How to find the earliest version of a web page

The earliest version of a web page is rarely sitting at the current URL. Pages move, domains change, redirects hide old paths, and archives capture only some moments. A reliable answer therefore needs two things: a search across the page's URL history and a record of exactly what the earliest capture proves.

The central rule is simple: an archive timestamp proves that a version was captured by that time. It does not prove the page was first published then.

Define the version you are trying to find

“Earliest version” can mean:

• the first archived capture of the exact URL
• the first version containing a particular claim, image, price, or policy
• the page before a later correction or redesign
• the original URL before a migration or redirect
• the earliest evidence that the page existed anywhere

Write the question down. If you need to know when a claim appeared, the first capture of the page may be irrelevant because the claim could have been added years later. If you need the original article, an early homepage capture is not enough.

Record the current page and its URL family

Start with the current page before opening an archive. Save:

• the full URL, including meaningful query parameters
• the final URL after redirects
• the canonical URL declared by the page, if any
• the visible title, author, publication date, and last-updated date
• one distinctive sentence from the body
• links to print, AMP, mobile, language, or syndicated versions

Then list plausible historical variants supported by what you can see:

• http and https
• www and non-www
• old and current domains
• a previous slug revealed by a redirect
• trailing-slash and index-page forms
• print, AMP, mobile, or query-based versions linked by the site

Do not generate dozens of random URL guesses. Follow redirects, canonical tags, old internal links, sitemaps, and feed entries. Each variant should have a reason to exist.

Check the exact URL in the Wayback Machine

Enter the full URL into the Wayback Machine and choose Browse History. Open the earliest available capture, then verify the page itself:

1. Does the archived address contain the URL you intended to check?
2. Does the page title and body match the subject?
3. Did the capture redirect to a different path or domain?
4. Is the content complete enough to support your conclusion?
5. What exact capture timestamp appears in the archive URL?

Record the stable archived URL, not just the year shown on the calendar. A capture can contain missing images, replayed assets from nearby dates, or a redirect rather than the page body. The first dot on a calendar is therefore a candidate until you open it.

Repeat the check for each justified URL variant. A site migration can make the current address look years newer than the same article at its old domain.

Compare captures, not only dates

Open the first usable capture and the next several captures around any important change. Compare the fields that matter to your question:

• title and byline
• visible publication and update dates
• the central claim or policy language
• images, captions, and credits
• prices, product names, or calls to action
• outbound source links
• canonical URL and redirects

For a consequential change, copy the relevant passage into your notes and keep the capture URL beside it. Do not say “the page changed in March” when all you know is that one wording appears in a February capture and another in an April capture. The honest interval is “after the February capture and by the April capture.”

Archive coverage is usually sparse. The Memento web-archiving standard explicitly allows the returned archived date to differ significantly from the date requested. A missing day is not evidence that the page did not exist that day.

When the exact URL has no useful capture

Work outward in a controlled order.

Check the site's archived URL listing. The Wayback Machine can list files it captured under a domain or path. Look for an older slug, date-based directory, print page, attachment, or index page that links to the target.

Inspect old sitemaps and feeds. Archived XML sitemaps, RSS feeds, author indexes, category pages, and newsletter archives often preserve the original URL and title even when the article itself was missed.

Search the exact title or a rare sentence. Quoted text can reveal a syndicated copy, citation, forum link, or feed mirror. An independently dated page linking to the target proves the target existed no later than that date, even if the target was not captured.

Check another web index. Common Crawl provides a public CDX URL index across its crawl collections. Search the exact URL or a narrow URL pattern. A record can reveal that a crawler saw a URL in a particular crawl and may provide the stored response location. It is a second coverage source, not proof that no earlier page existed.

Follow redirects backwards. Search the old domain, title, author page, and inbound links. Migration maps sometimes reveal the predecessor URL more clearly than the current site does.

Treat page dates as claims to verify

A visible “Published” date, structured-data date, HTTP Last-Modified header, or date inside a URL can help, but none proves first publication by itself. Content-management migrations can reset fields. Editors can backdate or correct them. Servers often generate Last-Modified from deployment or cache behavior.

Prefer independent agreement:

• the page's own publication date
• an archive capture at or shortly after that date
• a dated feed or sitemap entry
• a contemporaneous link from another site
• an author, issue, or category index containing the page

If three records agree and nothing earlier contradicts them, you can report a well-supported publication window. Keep the evidence attached to the conclusion.

Keep an evidence log

A small table prevents dates from turning into guesses:

| Evidence | Timestamp | What it proves | Limitation | | --- | --- | --- | --- | | Archived page URL | Capture time | This version existed by the capture | It may have existed earlier | | Dated feed entry | Feed date | The site listed the URL by this date | Feed dates can be migrated | | Independent inbound link | Publication date | Another page referred to it by this date | The linking page's date also needs checking | | Current page metadata | Visible or machine date | What the current site claims | It can be edited or reset |

For each item, record the exact URL, timestamp, page title, relevant text, and the date you accessed it. If the result matters legally or professionally, retain screenshots or files under the evidence rules that apply to your work; do not rely on a browser tab remaining available.

Preserve the current page without rewriting history

If the live page is important and has no current capture, the Wayback Machine's Save Page Now can create one. The Internet Archive says this saves the single page entered and returns a permanent URL; it does not crawl the whole site or its outbound links.

That new capture preserves the page from now onward. It does not fill the gap in the past, and it must never be presented as evidence of what the page contained before the capture time.

State the conclusion precisely

Use language that exposes the remaining uncertainty:

• Earliest archived version found: the oldest usable capture located for the checked URL family.
• Claim present by: the first capture in which the relevant text appears.
• Publication window: after the last capture without the content and by the first capture with it.
• Earliest independently verified existence: an archive, feed, or inbound link establishes that the page existed by this date.
• Unresolved: archive coverage or URL history is insufficient.

Never convert “the archive has nothing earlier” into “nothing earlier existed.” Password protection, crawler exclusions, unlinked pages, script-rendered content, crawl failures, removals, and simple bad luck all create gaps.

If the page contains a photo or video whose own origin is in question, continue with how to find the original source of a photo or video.