Clusty and Yippy Search Engine: What Happened

By Bill Hartzer · Updated 27 October 2026

What Clusty was, and where its results came from

Clusty was a web search engine developed by Vivísimo, Inc. and released in 2004. In 2010 it was sold to Yippy, Inc. and rebranded, which is why the same engine is searched for today under both names. It is listed here as no longer available. From late April 2021 yippy.com redirected to DuckDuckGo, and as checked on 27 October 2026 it serves a page of unrelated AI and technology content fragments with no search function. A request to clusty.com from this site's research environment failed, so its current state is unconfirmed.

The essential structural fact about Clusty is that it was a metasearch engine: it had no crawler and no index (the stored copy of the web that an engine builds by crawling) of its own. It sent each query to other search engines and worked on what came back. During its lifetime its verdict on this site would have been resold results. That is not a criticism; metasearch is a design choice, and Clusty's own contribution was what it did with the results after it received them.

Wikipedia's description is exact: "Vivisimo's public web search engine Clusty was a metasearch engine with document clustering." The clustering was the product.

How clustering at query time worked

Every ranked list of search results has the same weakness. A query such as jaguar or mercury carries several unrelated meanings, and a single column of ten links mixes them together. Clusty's answer was to sort the results by meaning before showing them.

It worked in three steps. Clusty fetched results from other engines. It then read the titles and snippets those results returned, found phrases that recurred across them, and built a tree of labeled folders down the left side of the page. For jaguar, that meant separate folders for the car, the animal, the Mac OS X release and the football team. A user opened a folder to see only that subset of the results.

The important detail is that the folders were built at query time, from the returned snippets, not drawn from a pre-computed taxonomy. The labels changed with every query. That is the distinction between clustering and a directory: a directory files the web into categories decided in advance by editors, while Clusty discovered the categories in each result set as it arrived.

Which engines Clusty queried is not documented in any dated primary source retrieved for this page, so no panel of source engines is named here. Accounts that list a specific set should be treated with caution unless they cite one.

The clustering engine itself came out of research at Carnegie Mellon University and was the same technology Vivísimo sold to enterprises and government. Vivísimo technology, together with MSN, was reported in September 2005 to power the US government portal FirstGov; that is recorded by Wikipedia, citing Search Engine Watch, and was not confirmed from the original article.

Why metasearch existed in the first place

Clusty arrived late in a tradition that began a decade earlier, and the original rationale is on the record. When MetaCrawler launched at the University of Washington on 7 July 1995, its authors justified it in their founding paper: "No single search service is sufficient... no single service is able to return more then 45% of the references followed by users." MetaCrawler queried six services and handled 50,878 completed queries between 7 July and 30 September 1995. Those are historical, point-in-time figures, and all six services it queried are now defunct or transformed.

The peer-reviewed measurement came soon after. Lawrence and Giles, writing in Science in 1998, found that as of December 1997 the publicly indexable web held at least 320 million pages, and that no single engine indexed more than about one-third of it, with HotBot leading at 34 percent. The 320 million figure is explicitly a lower bound for that month. On those numbers, combining several engines was a reasonable way to see more of the web.

Clusty, released in 2004, inherited the method but made a different case for it. Its contribution was not more coverage but organization: not "see more results" but "see the results sorted into meanings."

Who ran it: Vivísimo, Yippy and IBM

Vivísimo was founded in Pittsburgh in 2000 by three Carnegie Mellon computer scientists, Raul Valdes-Perez, Jerome Pesenti and Chris Palmer. Its business was enterprise search software, sold as Velocity. Clusty was a public showcase for that software, carrying advertising on its results. TechCrunch reported in April 2012 that Vivísimo had more than 140 customers in government, life sciences, manufacturing and finance, among them Airbus, the US Air Force, the Social Security Administration, the Defense Intelligence Agency, the US Navy, Procter & Gamble, Bupa and LexisNexis, and that it had raised $6 million in funding.

