Before brands spent April 1 producing fake product announcements through five social platforms at once, people were already using the Internet to lie recreationally.
2meta’s archive collects April Fools material from the older Internet, with yearly pages reaching back into the 1990s. The host site itself describes being online since 1996, and the prank archive preserves examples from an era when the joke might arrive through a website, mailing list, Usenet post, fake advisory, or deliberately ridiculous technical document.
The 1994 collection includes a fake Microsoft bid to acquire the Catholic Church, a NASA Congressional Budget Explorer, an “Internet in a Lunchbox” item, a joke about Cisco acquiring Sears, and several deliberately absurd RFCs.
By 1996, the archive includes a fake CERT security advisory warning about an “Alien/OS vulnerability,” complete with a Windows 95 angle.
Some jokes aged well.
Some require enough historical context that the explanation is now funnier than the prank.
The joke used to have a location
What makes the collection useful is not merely the comedy. It preserves where Internet humor lived before the modern feed flattened everything into screenshots and reposts.
A fake security bulletin worked because readers were accustomed to receiving real security bulletins in similar forms. A joke RFC worked because technical people recognized the dead-serious formatting. A bogus corporate acquisition could circulate because the Web was becoming fast enough for an invented announcement to escape its original page before everybody had instant access to verification.
The medium was part of the prank.
The archive also catches references that would otherwise disappear when old homepages, university accounts, and small organizations went offline. A list of old joke links is imperfect preservation—some destinations are inevitably dead—but even the titles, dates, and descriptions preserve the shape of what people were doing.
Modern April Fools material is often intentionally disposable. It appears in a feed, gets screenshotted, creates one day of confusion, and is buried by April 2.
An old-fashioned yearly index is less glamorous and considerably easier to research twenty years later.
CacheRat’s 1,967 Ancient Web Domains research list includes the kinds of independent sites where projects like this lived. The value of these pages is frequently not that every outgoing link survived, but that somebody kept the index.
History benefits from people who refuse to clean out the junk drawer.