Search engines other than Google

First decide what you are actually trying to change

People leave Google for four quite different reasons, and each has a different answer. Confusing them is why so many people switch, feel underwhelmed, and switch back within a fortnight.

• You object to being tracked. You want mainstream-quality results without a profile being built. The answer is a privacy front end over Google or Bing.
• You object to the results page itself — the ads, the AI summary, the Google-owned modules pushing links below the fold. The answer is a cleaner interface, which may still be Google's or Bing's results underneath.
• You want different results — genuinely different pages, from an index Google does not control. That is a much shorter list, and it involves real trade-offs.
• You are not in the anglosphere. In several large markets the local engine is not an alternative at all; it is the mainstream, and Google is the challenger.

The distinction that runs underneath all four is whether an engine has an index of its own — its own crawled copy of the web — or buys finished results from a company that does. An engine with no index cannot show you a page its supplier never crawled. It can change everything else: what is logged, what is filtered, what the page looks like. It cannot change coverage.

So the sections below are organised by that fact rather than by a ranking. There is no “best” answer here, and any list that gives you one without telling you where the results come from is not describing search engines, only websites. A reasonable first question is how many search engines exist at all, and the answer turns out to depend entirely on whether you are counting indexes or interfaces.

Engines that run a genuinely different index

Worldwide, only about a dozen organisations crawl the open web at scale. These are the ones a Western reader can realistically use day to day.

Microsoft Bing. The obvious one that people skip because it is not exotic. It is the second full independent index of the Western web, it is the default in Windows and Edge, and it is the wholesale supplier behind Yahoo, DuckDuckGo, Swisscows and others. Per StatCounter for July 2026 it held 4.47% worldwide across all devices but 8.84% on desktop worldwide and 8.63% in the United States — the desktop figure is the one that reflects how people actually use it.

Brave Search. The most credible independent alternative for general daily use. It launched in June 2021 with roughly 87% of results from its own index and announced the removal of its last Bing API calls on 27 April 2023, adding its own image and video indexes that August. Because it controls the index, it can offer Goggles — user-written re-ranking filters that boost, demote or exclude whole sets of sites, which is something no reseller can build. Its AI summary can be switched off.

Mojeek. A small British company that has never resold anyone's results at any point in its history, which makes it the cleanest example on this site of what independence looks like. Its own About page gave 9 billion pages for 2025. That is a real web index and it is also one to two orders of magnitude smaller than Google's, so it will not have the page you need on a good share of specific queries. Used as a second engine — the one you check when you want to see the web without Google's hand on it — it is excellent. Used as a Google replacement it will disappoint, and saying otherwise sets readers up to fail.

Marginalia Search. Built and run by one engineer in Sweden, with around 300 million documents as of 2024, and deliberately ranked backwards from commercial engines: heavy JavaScript, dense advertising, affiliate links and tracking scripts push a page down, and age is treated as a positive signal rather than staleness. What surfaces is personal sites, old blogs, university pages and unmonetised documentation. Its own description of itself is “a minority report that keeps them honest.” It cannot do news, shopping, maps or local, and it does not pretend to.

Qwant and Ecosia sit in between. Both blend partner feeds with Staan, the European index built by the 50:50 joint venture they own together, which went live on 6–7 August 2025 starting in France. Neither is independent yet — Ecosia's own help centre names Bing, Google and the joint venture as its three providers — but they are the only consumer engines in Europe putting money into building an index rather than renting one.

Mainstream results, handled differently

If your objection is surveillance rather than coverage, these deliver Google- or Bing-grade results without the engine profiling you. They are not independent engines and should not be sold as such.

DuckDuckGo is the mainstream answer. It runs its own crawlers and its own indexes behind the Instant Answer boxes, but its ranked web links come from Bing, as its own sources page states. It stores no search history against an identity, does not personalise, and offers the clearest AI opt-out of any mainstream engine. StatCounter put it at 0.65% worldwide and 1.58% in the United States for July 2026 — figures that almost certainly understate it, because DuckDuckGo strips referrers by design and StatCounter counts referrers.

Startpage has no crawler at all. It relays your query from its own servers to Google and Bing without your IP or cookies and returns their results, plus Anonymous View, a proxy that loads the destination site through Startpage so the site does not see you. It was Google-only from 2009 until March 2023, when a Microsoft partnership added Bing results and Bing ads. Worth knowing before you choose it: the operating company is Dutch, but the majority owner since December 2018 has been System1, a US-listed advertising company — a fact not disclosed publicly until October 2019.

Ecosia spends the large majority of its profits on tree planting and landscape restoration, and its steward-ownership structure with the Purpose Foundation means it legally cannot be sold and cannot pay dividends. It is a climate engine with decent privacy hygiene, not a privacy engine, and conflating the two is the commonest error made about it.

Swisscows is Swiss and family-friendly by design; it states an exclusive Bing partnership on its own blog. Yahoo is worth naming only to be clear about what it is: it has not crawled the web since 2010, and its January 2026 AI answer engine, Yahoo Scout, is Bing's index with Yahoo's ranking over it.

Outside the English-speaking web

In four markets the local engine is not an alternative to Google; it is a peer or the incumbent, and each runs its own crawler.

