The first search engine was Archie, created in 1990 by Alan Emtage, a student at McGill University in Montreal. It predated the World Wide Web entirely and indexed file names sitting on public FTP servers. It could not search page content, because there were no pages yet — only directories of downloadable files.
That last detail matters more than it sounds. We tend to imagine search as something that reads and understands text. Archie did nothing of the kind. It was a glorified card catalog for the pre-web internet, and everything that followed — Veronica, WebCrawler, Lycos, Google — was an argument about how much more a search tool should actually do.
What was the first search engine, exactly?
Archie launched in 1990. Emtage wrote a script to automatically log into anonymous FTP servers, list their contents, and store the file names in a searchable database. The name was simply “archive” with the “v” dropped, not a nod to the comic strip — though the comic-strip coincidence would haunt its successors.
To use it you needed to already know the file existed, or at least guess part of its name. There was no ranking, no relevance score, no notion of a “best” result. Archie returned a list of servers where a matching file lived. For the early-1990s internet — a place that looked nothing like the first website, which Tim Berners-Lee would not put online until 1991 — that was a small miracle.
What came after Archie? Veronica and Jughead
Once Archie proved file indexing worked, the Gopher protocol arrived in 1991 and needed its own search layer. Gopher organized the internet into nested menus, and two tools sprang up to index it, both leaning hard into the Archie comic-strip joke.
Veronica (1992, University of Nevada) indexed menu titles across the global Gopher space. Jughead arrived shortly after and searched a single Gopher server’s menus rather than the whole network. Between them they made Gopherspace navigable. Then the Web ate Gopher’s lunch, and the search problem moved to HTML.
It is easy to forget how plausible Gopher once looked. For a stretch in the early 1990s, Gopher traffic rivaled or exceeded web traffic, and Veronica was how most people found anything in it. The protocol’s licensing missteps and the Web’s richer formatting decided the contest, but the search-tool lineage carried straight across — the idea of a single bot crawling a network and stuffing names into a database never went away. It just changed what it crawled.
What was the first web search engine?
The Web needed tools that understood URLs and hyperlinks, not FTP directories or Gopher menus. Several contenders appeared in quick succession:
- W3Catalog (September 1994, Oscar Nierstrasz at the University of Geneva) is often cited as one of the earliest. It worked by reformatting existing curated lists of websites into a searchable index.
- Wandex came from the World Wide Web Wanderer, a bot Matthew Gray built at MIT in 1993 — originally to measure how big the web was, a question we still chase on how many websites are there.
- ALIWEB (1993-94, Martijn Koster) let site owners submit their own descriptions, sidestepping the need to crawl. Good idea, poor adoption, because it relied on humans bothering to register.
The breakthrough was WebCrawler, launched in April 1994 by Brian Pinkerton at the University of Washington. It was the first to index the full text of pages it crawled, not just titles or file names. You could finally search for a word and find pages that contained it anywhere. That is the model every modern engine still uses.
The commercial boom: Lycos, AltaVista, Yahoo
After WebCrawler, search turned into a land grab. The mid-1990s produced a cluster of engines that briefly defined the web.
| Search engine | Launched | Notable for |
|---|---|---|
| Archie | 1990 | First search engine; indexed FTP file names |
| Veronica | 1992 | Searched Gopher menu titles |
| WebCrawler | 1994 | First full-text web search |
| Lycos | 1994 | Carnegie Mellon spinout; huge early index |
| Yahoo | 1994 | Human-curated directory, not a true engine |
| AltaVista | 1995 | Fast, large index; advanced query syntax |
| 1998 | PageRank link-based ranking |
Lycos came out of Carnegie Mellon in 1994, named after the wolf spider genus Lycosa for the way it crawled, and grew one of the largest indexes of its day. AltaVista, launched by Digital Equipment Corporation in 1995, dazzled people with sheer speed and a then-enormous index, plus boolean operators power users loved. DEC partly built it to show off the muscle of its Alpha servers, and for a couple of years AltaVista was the answer when someone said “search the web.”
There were others — Excite, Infoseek, HotBot, Inktomi feeding results behind the scenes. The field was so crowded that “metasearch” engines appeared just to query several at once, because no single index felt complete. The clutter was a symptom of the same unsolved problem: plenty of engines could find pages, but none could reliably tell you which one was good.
Yahoo deserves an asterisk. Founded in 1994 by Jerry Yang and David Filo as “Jerry and David’s Guide to the World Wide Web,” it was a hand-built directory — humans sorting sites into categories — not a crawler-driven engine. It dominated discovery anyway, which tells you ranking was the unsolved problem. This was the same era that gave us the first banner ad, when the ad-supported web was being invented in real time.
How did Google change search? PageRank and BackRub
By 1997 the engines all had the same flaw: results were ranked mostly by how often a word appeared on a page, which spammers gamed mercilessly. Larry Page and Sergey Brin, then Stanford PhD students, asked a different question — not how often a page used a word, but how many other pages linked to it, and how trustworthy those linkers were.
That algorithm became PageRank, treating links as votes. The project’s first name was BackRub, because it analyzed the web’s “back links.” The name mercifully did not survive. Google launched in 1998, and within a few years the crowded field of the 1990s had collapsed into its shadow.
The leap from Archie to Google was a leap in ambition. Archie told you where a file lived. Google tried to tell you which of a billion pages was worth your attention. For more on how these milestones stack up, the full internet history timeline puts search alongside email, domains, and the web itself — and if you want to feel how the early internet grew one piece at a time, the clicker game on this site lets you build it from a single click.
Frequently Asked Questions
What was the first search engine ever made?
Archie, built in 1990 by Alan Emtage at McGill University. It indexed the file names on public FTP servers, predating the World Wide Web. It could not read page content because web pages did not exist yet.
Was Google the first search engine?
No. Google launched in 1998, nearly a decade after Archie and years after WebCrawler, Lycos, AltaVista, and Yahoo. Google’s contribution was PageRank, a link-based ranking system that produced far better results than its rivals.
What was the first full-text web search engine?
WebCrawler, launched in April 1994 by Brian Pinkerton. It was the first to index the complete text of the pages it crawled, letting users search for any word that appeared anywhere on a page rather than just titles or file names.
Why was the first search engine named Archie?
The name came from “archive” with the letter “v” removed. The comic-strip resemblance was a coincidence, but it inspired the later Gopher search tools Veronica and Jughead, which leaned into the joke deliberately.
What was Google originally called?
Google’s research project was first called BackRub, named for its analysis of the web’s back links. Larry Page and Sergey Brin renamed it Google, a play on “googol,” the number one followed by a hundred zeros.