Mining Wikipedia's Dead Links for Expired Domains

Published September 3, 2026

Wikipedia cites millions of external pages, and the web underneath those citations rots constantly. Editors flag the failures, which means one of the most authoritative sites on the internet maintains a public, continuously updated map of dying domains. Some of those domains are expired and catchable, and whoever catches one holds a domain still referenced by Wikipedia. The technique has been quietly used since the 2000s and remains underdocumented. Here is the whole workflow, including the honest parts most write-ups skip: the links are nofollow, the pool is shrinking, and the win rate is low but the wins are unusual quality.

Why are Wikipedia dead links worth mining?

Three reasons. First, curation: a domain cited by Wikipedia editors was, at some point, a source someone considered worth referencing, which filters out an enormous amount of junk before you ever look. Second, the reference itself: a rebuilt domain that still sits in a Wikipedia citation earns referral visits from a page people actually read, and Wikipedia content is endlessly scraped and republished, so one citation tends to propagate into copies across the web. Third, competition: this is manual work with no push-button tool in 2026, so almost nobody does it systematically. Be clear about what it is not: Wikipedia external links carry nofollow, so this is not a direct PageRank play, and anyone selling it as one is selling the wrong decade.

Where does Wikipedia keep its dead links?

When a citation target stops resolving, editors or bots tag it with a dead link template, and tagged articles collect in maintenance categories such as "Articles with dead external links". Every major language edition runs the same system, which matters later, because the German, French and Spanish pools are mined even less than the English one. You can browse the categories directly on Wikipedia, but search operators are usually faster for niche work:

• site:en.wikipedia.org "dead link" yourkeyword finds articles in your topic carrying at least one flagged citation.
• Adding intitle: terms narrows to articles whose subject matches your niche rather than passing mentions.
• Swapping the language subdomain (de, fr, es) repeats the hunt in less contested editions.

Tools in the WikiGrabber style used to automate these queries; most have come and gone over the years, and the manual operators above do the same job without depending on anyone's side project surviving.

From flagged citation to owned domain: the six steps

Vetting: most dead links are not opportunities

Expect a brutal funnel, and treat that as normal rather than discouraging. Most flagged citations point at pages that moved, sites that redesigned, or domains that are still registered and simply broken. Of the genuinely expired remainder, many spent their afterlife as parked pages or worse, which is why the Wayback check in step 4 is not optional: you are looking for a domain whose last real content matches what Wikipedia cited, with no casino chapter in between. For status checks in step 3, WHOIS status codes tell you exactly where a domain sits in the expiry pipeline; redemptionPeriod means the old owner can still restore it, while pendingDelete means it drops within 5 days and can be backordered. When a batch of candidates survives to step 5, a bulk metrics pass through an aggregator saves the domain-by-domain lookups; DomCop handles uploaded lists through its credits system and returns Majestic and Moz data in one table.

Keeping the citation alive after you catch it

Owning the domain is half the value; the citation is the other half, and it survives only if the reference still makes sense. Rebuild the cited content at its original URL, sourced from the Wayback copy, and the reference keeps doing its job. Redirect the domain to a sales page and an editor will eventually cut it, reasonably. Two honest caveats belong here. Wikipedia's own bots increasingly rescue dead citations by swapping in Wayback archive URLs, which shrinks the pool of live opportunities over time, so fresh flags are worth more than old ones. And Google's expired domain abuse policy, published March 5, 2024, draws the line at intent: genuinely rebuilding cited content is explicitly fine, while harvesting Wikipedia-linked domains purely to manipulate rankings is the exact behavior the policy names as spam.

Scaling the hunt without ruining it

Manual mining suits a niche site builder hunting a handful of domains. If you want volume, harvest citation URLs at scale with the same tooling covered in our Scrapebox walkthrough, then run the same six steps in bulk. Pair the finds with the topical filters from our guide to niche filtering so you only chase domains that fit what you are building, and remember this is one method of many; our ranked overview of how to find expired domains shows where it fits. As a rough illustrative expectation, an afternoon of mining in a mid-sized niche tends to yield a shortlist of a few domains, of which one may survive full vetting. Low volume, unusual quality.

Frequently asked questions

Sources

• Internet Archive Wayback Machine
• Wikipedia: domain drop catching
• ICANN: EPP status codes
• WhoisFreaks: domain lifecycle and pending delete
• Google Search Central: March 2024 spam policies