Skip to main content
Whole-site checking

Website Broken Link Checker

Checking one page is easy. Checking a site means deciding how deep to go, what counts as part of the site, and what to do with two thousand findings on Monday morning.

Whole-site checking is not open here yet

TagPipes is finishing continuous link monitoring first. Rather than show a scan box that cannot scan, this page is the useful part: how site-wide checking works and what to do with the output. There is a form below if you want telling when it opens.

The free check that is open today is the Website Accessibility Checker: paste a URL, get a report in a few minutes, no signup.

Where a website's broken links come from

Almost none of them are typos. On a site of any age, broken links arrive in clusters, and each cluster has a cause worth knowing because fixing the cause fixes dozens of findings at once.

A CMS migration is the largest single source: URL structures change, redirects are written for the pages somebody remembered, and everything else silently becomes a 404. Deleting or unpublishing content is the second: the page goes, the twelve links pointing at it do not. Then there is the slow one, outbound link rot, where other people's sites move, restructure or end, and nothing on your side changes at all.

The pattern matters for triage. A migration produces a burst of internal 404s that a handful of redirect rules will clear. Link rot produces a trickle of outbound failures that no rule can prevent and that only a schedule will catch.

Four decisions that change the answer

Two tools disagreeing about a site almost always disagree on one of these.

DecisionThe narrow choiceThe broad choiceWhich to start with
How deep to crawlThe pages in your navigation and sitemapEverything reachable, including archives, tags and paginationNarrow. The deep tail of a blog is where findings go to be ignored
Whether to check outbound linksInternal only: links to your own pagesEvery link, including to other people's sitesInternal first. They are your fault, your fix, and the highest value per hour
How to treat redirectsFollow them; a link that arrives is not brokenFlag every redirect as something to tidyFollow them. Redirect tidying is housekeeping, not repair
Whether to render pagesRead the HTML the server returnsRun the page in a browser and read what exists afterwardsDepends on the site. If your navigation is built in JavaScript, reading the HTML sees almost nothing

Start narrow. A first scan that returns forty findings gets fixed. A first scan that returns four thousand gets a meeting, and then gets forgotten.

Working through a site-wide result

The order that gets a real site fixed rather than surveyed.

1. Group by destination, not by source page

One deleted page produces one finding per page that links to it. Grouping by the broken destination turns a list of two hundred rows into a list of eleven actual problems, which is the number worth looking at.

2. Fix at the destination where you can

If forty pages link to a URL you deleted, a single redirect fixes all forty, including the copies of that link in emails and PDFs you cannot edit. Editing forty pages fixes only the forty pages.

3. Then fix at the source, for the rest

Links to somebody else's dead page have no redirect available. Update or remove them, page by page, worst page first.

4. Re-check, and keep re-checking

Verify the fixes landed, then put the check on a schedule. A site that was clean in March is not clean in June, and nothing about that is a failure of the March fix.

How often to look

Different parts of a site rot at different speeds, which is the real argument for a schedule rather than a habit.

After every release

A deploy that renames URLs is the single most reliable way to create internal 404s. Checking after a release catches them while the change is still fresh in somebody's head.

Monthly for outbound links

Other people's sites break on their own schedule. A monthly pass over external links keeps the rot from accumulating into a project.

Continuously for pages that matter

Pricing, checkout, the main navigation, the top ten landing pages. A handful of pages where a dead link costs money on the day, rather than at the next audit.

After a content migration

Not once. Immediately, a week later, and a month later. Migrations surface problems in waves as caches expire and search engines recrawl.

Before a campaign goes out

The landing page and everything it links to. This is the cheapest check on the list and the one most often skipped.

Never, by hand, forever

The honest failure mode. Manual checking works for about two months, then somebody gets busy. It is the reason monitoring exists as a category.

Frequently asked questions

Tell me when site-wide checking opens

One email when the checker and the monitor are open to use. Nothing else, and no newsletter.

We use your address to tell you when this is available and nothing else. No newsletter.

TagPipes is the platform this will be part of.

Site-wide link monitoring is still being finished and is not in the product yet. What is in it today is tagging, measurement, consent and privacy tooling.

See what TagPipes does today

Free 14-day trial. No credit card.