In May 2010 Vivísimo sold Clusty to Yippy, Inc., formerly Cinnabar Ventures Inc., for $5.55 million, a deal that included a license to Velocity, per Wikipedia citing the Florida Technology Journal. The exact date Clusty became Yippy is unconfirmed. Yippy, Inc. traded over the counter, and in 2016 it marketed a Yippy Search Appliance as a replacement for the Google Search Appliance. A 2019 homepage claim that Yippy searches were "powered by IBM Watson" is flagged on Wikipedia as needing a citation and is not confirmed here.

Vivísimo itself followed a separate path. IBM announced it would acquire the company on 25 April 2012, completed the deal in May 2012 and turned the software into IBM Watson Explorer. Terms were not disclosed. The clustering lineage therefore ends inside an IBM enterprise product, not in consumer search.

How Yippy ended

Yippy's end as a search engine was gradual and undramatic. From late April 2021 to early June 2022, yippy.com redirected visitors to DuckDuckGo. A search brand that sends its users to another engine has stopped being a search engine in any practical sense, and late April 2021 is the most defensible end date for the Clusty and Yippy line.

The corporate record points the same way. Georgia's Secretary of State listed Yippy, Inc.'s status as "Revoked" as of 1 August 2021, per Wikipedia. The company's SEC filings, which would be the best primary source for the 2010 purchase terms, were not located for this page.

As of 27 October 2026, yippy.com shows fragments of unrelated AI and technology content and no search box. It should be treated as a repurposed or parked domain with no connection to the engine described here.

What clustering got right, and why it lost

Clustering solved a real problem, and the problem has not gone away. Ambiguous queries still return mixed results, and a reader still has to scan past meanings they did not intend. Clusty's folders made that sorting visible and put the reader in charge of it.

It also had structural weaknesses that this page should state plainly. As a metasearch engine, Clusty could only cluster what its sources returned; it could not reach a page none of them ranked, and it depended entirely on engines it did not control. Its folder labels were generated from snippets, so their quality varied with the query. And a tree of folders asks the user to do more work than a single list.

Google and Bing kept the single ranked list and added refinements to it. Clustering search became the road not taken in mainstream web search, surviving mainly inside enterprise software such as the IBM product Vivísimo's technology became. For anyone looking for a working clustering web search engine today, Clusty and Yippy are not options.

Frequently asked questions

What happened to the Yippy search engine?

Yippy stopped working as a search engine in 2021. From late April 2021 to early June 2022 yippy.com redirected to DuckDuckGo, and Georgia's Secretary of State listed Yippy, Inc. as "Revoked" from 1 August 2021. As checked on 27 October 2026, yippy.com shows unrelated AI and technology content with no search function.

Is Clusty still available?

No. Clusty was sold to Yippy, Inc. in May 2010 and became Yippy, which itself ceased operating as a search engine in 2021. A request to clusty.com from this site's research environment failed on 27 October 2026, so the domain's current state is unconfirmed, but no Clusty search service is known to exist.

Was Clusty the same as Yippy?

Yes, in lineage. Vivísimo released Clusty in 2004 and sold it to Yippy, Inc., formerly Cinnabar Ventures, in May 2010 for $5.55 million, including a license to Vivísimo's Velocity software, according to Wikipedia. The engine continued under the Yippy name. Vivísimo the company was separately bought by IBM in 2012.

Did Clusty have its own index?

No. Clusty was a metasearch engine: it sent queries to other search engines, merged what they returned and then clustered those results into labeled topic folders. It never crawled the web itself. Which engines it drew on is not documented in any dated primary source found for this page, so no list of source engines is given here.

What is a clustering search engine?

A clustering search engine groups results by topic instead of showing one ranked list. Clusty read the titles and snippets of returned results, found recurring phrases and built folders such as car, animal and software for a query like jaguar. The folders were generated fresh for each query, unlike a directory's fixed categories chosen in advance by editors.

Who made Clusty?

Vivísimo, Inc., a Pittsburgh company founded in 2000 by Carnegie Mellon University computer scientists Raul Valdes-Perez, Jerome Pesenti and Chris Palmer. Its business was enterprise search software, and Clusty served as a public showcase. IBM announced its acquisition of Vivísimo on 25 April 2012, and the software became IBM Watson Explorer.

Clusty and Yippy search engine