MetaCrawler

By Bill Hartzer · Updated 19 August 2026

What MetaCrawler is, and where its results come from

MetaCrawler is a metasearch engine: it sends your query to other search services and returns a merged list of what they send back. It does not crawl. It does not hold an index — an engine's own stored copy of the web, built by its own crawler. There is no MetaCrawler bot and no MetaCrawler user-agent, because there is nothing for one to fetch.

This is not decline; it is design. The 1995 paper that introduced MetaCrawler describes the system as one that "posts the query to multiple search services in parallel, collates the returned references" — the explicit alternative to maintaining an index. The name is the misleading part. "MetaCrawler" does not mean it crawls; it means it works one level above the crawlers, on their output.

What the site is today is a different question from what it was. MetaCrawler is now owned by InfoSpace Holdings LLC, a subsidiary of System1, Inc., an advertising company based in Marina del Rey, California. You get a results page assembled from third-party feeds and paid listings, with tabs for web, images, video and news. It is a sibling of Dogpile and WebCrawler: same owner, same underlying infrastructure, different skin. The research instrument became a monetised legacy brand, and the two should not be confused with each other.

The 1995 research project that invented metasearch

MetaCrawler is the founding academic work of metasearch, and unusually for a 1990s search engine there is a primary source: Erik Selberg and Oren Etzioni's paper "Multi-Service Search and Comparison Using the MetaCrawler," presented at the WWW4 conference in 1995. Development began in 1994 at the University of Washington as Selberg's PhD qualifying project, supervised by Professor Etzioni. It launched publicly on 7 July 1995.

The problem it solved was real and specific. In 1995 no single engine covered much of the web, each had its own incompatible query syntax, and a large share of the links any of them returned were already dead. MetaCrawler's answer had three parts, and the third is the one modern metasearch dropped:

Parallel dispatch — fire the query at several services at once. At launch it queried six: Galaxy, InfoSeek, Lycos, Open Text, WebCrawler and Yahoo.
Collation — merge the references and remove duplicates, presenting one ranked list instead of six.
Verification — actually fetch each returned URL to confirm it still existed, and use what came back to score it.

The concrete numbers from the paper are worth keeping because they are primary and they are small in a way that makes the period legible: between 7 July and 30 September 1995 MetaCrawler processed 50,878 completed queries, growing to over 7,000 queries per week. The entire system was 3,985 lines of C++. Those are launch-period research figures and say nothing about any later scale.

The template MetaCrawler set — parallel dispatch, collation, deduplication — is the one every later metasearch engine follows, from Dogpile and Mamma to Ixquick and today's SearXNG.

Whose results you are seeing now

MetaCrawler's own About page still makes a specific claim: that it searches Google and Yahoo! Search for web results, Yahoo! Search for images, Yahoo! Search and Truveo for video, and Yahoo! News for news.

Treat that as what MetaCrawler says, not as what MetaCrawler does. Truveo was AOL's video search service and was shut down years ago. The same page opens with an unreplaced internal template string — "MetacrawlerSemElevatedMedia has been created to give you the best search results…" — which is a leaked variable name sitting in production copy. A live About page carrying a dead partner's name and a piece of unrendered template is a good measure of how much maintenance the property receives.

The actual current feeds are not documented. System1 does not disclose syndication contracts by property. The safe statement is the structural one: no index of its own, results sourced externally, mixed with paid listings.

Over its commercial life MetaCrawler has at various times blended results from Google, Yahoo!, Bing, Ask.com, About.com, MIVA and LookSmart. The last two in that list are not organic engines at all — they were paid-listing networks. For much of the 2000s, a meaningful share of what MetaCrawler displayed was sponsored inventory, which is worth knowing before treating any historical MetaCrawler results page as a picture of the web.

Who owns it — and the years it did not exist

MetaCrawler has had six owners and one disappearance.

