How Bookmend recovers dead bookmarks from the Internet Archive

Install Bookmend free — dead-link detection and Wayback recovery, no account needed

When a bookmarked page goes dead, there is often still a copy of it somewhere. The Internet Archive’s Wayback Machine has been crawling the web since 1996 and holds archived snapshots of hundreds of billions of pages. Bookmend connects dead bookmarks to the right snapshot automatically, using the date you originally saved the link as the anchor point. This post explains the full mechanism — how the Wayback Machine works, how Bookmend queries it, and what the recovery process looks like from start to finish.

Why Dead Bookmarks Need a Recovery Mechanism

Chrome stores bookmarks as URLs. A URL is a pointer — it says where content lives on the web, not what the content is. When the content at that location moves, gets deleted, or the domain stops resolving, the pointer breaks. Chrome has no way to know this has happened unless you click the link and see an error page.

Unlike a service that pre-caches content at save time (which requires server infrastructure and ongoing storage), Bookmend’s approach is to monitor links actively and recover them on demand using an existing public archive. The Wayback Machine provides the archival layer; Bookmend provides the detection, matching, and workflow to surface the right snapshot at the right time.

The Two-Check Rule: Confirming a Bookmark Is Dead Before Recovery

Bookmend does not attempt a Wayback Machine recovery the first time a URL fails. Servers go down temporarily. A scheduled maintenance window, a brief DDoS event, a DNS propagation delay, or a CDN outage can all make a healthy page temporarily unreachable. Flagging those as “dead” and attempting recovery would create constant false positives.

Instead, Bookmend waits. After the first failed check, the bookmark is marked as a candidate for re-verification. A second, independent check is scheduled at least 48 hours later. Only if both checks independently fail — same URL, different times, at least two days apart — is the bookmark marked as confirmed dead and advanced to the recovery stage.

This two-check rule means that by the time Bookmend offers a Wayback Machine recovery, it has already verified the failure twice. The 48-hour gap is long enough to outlast nearly all temporary server issues. When you see a recovery option in Bookmend, the underlying link is genuinely dead.

How the Wayback Machine Works

The Internet Archive’s Wayback Machine stores point-in-time copies of web pages. It operates automated crawlers that continuously traverse the web, following links, fetching pages, and storing copies of what they find. These stored copies are called snapshots.

Each snapshot is indexed by two pieces of data: the original URL and the timestamp at which the page was fetched. Timestamps follow the format YYYYMMDDHHMMSS — year, month, day, hour, minute, second. A URL in the Wayback Machine looks like:

https://web.archive.org/web/20190315142233/https://example.com/article

The long number in the middle (20190315142233) is the timestamp: this particular snapshot was taken on March 15, 2019, at 14:22:33 UTC.

A popular page might have hundreds of snapshots spanning many years, taken at different points in time as the page evolved. A rarely-visited page might have only a handful of snapshots, or none at all.

The Wayback Machine provides an API that accepts a URL and a target timestamp, and returns the closest available snapshot to that timestamp. This is the API Bookmend uses.

What a Snapshot Contains

A Wayback Machine snapshot contains a copy of the HTML of the page at the time it was fetched, along with copies of many of the page’s associated assets — images, CSS stylesheets, and JavaScript files — that were also crawled at around the same time.

When you load a Wayback Machine snapshot in your browser, the Wayback Machine serves the archived HTML and rewrites links within the page to point to other archived versions where possible. This means that clicking links within an archived page often takes you to other archived pages rather than back to the live web.

What a snapshot may be missing: dynamic content loaded via JavaScript API calls that happened after the page first loaded, authenticated or personalized content, embedded third-party resources that the crawler did not capture, and media (video, audio) that was not crawled. For most text-based articles and reference pages, the core content is usually intact in the snapshot.

Why Bookmend Uses Your Original Save Date, Not the Most Recent Snapshot

This is a specific design decision with a concrete reason behind it.

A web page can change substantially over time. A news article published in 2018 might have been updated multiple times since, with corrections, additions, or complete rewrites. A product page might now show a completely different product. A blog post might have been edited to add a paywall, a different conclusion, or additional sections the author added years later.

When you saved a bookmark, you saved it because of what the page contained at that moment. The most recent Wayback Machine snapshot of that URL might show the page as it appeared in 2023 or 2024 — which may be quite different from the version you saved in 2018.

Bookmend uses Chrome’s bookmark metadata, which includes the date you originally created the bookmark, to query for the snapshot closest to that specific date. If you saved a link on November 12, 2019, Bookmend requests the snapshot from the Wayback Machine closest to November 12, 2019 — not the latest snapshot it can find.

