Search Engines That Are Not Google

There are plenty of Google alternatives and most articles about them list logos without explaining the important distinction: whether the engine crawls the web itself or repackages results from someone who does.

That difference decides everything about what you will and will not find.

The thing to understand first

Running a web index is enormously expensive. Only a handful of organisations do it at scale. Many “alternative” search engines are interfaces on top of someone else’s index, usually Bing’s, with privacy protections and a different design added on top.

That is a completely legitimate product — the privacy benefit is real — but it means the results will be broadly similar to the source. If your complaint is about the results rather than the tracking, you need an engine with its own crawler.

Independent indexes

Mojeek is a UK engine that crawls the web itself. Its index is far smaller than Google’s, which is exactly why it is interesting — the results are noticeably less commercial, and it will surface small sites that are buried elsewhere.

Marginalia Search deliberately ranks against the commercial web. It favours pages that are text-heavy, non-commercial and lightly optimised. For finding personal sites, hobbyist documentation and the handmade end of the web, nothing else comes close. It is genuinely bad at finding a shop, which is the point.

Brave Search runs its own index and is a reasonable general-purpose default.

Wiby indexes pages built in the older, simpler style of the web. It is closer to a curated directory than a search engine and it is a delight for the same reasons covered in where to find the old web.

Privacy front-ends

DuckDuckGo is the best-known, drawing on Bing among other sources alongside its own crawling, with no tracking profile. Startpage serves Google’s results through a proxy, so you get the familiar index without the profile. Ecosia is built largely on Bing and puts its advertising revenue into tree planting.

Pick one of these if your objection is to being tracked. Do not expect different results, because the index is largely the same.

Paid search

Kagi charges a subscription and shows no ads. The argument is straightforward: an engine paid by users optimises for users, and an engine paid by advertisers optimises for advertisers. It also lets you demote or block domains permanently, which is unexpectedly satisfying if there are content farms you never want to see again.

Whether it is worth money is a personal question. The experiment is at least honest about the trade.

How to actually use these

The mistake is picking one and switching wholesale. Search engines are tools with different shapes, and the sensible approach is to route by question type:

• Commercial, local or current — the big indexes, because they have the coverage.
• Personal sites, hobby documentation, old writing — Marginalia or Mojeek.
• Old-web aesthetics and small handmade pages — Wiby.
• A page you know exists but cannot surface — try two or three indexes, since coverage of small sites varies enormously.

Setting these up as address-bar keywords makes switching nearly free, which is the only way you will actually do it. That is covered in hidden browser features you are not using.

Whatever you use, use the operators

Exact phrases, site restrictions and date filters change results far more dramatically than changing engine does. Search operators that still work is the shorter path to better results, and it applies everywhere.

For how all this came to be, search engines before Google covers the eight years when the market looked completely different.