How to find a Wayback Machine capture and cite the right date

A Wayback Machine link can show which version of a page was archived at a particular time. To use one as evidence, find a capture for the exact page, copy its final archived URL, and keep its capture date separate from any publication date printed on the page. A dated capture can support what appeared in that copy; it cannot, by itself, tell you when every change happened.

Find the exact page

Enter the original page URL in the Wayback Machine’s Browse History field. If you are researching a removed explanation, start with its old article address rather than the publisher’s homepage. The Internet Archive’s Wayback guide describes the calendar’s color coding: blue marks a 2xx HTTP response, green a 3xx redirect, orange a 4xx response, and red a 5xx response. Choose a capture and check what it actually displays.

Copy the archived URL from the address bar after the page loads. Check it again after clicking a link inside the capture: a linked page may open at the nearest available archived date or take you to a live URL. If the address is back on the original domain, you are looking at a live page, not the capture you meant to cite. If its timestamp changed, record the timestamp in the address you actually reached.

Read the capture date

A standard capture URL contains a 14-digit timestamp in YYYYMMDDhhmmss order: year, month, day, hour, minute, and second. The Internet Archive’s documented example is:

https://web.archive.org/web/20000229123340/http://www.yahoo.com/

Its timestamp reads February 29, 2000, at 12:33:40 UTC. This is a documentation example, not a freshly verified replay of that page.

Digits	Meaning	Example
2000	Year	2000
02	Month	February
29	Day	29
12	Hour	12 UTC
33	Minute	33
40	Second	40

The capture date says when that archived copy was recorded. A publication date printed in the page is a separate claim about the original work. Your access date says when you consulted the capture. Keep all three distinct when more than one is relevant.

For a research note, a reusable citation is: [actual page title], [author or organization if shown], captured [capture date], Wayback Machine, [exact capture URL], accessed [access date]. Fill in the title and attribution only if you can see them in the material you are citing. Do not turn the original URL or an assumed site name into an invented author.

Compare what the captures show

Suppose you are investigating a café menu change. You find one capture dated Monday, September 7, 2026, and another dated Friday, September 11, 2026. Those dates are illustrative; they are not a claim that either capture exists. For each real capture you find, note the exact menu page URL, the item you checked, the amount actually shown, and the capture URL.

If the two copies display different amounts, you can say what each archived copy shows on its respective date. You cannot conclude from those two copies alone that the café made the change on Wednesday. The same care applies to a missing chart or paragraph: its absence from one archived view does not, on its own, prove that the publisher deleted it.

Keep a citation note

The Wayback citation workbench is a standalone HTML file you can save and open locally. Paste the final capture URL, then enter the title you observed, an author or organization if one is shown, your access date, and an observation in your own words. Use its missing-details field for anything you could not establish. Select Build note to see a plain citation draft and the JSON research note. Copy the citation into your notes, or select Download JSON to save wayback-citation.json beside your research.

The workbench reads a Wayback URL’s timestamp and original page address, including the original query string and fragment. It checks that the timestamp is a valid UTC calendar date. It does not confirm that a capture exists, that its page opens, or that your observation is true. An empty observation supplies no evidence; if retrieval failed, record that plainly instead of describing unseen content.

A calendar address containing * does not identify a specific archived copy. The notebook rejects it, along with short timestamps, impossible dates such as February 29 in a non-leap year, and addresses hosted outside web.archive.org. Return to the opened capture and copy its complete address before building the note. The title and observations remain yours to supply; the notebook never retrieves the archived page.

If the page will not open

First check the precise address you entered. An old link may give you a reason to try an http or www variant, or a different documented path. Keep the specific article path in the search where possible: a capture of the homepage does not establish what appeared on a removed article.

Treat an unavailable capture and a request restriction differently. The Internet Archive’s availability API accepts a URL and an optional timestamp and can return the closest capture if one is available. “Closest” does not mean the exact date requested. The recorded API requests for example.com and info.cern.ch returned HTTP 429, so they supplied no capture data; that result cannot establish that those pages lack historical captures. You can use the calendar without using the API. If the archive does not provide the page you need, a publisher’s changelog or an explicitly dated static version may provide another source.

For a page that is available now, Save Page Now is for a single page, not its outgoing links or an entire site. It cannot reconstruct an earlier version that was never captured. Avoid submitting private account pages or URLs containing tokens.

Source check: Internet Archive documentation was reviewed October 7, 2026. The offline notebook was checked in isolated Chrome 154 with five URL cases, visible citation output, an actual JSON download, and a 390-pixel view. The Wayback calendar, historical page contents, and current Yahoo replay were not verified.