Find the links that lead nowhere
A page you linked to five years ago is gone, the company behind it folded, and the sentence in your article still points at it. Nobody tells you. This walks your pages, follows every link that leaves them, and reports the ones that are dead — plus, where a whole domain has been given up, whether it is free again.
- 1Enter a page or a domain
- 2Watch the findings arrive live
- 3Take the list as CSV
Nothing checked yet
Enter a page or a domain above. Works on any publicly reachable site, yours or somebody else's.
How a link is judged
A checker that cries wolf is worse than none, because you start pulling good links out of your own texts. So here is exactly where the line sits.
What gets read
The address you enter, then the internal links found on it, breadth first, until the page budget runs out. If those links run out first, the rest comes from the sitemap. robots.txt is read before anything else and its disallowed paths are skipped. Only the delivered HTML is parsed: links a site adds with JavaScript after loading are not in the source and cannot be found.
Dead, and not dead
A HEAD request first, then — for servers that refuse HEAD as a matter of course — a second, ordinary GET with a browser signature. A finding needs an unambiguous answer: 404, 410, a 5xx, or a hostname that no longer resolves. Everything that says more about us than about the link stays out of the list. 401 and 403 mean 'exists, but not for you' — a paywall, a bot block, a country block — and a timeout, an expired certificate or a dropped connection describe the moment, not the link.
Domains that dropped
When a hostname stops resolving, the domain behind it may have been given up rather than merely moved. That is checked against the registry itself over RDAP, the official successor to WHOIS: no entry means nobody holds it. Where the registry gives no clear answer, the result stays 'unclear' instead of turning into a promise.
Why this is free, and stays free
There is no paid API behind this tool, so there is nothing to meter and no account to create. It is HTTP requests and one registry lookup, and both are free at this volume. The plugin this grew out of also called Ahrefs for a domain rating; that endpoint has wanted a key since August 2026 and is deliberately not rebuilt here.
- Fetching pages and testing link targets is plain HTTP from our server. It costs bandwidth and nothing else.
- Whether a domain is free comes from RDAP, run by the registries themselves — DENIC for .de, Verisign for .com, and so on. The IANA bootstrap directory says which server is responsible. No key, no account, no fee.
- The limits — five runs an hour, three at a time, 80 pages, 1,200 links — exist so this stays a tool and does not become somebody's free crawler.
Questions
Nothing is stored. The run happens inside your request and the findings go to your browser only.
Brand kit from a domain: Logo, colours and typeface in one request
Does it check internal links too?+
No, only outbound ones — links that leave your domain. Those are the ones that rot silently, because nobody writes to tell you their site is gone. Internal links break when you change something yourself, and your CMS usually knows about that.
A link shows as fine, but my browser shows an error page.+
Then the server answered with 200 and wrote 'page not found' into the page — a soft 404. Telling that apart from a real page means reading and judging the text, and guessing there would produce false alarms. The gap runs the other way as well: a site that turns away automated visits answers 403 to us, that answer is ambiguous, and so we stay quiet — a genuinely dead link behind a bot block will not appear in the list.
Why is a timeout not counted as dead?+
Because usually it is not. A slow server, a maintenance window, a rate limit at their end: the link works again tomorrow. Only a definite refusal or a hostname that no longer exists gets flagged.
What does 'free to register' actually mean?+
That the registry for that ending reports no entry for the domain at the moment of the check. It is not a reservation and it is not advice. Before you spend money, look the domain up at a registrar and check its past — a dropped domain can carry a history you would rather not inherit.
Do you store what I check?+
No. The run lives inside your request and the findings go straight to your browser. No database, no account, no record of who checked what. Close the tab and it is gone, so download the CSV if you want to keep it.
Can I run it on a site that is not mine?+
Yes. Old link lists on other people's sites are exactly where dropped domains turn up and where outreach starts. The tool reads only what any public visitor sees, respects robots.txt and keeps the volume deliberately low.
It found nothing on a large site. Why?+
Most likely the pages assemble their links in the browser. If the delivered HTML holds no anchor tags, there is nothing to follow, and that is true of many app-style sites. It can also simply mean the site is well maintained, which happens more often than people expect.
Eighty pages here. The same chain runs on eighty thousand.
Reading pages, judging what is on them and producing a clean list is the work underneath most of what we build: catalogues out of supplier data, brochures out of a database, product records checked before they go live. Tell us what you would point it at.