• Yandex held 72.85% of Russian search in July 2026 per StatCounter, against Google's 25.48%. It is a genuinely strong index in Russian. It is also a Russian company under Russian data-localisation and lawful-interception law, and independent research has repeatedly found its results filtered in line with Russian legal designations without that filtering being disclosed on the page.
• Baidu is the Chinese incumbent. StatCounter put it at 50.83% of Chinese search in July 2026, but that panel is small and skewed — it also reports implausible figures for other engines in China — and domestic estimates run substantially higher. Results are filtered in line with Chinese law and users are not told what is missing.
• Naver is the case where measurement matters most. StatCounter gave it 40.9% of South Korean search in July 2026, behind Google; the Korean domestic panel InternetTrend gave it 63.8% for March 2026, ahead of Google at 28.7%. Both are defensible and they measure different things — a large share of Naver searches are answered inside Naver's own blogs and Q&A and never produce an outbound referral, which is exactly what StatCounter counts.
• Seznam is the Czech engine, and one of the very few independent indexes operated by an EU-headquartered company. It has been in a long, unreversed decline: a majority of the Czech market in 2000, roughly level with Google in 2012, 26% at the end of 2018, and 14.05% in July 2026 per StatCounter.
• Cốc Cốc in Vietnam runs a documented crawler fleet that deliberately limits itself to Vietnamese content — a real index, but a national one. StatCounter put it at 3.38% of Vietnamese search in July 2026 against Google's 95.95%, down from a desktop peak above 32% in 2017.

Answer engines instead of a results page

A growing share of people leaving Google are not switching to another list of links at all. Perplexity began as a Bing reseller with a summarisation layer and, by its own September 2025 engineering account, now runs an exabyte-scale crawl tracking over 200 billion unique URLs. ChatGPT Search runs OpenAI's own OAI-SearchBot and also buys results from third-party providers; the mix differs between tiers. Microsoft Copilot has no index of its own and grounds its answers by sending generated queries to Bing.

Kagi is the outlier worth understanding, because it changes the business model rather than the technology. It is subscription-funded and shows no advertising, which removes the structural incentive to fill a results page with paid units. It runs its own crawler into its own Teclis web index and TinyGem news index, and its documentation says the bulk of a results page comes from anonymised requests to third-party commercial indexes, Brave's among them. It also treats AI as something you invoke rather than something imposed above every result.

The honest caution about all of these: an answer engine gives you a paragraph instead of a page of sources, and the paragraph is only as good as the retrieval underneath it. If you need to know that something is true rather than that a model asserted it, you still need the links.

Metasearch, and running your own

SearXNG is not a website; it is free software you install. It forwards your query to a set of engines you choose, strips your identity from those requests, and merges what comes back. Self-hosted, its privacy guarantee is a property of your own machine rather than a promise from a company, which is a categorically stronger thing. There are two catches worth stating plainly. On a public instance run by an anonymous volunteer you have transferred your trust rather than eliminated it — nothing in the software stops an operator logging everything. And upstream engines actively block it: a 2023 incident saw the number of tracked public instances with working Google results collapse from 66 of 91 to 25 within days, and rate limiting remains an active problem. Its results can never be better than the engines it queries, and are often thinner because engines drop out mid-session.

The 1990s metasearch brands — Dogpile, WebCrawler, MetaCrawler — still resolve, and all three are now owned by System1, buying syndicated feeds. WebCrawler, named after the web's first full-text crawler, stopped crawling in 2001.

What you actually give up, and what has closed

Nobody switching should be surprised by this, so here it is directly. Leaving Google for an independent index costs you long-tail recall first — the assurance that if a page exists, you will find it. It costs you freshness, because a small crawl budget cannot recrawl news-velocity sites at Google's cadence. It costs you local and maps, which are Google's strongest verticals by a distance. It usually costs you non-English coverage. And increasingly it costs you the platform web: when Reddit blocked non-Google crawlers in 2024 while Google retained access under a paid deal, every small engine lost Reddit at once. That is a licensing barrier, not an engineering one, and no independent engine can solve it alone.

Several names that still appear on alternatives lists are no longer search engines at all. Ask.com closed on 1 May 2026; IAC discontinued its search business and the domain now serves a single farewell page. AltaVista was shut down by Yahoo on 8 July 2013. Neeva, the ad-free subscription engine with its own crawler, shut its consumer product on 2 June 2023 and was absorbed into Snowflake. You.com still exists as a company but has discontinued its consumer search box and now sells its index as an API. HotBot is a repurposed domain running an AI chat aggregator; Excite and Lycos are still online with search boxes fed by third parties, in Excite's case System1 by its own privacy policy's admission.

The practical conclusion most readers land on is not a single replacement but a pair: a mainstream engine for the queries where coverage decides the answer, and an independent one for the queries where you want to see what the mainstream is not showing you. That is a more honest recommendation than any ranked list, and it is what most people who successfully leave Google actually end up doing.

Frequently asked questions

Sources

• gs.statcounter.com/search-engine-market-share
• duckduckgo.com/duckduckgo-help-pages/results/sources
• brave.com/blog/search-independence/
• support.ecosia.org/article/579-search-results-providers
• mojeek.com/about/
• about.marginalia-search.com/article/about/
• gs.statcounter.com/search-engine-market-share/all/south-korea
• ppc.land/ask-com-closes-after-nearly-30-years-as-iac-exits-the-search-busine…