Browsing

Search Engines

Search queries are among the most sensitive data you produce. The alternatives, their sources, and their business models.

On this page
  1. What a search engine can log
  2. The alternatives
  3. Setup that makes it stick
  4. Logged-out Google is not private

Search history is a diary you didn't know you were keeping: health worries, legal questions, relationship trouble, employer names, travel plans. Logged against your account or IP address, it is one of the most complete behavioral records that exists about you. Moving it away from your ad profile is a one-minute change.

What a search engine can log

Query text, timestamp, IP address, browser fingerprint, which results you clicked, and (if you're signed in) your identity, joined against everything else that account does. Google's own account controls page is a candid inventory of what gets retained. The default is: everything, indefinitely tied to you, unless you configure otherwise.

The alternatives

Most private search engines are frontends: they buy results from Bing, Google, or Brave's index and interpose themselves so the upstream provider sees aggregate queries, not you.

Engine Results from Model Notes
DuckDuckGo Bing + own crawler Non-tracking ads The mainstream option; no query logs tied to users
Startpage Google Non-tracking ads Google results without a Google login; owned by System1, an ad-tech firm, its no-logging policy has held up but the ownership is worth knowing
Brave Search Own index Ads / paid tier Independent index, no upstream dependency
Mojeek Own index Ads Fully independent crawler; results are weaker, transparency is excellent
SearXNG Aggregates many Self-hosted / public instances Open-source metasearch; you trust the instance operator instead

"Private" here means the engine's stated policy plus its business incentives. None of these can prove a negative about their logging; what you're choosing is a provider whose revenue does not depend on profiling you, which is a structural improvement over one whose revenue does.

Setup that makes it stick

  1. Change the default engine in your browser settings, on desktop and phone, which everyone forgets.
  2. Disable "search suggestions" from the network so keystrokes aren't streamed to the provider as you type (see Browser Hardening).
  3. Expect an adjustment period. For most queries the results are equivalent; for the occasional local or long-tail query where they aren't, using Google deliberately for one search is still a privacy win over using it for all of them.

Logged-out Google is not private

Searching Google without an account still ties queries to your IP address and fingerprint, and consent banners notwithstanding, still feeds the profile attached to your browser. If you need Google's results, Startpage gets you most of them without the profiling.