1994–1996 — University of Washington research project.
1996 — commercialised through NetBot, Inc., a spinout co-founded by Etzioni.
January 1997 — MetaCrawler is sold to Go2Net, Inc.; Paul Allen takes a roughly 54% stake in Go2Net. Go2Net IPOs on Nasdaq in April 1997.
October 1997 — Excite acquires NetBot for around $35 million. This is a different transaction from the one above: Excite bought the company, months after the MetaCrawler asset had already gone to Go2Net. Secondary sources merge the two constantly.
August 1999 — Go2Net acquires Dogpile, putting both leading metasearch brands under one roof.
July 2000 — InfoSpace, Inc. acquires Go2Net for a reported $4.2 billion in stock, at the peak of the dot-com market.
June 2012 — InfoSpace, Inc. renames itself Blucora; search becomes a segment held in an InfoSpace subsidiary.
July 2016 — Blucora sells InfoSpace, including MetaCrawler, Dogpile and WebCrawler, to OpenMail LLC for $45 million cash.
2017 — OpenMail renames itself System1. It lists on the NYSE as SST in January 2022 via the Trebia Acquisition Corp. merger.

The gap most write-ups miss sits in the middle of that. From around 2014 MetaCrawler was folded into Zoo.com, another InfoSpace property; metacrawler.com stopped serving its own search and was redirected away, later to Excite's search page. For a period there was no MetaCrawler product at all. It was relaunched as a standalone search site in 2017 under System1. So "MetaCrawler is dead" and "MetaCrawler has run continuously since 1995" are both wrong; the truth is a retirement and a revival, and the revival is a brand, not a return of the original system.

How it makes money

Advertising. Paid search listings and display advertising on the results page, monetised through System1's advertising and search-syndication arrangements. There is no subscription and no consumer API.

The paid element is not new. As noted above, MetaCrawler carried explicit paid-inclusion inventory from MIVA and LookSmart during the 2000s, which means sponsored results have been part of the product for most of its commercial life rather than an addition at the end of it.

System1's own description of its business is buying traffic and monetising it across an owned-and-operated portfolio of search and content sites. That framing is useful for a reader because it names the thing being optimised: the yield of the results page. An engine with its own index has a reason to publish facts about that index. A property monetising bought feeds has no equivalent incentive, and MetaCrawler publishes neither its suppliers nor its paid-to-organic ratio.

Privacy and data handling

MetaCrawler makes no privacy claim and is not marketed as a privacy tool. It publishes no search-specific privacy policy and no no-logging commitment; users are governed by System1's general corporate privacy policy, under United States jurisdiction and United States legal process.

There is also a point here that applies to metasearch in general and is worth stating once, plainly: a metasearch engine transmits your query to third parties by definition. That is the mechanism. It cannot offer end-to-end query privacy unless the operator proxies the request and strips identifying information — which some engines do and say so. MetaCrawler makes no such claim, so no such assumption is warranted.

For the record, MetaCrawler was named alongside Dogpile, WebCrawler and Startpage in November 2019 commentary about data flowing into System1's advertising infrastructure. That is advocacy writing rather than a regulator's finding and should be weighted accordingly; the corporate facts beneath it — common System1 ownership, and System1's Privacy One Group acquiring a majority stake in Startpage in October 2019 — are documented and undisputed. No regulatory enforcement action against MetaCrawler was found.

Who it suits, and who it does not

What it was genuinely good at. In 1995, exactly what it claimed. Parallel querying with link verification was a real solution to a real problem at a moment when each engine indexed only a slice of the web and returned a lot of dead links. It is also one of the few search engines of that era with a proper academic paper behind it, which makes it unusually well documented for anyone studying search history.

What it is bad at. Being a 2026 search engine. The problem it solved no longer exists: modern index coverage and freshness make "combine six partial indexes" unnecessary, and the link-verification step that made the design clever is gone. What remains has no index, no disclosure of current sources, no privacy posture and no AI layer, operated by a company whose incentive is the monetisation of the results page. The name actively misleads people into thinking it crawls.