This is a materially different approach from simply fetching the most recent archive, and it is more likely to return the content you actually saved rather than a later, possibly very different version of the same page.

How Bookmend Queries the Wayback Machine

Bookmend uses the Wayback Machine’s Availability API. The API accepts a URL and an optional timestamp and returns the closest available snapshot. The query process is:

1. A bookmark is confirmed dead (two checks, 48 hours apart, both failing)
2. Bookmend retrieves the original save date from Chrome’s bookmark metadata
3. Bookmend queries the Wayback Availability API with the dead URL and the save date as the target timestamp
4. The API returns the URL of the closest available snapshot, if one exists
5. Bookmend presents this snapshot to you

This query happens over your internet connection from your device. No bookmark data is sent to any Bookmend server — the query goes directly from the extension to the Wayback Machine’s public API.

What You See Before Confirming

When Bookmend identifies a recovery candidate, it presents:

• The dead bookmark’s original URL and title
• The proposed replacement: the Wayback Machine archive URL for the closest snapshot to your save date
• The snapshot date — when that archived version was captured

You can inspect the archive link before confirming. Clicking through to the archive.org URL lets you verify that the captured version of the page is the content you wanted. If the snapshot is from a significantly different time period than your original save, or if it captured an incomplete version of the page, you can decline the replacement and leave the bookmark as-is.

Nothing in your bookmark folder changes until you explicitly confirm the replacement.

What Happens to Your Bookmark After the Replacement

After confirmation, Bookmend updates the bookmark’s URL in your Chrome bookmarks from the dead original URL to the Wayback Machine archive URL. The bookmark title is preserved. The bookmark folder location and all other metadata remain unchanged.

From that point forward, clicking the bookmark opens the archived version of the page at web.archive.org. The archived version requires an internet connection to load, since it is being served from the Internet Archive’s servers.

If you later want to reverse the replacement — if, for example, the original site came back online — you can manually edit the bookmark in Chrome’s Bookmark Manager (right-click the bookmark and select Edit) to restore the original URL or update it to a new working address. Bookmend does not prevent this or lock the bookmark in any way.

What to Do When No Snapshot Exists

Not every URL has a Wayback Machine snapshot. Pages that were only briefly accessible, hosted on platforms that blocked the Wayback Machine’s crawler, required authentication, or never received significant external links may have no archived copy.

When Bookmend cannot find a snapshot for a confirmed dead URL, it reports this directly: no recovery is available. The bookmark is left unchanged. Bookmend does not substitute an unrelated page, silently delete the bookmark, or offer a misleading recovery path.

In this case, your options are:

• Accept that the content is gone and manually delete the bookmark
• Search for the content through other means (a Google cache, a different archive service, or the content republished elsewhere)
• Keep the dead bookmark in place as a record, marked as confirmed dead

Bookmend surfaces the dead-link flag so you know the status, but what you do with it is your decision.

What Kinds of Dead Links Can Be Recovered

Recovery is most reliable for:

• Articles from established news publications and large content sites
• Wikipedia pages and other frequently-crawled reference content
• Blog posts from personal or independent sites that were publicly linked from other sites
• Documentation pages, tutorials, and reference guides from developer-focused sites
• Government and institutional pages

Recovery is less reliable or unavailable for:

• Content that was behind a paywall or authentication at the time of crawling
• Pages on sites that blocked the Wayback Machine via robots.txt
• Dynamically generated pages with no stable HTML version
• Social media posts and profile pages (most platforms block archiving)
• Small, obscure sites that never attracted enough links to be crawled
• Pages that were very recently published and deleted before the Wayback Machine could crawl them

The Full Recovery Flow End-to-End

To summarize the complete process:

1. Bookmend performs an initial check on each bookmarked URL
2. A failing URL is re-checked at least 48 hours later
3. If both checks fail, the bookmark is flagged as confirmed dead
4. Bookmend queries the Wayback Availability API with the dead URL and the original save date
5. If a snapshot exists: Bookmend shows you the archive URL and the snapshot date
6. You review and confirm (or skip)
7. On confirmation: the dead URL in your Chrome bookmarks is replaced with the archive URL
8. If no snapshot exists: Bookmend reports this, leaves the bookmark unchanged

The only manual steps required are the final confirmation — or the decision to skip a recovery when the snapshot is not what you need.

Install Bookmend free from the Chrome Web Store — dead-link detection and Wayback Machine recovery are both fully free, with no bookmark limit and no account required.