Outbound links
External link checker for a whole website
Links to other sites rot quietly. The article you cited moves, the supplier rebrands, the tool you recommended shuts down, and nothing on your own site changes to tell you.
Enter a domain. After the crawl, every destination on another site is asked whether it still answers, and the ones that do not are listed with the pages that link to them.
What you get
The Links to other sites check runs once the crawl has finished, usually within a few minutes. Each row is one link that needs a look:
- The linking page, where the edit has to be made.
- The destination, as written in your HTML.
- The outcome and the status code.
- The anchor text, which usually says what the link was for.
Confirmed dead destinations also appear as tasks, grouped so that one dead address linked from thirty pages is one job.
What each outcome means
| Error | The destination answered with a 4xx or 5xx: the page is gone or the server is failing. This is a broken link. |
|---|---|
| Redirects | The link works and lands somewhere else. Worth updating if the new address is permanent. |
| Refused the request | The site refused the check (401 or 403). It may be fine in a browser; open it to see. |
| Rate limited | The site asked us to slow down (429). Undetermined, not broken. |
| Timed out | No answer in time. Often temporary. |
| Could not be reached | The host did not resolve or would not connect: an expired domain, or a typo in the address. |
Destinations that answered normally are not listed: the table is the work, not the inventory. Until the check has run, the card reads “Not checked” rather than passing.
How the check is made
The crawl records every outbound link and follows none of them. Afterwards the distinct destinations are taken, most-linked first, and each is asked with a light request that fetches no page body; if a server refuses that kind of request, it is asked again the ordinary way. Redirects are followed up to five steps so the answer is about where the link ends up.
Different sites are checked side by side and each site one request at a time with a pause between, so a page that links to one host two hundred times does not turn into two hundred requests at once.
Fixing what it finds
The page moved. Find its new address and update the link. If the outcome is “Redirects”, following the link shows the new address.
The page is gone. Link to an equivalent source, or to an archived copy if the reference matters, or remove the link and adjust the sentence.
The domain expired. Remove the link promptly. Expired domains are bought and repurposed, and a link from your site then points wherever the new owner likes.
It is in a template. A dead link in the footer shows as one destination with every page linking to it. One edit clears it.
Where it stops
It checks 300 destinations per scan and says how many it left. It can tell that a destination answers, not that it still says what you cited it for: a page replaced by something unrelated, or a parked domain that answers 200, passes. A destination that blocks automated requests has to be opened by hand.
Links inside your own site are checked in full by the same crawl: see the broken link checker and the internal link checker. It runs 39 checks in all.
Questions people ask
- How do I find all the external links on a website?
- Crawl the site: every link to another site is recorded with the page it is on. Enter the domain above. After the crawl, Links to other sites lists the destinations that did not answer cleanly, and the downloadable database holds every outbound link, working or not.
- Do broken external links hurt SEO?
- A few do no measurable harm. They cost you with readers, who follow a recommendation and reach an error, and a page full of dead references reads as abandoned. That is why the check scores them lightly: other people's pages disappear for reasons that are not your fault.
- Why are some links marked refused or timed out instead of broken?
- Because that is all that is known. Many sites refuse automated checks or answer slowly, and their pages are perfectly alive in a browser. Reporting those as broken would send you to fix links that work. Only a destination that answered with an error is counted as broken.
- How many external links does it check?
- The 300 destinations your site links to most, per scan. A site with more has the rest reported as unchecked, never as fine. The crawl itself covers up to 5,000 pages.
- Does it crawl the other websites?
- No. It never follows a link off your site. Each destination gets a single request asking whether it answers, with hosts checked one request at a time. An embedded video or map is not treated as a link at all.
Check your links to other sites
Read next
- Website structure visualizer — an interactive map of the internal links: rings by click depth, pages sized by inbound links, orphans and errors coloured.
- Search a website for a word — four ways to search a whole website for a word or phrase, what each one misses, and a crawl that checks every page.
- Find all PDFs on a website — four ways to list every PDF a site links to, and a crawl that gives each file with its size, status and linking pages.
- The free site crawler — the crawl all of these checks read from, and what else it records.