Who should look elsewhere. Anyone who wants an index independent of Google and Bing should look at Mojeek or Brave. Anyone who wants privacy should look at engines that make an explicit, documented commitment. Anyone who needs to cite where a result came from cannot use an engine that will not say. And anyone who specifically wants modern metasearch done honestly should look at a public or self-hosted SearXNG instance, which lists the engines it queries and whose code can be inspected. MetaCrawler in 2026 is a brand; SearXNG is the idea.

Common misconceptions

"MetaCrawler crawls the web." It never has, by design. The "crawler" in the name refers to working across other crawlers' results, and the original paper is explicit about it.

"Excite bought MetaCrawler in 1997." Excite bought NetBot in October 1997. MetaCrawler itself had been sold to Go2Net in January 1997. Two transactions, routinely merged into one.

"MetaCrawler and Dogpile are rivals." Same owner since Go2Net bought Dogpile in 1999 and InfoSpace bought Go2Net in 2000. Both are System1 properties today.

"It was shut down." More precisely: effectively retired around 2014, when it was folded into Zoo.com and its domain redirected to Excite, then relaunched as a standalone site in 2017. Neither "dead" nor "continuously running" is accurate.

"Metasearch returns more results." It returns a merged, deduplicated top slice of several engines — usually fewer results than any single one of them would have shown you.

"It is a privacy option." Metasearch necessarily forwards your query to third parties, and MetaCrawler makes no privacy claim at all.

Frequently asked questions

Does MetaCrawler have its own crawler?

No, and it never has. MetaCrawler was built in 1995 specifically to avoid maintaining an index: it dispatches your query to other search services in parallel and merges the references they return. There is no MetaCrawler bot and no MetaCrawler user-agent. The name refers to working above other crawlers, on their results, rather than fetching pages itself.

Who invented metasearch?

Erik Selberg and Oren Etzioni at the University of Washington, with MetaCrawler. Development started in 1994 as Selberg's PhD qualifying project and the service launched publicly on 7 July 1995, querying Galaxy, InfoSeek, Lycos, Open Text, WebCrawler and Yahoo. Their WWW4 paper is the founding document of the field, and it describes the parallel-dispatch and collation model every later metasearch engine uses.

Who owns MetaCrawler now?

InfoSpace Holdings LLC, a subsidiary of System1, Inc., an advertising company in Marina del Rey, California. The chain runs from the University of Washington to NetBot in 1996, Go2Net in January 1997, InfoSpace in July 2000 for a reported $4.2 billion, Blucora after the June 2012 rename, OpenMail in July 2016 for $45 million, and System1 from 2017.

Which search engines does MetaCrawler use?

It does not reliably say. Its About page claims Google and Yahoo! Search for web results, but that page also names Truveo, a video service shut down years ago, and contains an unreplaced internal template string, so it is stale legacy copy rather than a current disclosure. System1 does not publish syndication contracts by property. What is certain is that there is no MetaCrawler index.

Is MetaCrawler still online?

Yes, as a System1-owned search site, though not continuously. Around 2014 MetaCrawler was folded into Zoo.com and its domain redirected away, eventually to Excite's search page, so for a period the product did not exist. It was relaunched as a standalone search site in 2017 under System1 and was resolving as of 19 August 2026.

Is MetaCrawler the same as Dogpile?

They are separate brands with the same owner and the same business model. Go2Net acquired Dogpile in August 1999, having bought MetaCrawler in January 1997, and InfoSpace acquired Go2Net in 2000; both are System1 properties today, alongside WebCrawler. The interfaces differ. The ownership, the advertising model and the absence of any index do not.

Is MetaCrawler good for privacy?

No. It makes no privacy claim, publishes no search-specific policy, and is governed by the general corporate policy of an advertising company under United States jurisdiction. There is also a structural limit: any metasearch engine forwards your query to third-party engines in order to work at all, so privacy depends entirely on whether the operator proxies and strips that request. MetaCrawler does not claim to.