A library's snapshot of your site is not your site kept alive

The Internet Archive has been saving copies of the web since 1996, unasked and free. If you have ever found a dead website again in the Wayback Machine, you owe that moment to a nonprofit library that decided the web was worth keeping. This page is about what that library can and cannot promise a small site. Every claim was tested against one real website, with commands and sources published so anyone can repeat the checks.

One small site, read closely

Anne R. Hanson made silver and gold jewelry in New Hampshire, and her website, annerhanson.com, showed the work: a page for each piece, with a close-up photograph behind it. It was the first site we archived, so we can set the library's copy beside our own and count.

The Wayback Machine's index for annerhanson.com holds 1,574 captures of 258 distinct URLs, from November 2009 to March 2023. Credit where it is due: every one of the site's 87 pages appears at least once. What the index cannot give you is the site.

No single day of the site is whole

Captures accumulate over years, but they do not line up. Of the 87 pages, 55 had their content stored exactly once in thirteen years; where the crawler came back to them, it recorded only that nothing had changed. Computed over every calendar day from 2009 to 2023, the best single day covers 51 of the 87 pages. There is no date on which the archived site is all one site.

Our copy was taken in a single crawl on September 4, 2026: 87 pages and roughly 250 asset files, all from the same afternoon. One moment, complete.

The pictures behind the pages

A jewelry site is its photographs. Crawlers reached the index pages often and captured their thumbnails; the medium close-up photographs fared worse, and 15 of the site's 153 images were never captured at all. The Archive holds the close-up page for bracelet B17, saved in 2016. Open it today and the page loads, but the photograph is a 404, forever, because the image was never crawled on any date. In our copy the same page loads with its photograph, because the crawl that took the page took the image in the same pass.

The record ends before the site did

The last Wayback capture of any annerhanson.com URL is from March 28, 2023. The site stayed up for years after that, and now that it has gone dark, nothing can put those years back. None of this is negligence. A small jewelry site is one URL among billions; the library samples, and it cannot sit with any one site for long. Save Page Now lets anyone request a capture of a single page on demand, and it depends on someone remembering to use it before the site goes away.

Two shorter facts sit in the ledger with full evidence. A capture in the Wayback Machine is not a settled fact: the Archive accepts removal requests and decides them at its own discretion, stating, "We do not make any guarantees beforehand about the outcome of a request." And everything above assumes the Wayback Machine is up: in October 2024 a breach exposed 31 million user accounts and separate denial-of-service attacks knocked the Wayback Machine offline for days. A small site archived only there has tied its afterlife to a single nonprofit's uptime and legal fortunes.

What taking your own copy looks like

Our archive of annerhanson.com is at foreverhost.ing/annerhanson.com: all 87 pages and all of the photographs, with every link resolving inside the copy, served plainly at its own address with no timestamp picker. The site as it was is the site you get.

Use both. Ask the Wayback Machine to save your pages; the library's record is worth being in, and it costs nothing. If the site matters to you, take a real copy while the taking is easy. A library's snapshot of your site is a snapshot. Keeping the site is your job, or the job of someone you choose.