An archive result is a lead, not proof that you have the complete article. Before quoting or sharing it, establish which page was saved, when it was saved and whether the important material survived.
1. Match the article, not just the headline
Start with the exact public article URL. A canonical URL is the publisher's preferred address; a redirect sends a visitor to another address. Neither proves that two saved pages contain identical text. A moved article, homepage redirect or reused headline can produce a convincing wrong match.
Compare the replay's original domain and path with the requested article, then check the title, byline and subject. Try a publisher-confirmed canonical address separately if the original has moved. Do not silently accept a different article simply because an archive returns a working link.
2. Keep three dates separate
Record the article's publication date, any displayed update date and the archive's capture timestamp. A capture timestamp describes the saved copy; it is not the publication date. A later snapshot might contain a correction that an earlier one lacks.
For a historical claim, choose a capture relevant to the period and compare adjacent captures when available. For current facts, check the publisher's latest version too. A preserved page can accurately reflect old reporting while being outdated today.
3. Check the ending and the evidence
Read past the opening paragraphs. Look for an abrupt ending, subscription prompt, missing table, empty image or references to material you cannot see. A saved paywall is still a valid capture of a page, but it does not supply the hidden article.
Compare the ending, section headings, captions and source links with another authorized version when available. A long word count can include menus and comments; a short count can describe a complete brief. Judge the article's structure and evidence, not length alone. See our reader limitations guide for similar extraction problems.
4. Interpret different services carefully
An empty search means no matching result was returned by that query. A timeout or blocked request leaves the answer unknown. A service homepage loading tells you little about one article's availability.
When the newest Wayback capture is gated or weak, the lookup can try a limited set of older captures of the same article. Arquivo can continue past a weak match within the same request budget. Checks may use cached evidence; they do not inspect every capture or guarantee a complete copy.
Archive.today automatically tries archive.ph, archive.today, archive.is, archive.li, archive.md, archive.fo and archive.vn within the lookup’s request and time limits. These domains share one archive collection. The check follows matching dated or short snapshot links, fetches the saved page and assesses its article text. A working mirror, CAPTCHA, search listing or preview alone does not confirm full article access; even a strong text result does not guarantee completeness.
When connected, the automatic browser service renders the exact saved snapshot within a limited check that may reuse recent results. A matching, substantial article body without an obvious reading gate can support an estimated High Chance rating; it cannot prove every paragraph is present. New snapshots can qualify automatically. A CAPTCHA, teaser, insufficient text or a different original article stays Medium Chance or lower. Completing a CAPTCHA in your browser only affects that browser session; it does not clear a challenge in the server browser.
Site testing notes, 20 September 2026: We used a freely accessible Register article as a control. Our application rated the Wayback result “found / high”; its capture timestamp was 20251112090830. Arquivo's follow-up returned no captures. Archive.today could not be reached by our automated request, preventing verification. Ghostarchive's direct search reported no results. These are dated observations from our technical audit, not current uptime measurements or paywall recovery rates. We did not request new captures.
5. Prepare a safe public URL
Remove ordinary section fragments such as #comments before a lookup. Review query parameters: some identify the article, while others contain tracking information or private access tokens. Do not submit password-reset links, signed account links or private document URLs. If uncertain, navigate to the publisher's public article page and copy that address.
6. Save deliberately and keep a fallback
Treat archive saving as a manual public submission, never private storage or a promise of a complete copy. Our lookup does not guarantee preservation. Keep the original and replay URLs together with the capture date when recording a source.
The Internet Archive's official FAQ explains replay and saving limitations. Arquivo's official URL-search documentation describes its capture lookup. If no usable copy exists, try library and publisher access.