SearXNG
FreeAggregates dozens of search engines, without logging you or building a profile
An open-source metasearch engine that aggregates results from dozens of sources — Google, Bing, Wikipedia, GitHub and more — onto one page with no injected ads, no query logging and no user profiling. Use a public instance someone else runs, or bring up your own with Docker.
Why we picked it
AGPL-3.0, 36,700 stars.
A search engine's business model requires it to know you: log the queries, assemble a profile, sell the ad slot. SearXNG removes that middle layer — it forwards your request to each upstream engine and aggregates the results, so the engines see the instance rather than you, while the instance itself keeps no query logs, builds no profile and injects no ads.
Two ways to use it. Take a public instance — searx.space lists a set of them, open one and search, no account. Or run your own; official container deployment makes a small VPS or a home NAS enough.
Result quality depends on which sources you enable, and the configuration is per-category: code searches through GitHub and Stack Overflow, papers through arXiv and Google Scholar, each toggled independently.
To be clear about:
- Public instances vary a lot. Some get rate-limited upstream and start returning empty results or captchas after a few searches; some are maintained loosely and lag behind. For sustained use, self-hosting is far steadier.
- Self-hosting hits rate limits too. A residential IP or small VPS being flagged as a bot by Google is routine, though you can weight the configuration toward sources that do not block.
- It is not an independent crawler. Anything the upstream engines cannot find, it cannot find either.