Broken Link Checker
Find dead ends without turning uncertainty into a false alarm. This checker reads a live page, selects a bounded sample of internal destinations and reports which return 404 or 410, which respond, and which remain unknown.
Follow the path. Report what happens.
PUBLIC PAGE / BOUNDED CHECKGood input. Useful output.
Check a bounded sample of links on a live page.
Try the example, or make it your own.How to use the
Broken Link Checker
- 01
Enter the page containing the links
Use the final public HTTP or HTTPS address of a page whose navigation or references you want to inspect. The submitted page is the source, not a bulk list of destinations.
- 02
Review the sample and statuses
The checker requests up to 12 unique same-origin links in groups of three. It uses HEAD requests and respects robots rules, so some otherwise browsable destinations may remain unknown.
- 03
Fix the source link or destination
For a confirmed dead destination, find a relevant replacement, restore the intended page or remove an obsolete reference. Retest and check the actual click path in a browser.
Read the result without over-interpreting it
A 200 response can still be a soft error page or the wrong destination. Conversely, a 403, 405 or timeout can occur while a normal browser visit works. The status categories deliberately separate confirmed HTTP absence from incomplete evidence.
| Observation | What was observed | Next step |
|---|---|---|
| Broken | The destination returned 404 or 410. | Confirm the intended resource and repair or remove its link. |
| Reachable | The request resolved with a 2xx or 3xx status. | Still inspect relevance and visible content manually. |
| Unknown / blocked or server error | The response was not a confirmed 404/410 failure. | Retry and inspect bot protection, HEAD support or server health. |
| Skipped | Robots rules disallowed the request. | Review the intended crawl policy; do not assume the page is broken. |
A mixed sample needs three different actions
Suppose an article links to an old guide, a current resource and a protected area. The report should lead to a specific repair for the guide and a manual check for the protected area, rather than labeling both broken.
/old-guide → 404
/current-resource → 200
/account → 403Broken: /old-guide
Reachable: /current-resource
Unknown / blocked or server error: /accountWhy this is a sample, not a whole-site crawl
The tool extracts links from the returned server HTML and checks at most 12 unique destinations with the same origin. Different subdomains count as different origins. External links are counted but not requested, and links inserted only by JavaScript may not be present in the extracted source.
The bounded sample keeps the check quick and limits unnecessary requests to other sites. It also means a clean report cannot certify every link on a large page. Inspect the reported count of additional internal links and use a full crawler or your publishing checks when you need exhaustive coverage.
Choose a repair that preserves the reader's intent
Update the link directly
If the resource moved, point current content to its final address. A redirect can protect old references, but it need not remain the normal internal navigation path.
Restore missing resources when appropriate
A deleted download or guide may still be the right answer. Restoring it can be more useful than sending every visitor to a general landing page.
Investigate repeat unknowns
Check whether your server rejects HEAD, requires a session or uses bot protection. Do not disable security simply to improve a checker result.
Retest representative templates
One page does not represent the whole site. Check a navigation-heavy page, a recent article and any template changed during a migration.
Broken Link Checker:
questions, answered.
Why are only 12 links checked?
The service intentionally limits the sample and concurrent requests. It reports additional internal links that were not checked, so a successful sample is not presented as complete site coverage.
Why is a working link marked unknown?
The server may block HEAD requests, require authentication, time out or apply bot protection. Unknown means the checker could not establish a normal reachable or confirmed missing result.
Are external backlinks checked?
No. This tool samples outgoing same-origin links found on your submitted page. It is not a backlink index and does not crawl external destinations.
Can the checker detect a soft 404?
Not reliably. A page returning HTTP 200 can display an error message. The report is based on bounded HTTP checks, so review suspicious reachable pages manually.
The calculations and checks above describe this tool's implementation. Use these primary references for the underlying formats and publishing guidance.
HTTP response status codes ↗ (opens in a new tab)HTTP HEAD method ↗ (opens in a new tab)Replace a dead end with a useful next step.
Check a page, investigate the evidence and repair the link that matters to the reader.
Use the Broken Link Checker↑