images, stylesheets and files belong in the same bucket. A product photo that returns 404 leaves an empty box on the page, and a PDF link that returns an error looks careless on a pricing or contact page. When you plan a check, include every type of reference, not only the blue text links in your articles.

• 404 Not Found: the address does not exist on the server.
• 410 Gone: the page was deliberately removed.
• 5xx errors: the server failed to answer properly.
• Timeouts and DNS failures: the destination could not be reached at all.
• Redirect chains and loops: the visitor is bounced through several addresses or never arrives.
• Broken images and file links: media that no longer loads.

Step one is to crawl the whole site

Clicking through pages by hand finds the broken links on the pages you happen to open, and nothing else. Real coverage needs a crawl that starts from your sitemap or home page, follows every internal link, and records the status code of every link and image it meets. TypoMonitor does this for you: give it a start URL or sitemap.xml and it walks the site, then lists each problem with the page it was found on.

Make sure the crawl includes both internal links (pointing to your own pages) and external links (pointing to other sites). Internal breakage is entirely your responsibility and usually the easiest to fix. External breakage happens without any action on your side, because the other website moved or closed a page, so it needs to be rechecked on a schedule rather than once.

Step two is to sort results by status code and by page

A raw list of errors is overwhelming on a large site. Group it two ways. First by status code, so you can treat all 404 responses with one approach and all 5xx responses with another. Then by source page, so that a single edit session on one article can clear several links at once. The source page matters more than the destination, because that is where the fix is made.

Pay special attention to links that appear on every page, such as the main menu, the footer and the cookie banner. One broken link in a footer is reported hundreds of times in a crawl, yet it is a single edit. Fixing shared elements first removes the largest share of the list in the shortest time.

Step three is to choose the right fix for each link

There are only a few honest answers for a broken link. If the content moved, update the link to the new address, and add a 301 redirect from the old address so that other sites and bookmarks still arrive in the right place. If the content was replaced by something equivalent, point to the replacement. If the content is gone and nothing replaces it, remove the link or the sentence around it. Do not redirect everything to the home page, because visitors who expected a specific article feel misled.

For external links, open the destination in a browser first. A timeout can be temporary, and a page that blocks automated visits may still work for people. If the page really is gone, look for the same information on another reliable source, or link to an archived reference only when it is appropriate. Otherwise remove the link and keep the sentence useful without it.

• Moved content: update the link and add a 301 redirect from the old address.
• Replaced content: point the link at the closest equivalent page.
• Removed content with no replacement: delete the link and tidy the sentence.
• Temporary server error: recheck after some time before changing anything.
• Typo in the address: correct the URL, which is the cheapest fix of all.

Step four is to verify and mark the result

After edits, rescan the affected pages and confirm each fixed link now returns a 200 status. A redirect that was added should end on a 200 page in a single hop, not in a chain of several. In TypoMonitor you can mark an issue as fixed or ignored, so your list always reflects what still needs attention. Ignoring is useful for links you know are fine, such as a partner page that blocks all automated requests.

Keep a short note of larger changes, for example when a whole section of the site was reorganized. When a future scan shows a cluster of new errors, that note tells you immediately whether it is a known migration side effect or a new problem.

Keep it clean with scheduled checks

A one-off cleanup decays quickly. Content editors publish new pages, external sites reorganize, and old campaigns link to offers that ended. The reliable approach is a regular scan with an alert whenever a new issue appears, so you fix one broken link the week it breaks instead of facing a hundred after a year. TypoMonitor runs scheduled scans, sends an email alert for new issues and a weekly summary digest, and there is no free plan, so you pick the plan that fits and start from a paid account.

Starter is 49 USD per month for one site and up to 2,500 pages per scan, Team is 149 USD per month for up to 10 sites with daily scans, and Agency is 499 USD per month for up to 50 sites with client workspaces. Annual billing takes 50 percent off. Subscriptions are not refundable, so choose the plan with your page count and number of sites in mind. Create an account and run the first scan on your site today.