Replace Broken Links With Wayback Machine Copies

A dead external link in a WordPress post being swapped for an archived Wayback Machine copy

You wrote a post in 2019. It cites a study, a news story, a vendor’s docs page. Five years later half of those links go nowhere. The study moved, the newspaper rebuilt its site, the vendor got bought. Your post still reads fine, but the sources behind it are gone.

Video: Fix Dead Links With an Old Saved Copy of the Page (Pro)
Video: Fix Dead Links With an Old Saved Copy of the Page (Pro) (1:37)

The usual fix for a dead link to someone else’s site is to point it at a saved copy in the Internet Archive’s Wayback Machine. Some link checkers used to suggest that copy for you. On the WordPress.org forum for Broken Link Checker, a user asked this year: “Does the Cloud version still offer the very helpful support of suggesting an Internet Archive link for a discovered dead link?” The support reply was that the plugin “does not currently offer an option to create an Internet Archive for broken links.” The thread was closed with a note that the team would review the request.

So if you’re stuck there, here’s how to do it yourself, when it’s the right call, and when it really isn’t.

When an archived copy is the right fix

An archived copy is a snapshot of what a page looked like on a given day. That’s exactly what you want when the link was there as evidence. The reader clicks to check what the source said, and the source said it on a specific date.

  • Citations to studies, reports, statistics.
  • Quotes from news articles or press releases.
  • “Here’s the announcement” links to a blog post or changelog.
  • Docs pages you referenced for how something worked at the time you wrote.
  • Anything on a site that has shut down for good.

In all of these, the old version is arguably better than the live one. Even if the page still existed, it might have been edited since, and your post would now be quoting text that isn’t there.

When it isn’t

An archive copy is a museum piece. Nobody can buy from it, sign up on it, or download the current version from it. So for these, don’t reach for the Wayback Machine:

  • Product and pricing pages. Find the product’s new URL. If it’s discontinued, link the replacement or remove the link.
  • Tools and apps. A snapshot of a web app’s homepage doesn’t run the app.
  • Affiliate links. An archived affiliate link earns nothing and the price on it is years old.
  • Downloads. The archive may not have the file, and if it does it’s the old version.
  • Pages that moved, not died. Search the site for the title first. A lot of “dead” links are just a site that changed its URL structure without redirects.

And sometimes the honest answer is to unlink. If the sentence still stands on its own, keep the text and drop the <a>. If the sentence only made sense because of the link, rewrite it.

One more thing before you start replacing: make sure the link is actually dead. Plenty of sites block bots or rate-limit checkers and show up as “broken” when a human gets the page just fine. We went through that in why link checkers report working links as broken. Open each one in a normal browser before you swap it.

How to find the archived copy by hand

The Wayback Machine has a URL pattern you can type straight into the address bar:

https://web.archive.org/web/YYYYMMDD/https://example.com/the-dead-page/

Put the date you want in place of YYYYMMDD. You don’t need an exact match. The Wayback Machine sends you to the capture it has nearest that date. So if your post went out on 14 March 2019, try 20190314.

If you’d rather get a straight answer without clicking around, there’s a free public lookup that returns JSON. No account, no key:

https://archive.org/wayback/available?url=example.com/the-dead-page/&timestamp=20190314

The timestamp takes one to fourteen digits in YYYYMMDDhhmmss form, so a plain date works. If you leave it out you get the most recent capture. The response has an archived_snapshots.closest object with the snapshot’s url, its timestamp, and the HTTP status the page returned when it was captured. If nothing was ever saved, archived_snapshots comes back empty. (Archive.org’s own help page for this is old and says the APIs can change, so if it ever stops answering, check their current docs.)

Pick a copy close to when you wrote the post

This matters more than people think. Use the post’s publish date, not today’s date. The latest capture of a dead page is often the worst one: it might be the domain parking page, a “this site has moved” notice, or the 404 itself. The capture from around the time you linked it is the version you actually read and cited.

Check that the copy isn’t itself an error page

The archive saves whatever the server sent back. That includes 404 pages, login walls, cookie banners that cover everything, and “access denied” screens. Before you use a snapshot:

  1. Open it and look. Is the content you cited actually on the page?
  2. If you used the lookup, check status. Anything other than 200 is a red flag.
  3. If the copy is broken, step back a few months on the Wayback calendar and try an earlier capture.
  4. If no copy shows the content, that link can’t be rescued this way. Unlink or find another source.

Tell the reader it’s an archive

A reader who clicks a citation and lands on a page from 2019 with an archive toolbar on top might be confused. A short label fixes that:

<a href="https://web.archive.org/web/20190314000000/https://example.com/report/">the 2019 report</a> (archived)

Some people also add a rel value or a title attribute. That’s fine, but the visible “(archived)” is the part readers actually see. Keep it consistent across the site so it doesn’t look like a typo.

The manual route is fine for five links. It’s tedious for fifty. You need a list of dead external links first, then the publish date of each post they sit in, then a lookup per link, then an edit per post. If you haven’t got the list part sorted, our guide to finding and fixing broken links in WordPress covers the checking side.

There are also standalone plugins that rewrite dead links to archive copies automatically. The trade-off with fully automatic rewriting is that a link checker’s false positive becomes a real change in your content, and you may not notice. If you go that way, look at what it changed afterwards.

How Linkweave handles it

In Pro, links to other sites are checked after each full link check (daily, or when you click “Check all links now”). It sends one request per site every few seconds, uses short timeouts, and sends a user agent that names your site. If your host blocks outgoing requests, it sends nothing at all.

On the Broken links screen, Find archived copy asks the Wayback Machine (the same free public lookup above, no account) for the copy saved closest to the day the post was published. Use archived copy points the link there, and you can undo that in the URL changer’s history. Find archived copies runs the whole list, one address a second, and the answers are kept for a week. You still pick which ones to use, so the “is this the right fix?” call from earlier stays with you, and you should still open the snapshot to check it isn’t an error page.

There are also broken-link emails, off by default. They go out after each check, or at most daily or weekly, to the addresses you choose. Links that only “did not answer” aren’t mailed, so a slow server on one night doesn’t fill your inbox.

You can get it at the free download page.

A quick routine

  1. Get a list of dead links to other sites.
  2. Open each in a browser. Drop the false positives.
  3. Sort the rest: evidence (archive it), product or tool (find the new URL), nothing worth keeping (unlink).
  4. For the archive pile, look up the capture nearest each post’s publish date.
  5. Open the snapshot. Content there? Status 200? Use it. Otherwise try earlier, or unlink.
  6. Add “(archived)” next to the link.

Do it once a quarter and it stays small. Leave it for three years and it’s an afternoon.

Frequently asked questions

Yes, for citations and sources it’s a common practice, and Wikipedia does it all the time. It’s a poor fit for product pages, tools, downloads or affiliate links, because the archived copy is frozen and nobody can act on it. For those, find the new URL or remove the link.

Which archived copy should I use if there are many?

Use the capture closest to the date your post was published, because that’s the version you cited. The latest capture of a dead page is often a parking page or an error. Open the snapshot and confirm the content you quoted is actually there before you use it.

What if the Wayback Machine has no copy of the page?

Then that link can’t be fixed with an archive. Look for the same content elsewhere (a new URL on the same site, a press release, another source that reports the same thing), or keep the sentence and remove the link. If the sentence depended on the link, rewrite it.