SearXNG: Private Metasearch Without the Google Profile
🛠️ SearXNG
| What it is | Privacy-focused metasearch: one query, many backends, no user profile |
| Platform | Web (any browser); self-host on Linux, Docker, or a VPS |
| Price | Free, open source (AGPL-3.0) |
| Link | GitHub · Docs · Public instances |
Google and Bing are fine until you remember they remember you. Every query feeds a profile; every profile feeds ads. SearXNG sits in the middle instead: you type once, it fans the query out to many search engines (Google, Bing, DuckDuckGo, Qwant, Brave, Wikipedia, and dozens more), then shows you a merged result page. The instance you use sees your IP, but SearXNG itself does not build a search history or sell attention. For most people the fastest path is not installing anything—open searx.space, pick an online instance, and search.
- Start in under a minute. searx.space lists public SearXNG instances with uptime, country, TLS grade, and response time. Click a green row, bookmark it, set it as your browser’s default search engine. No account, no extension required.
- One box, many engines. SearXNG is a metasearch engine: backends are configurable per instance. Images, news, science papers, and map tiles can each route to different providers. You get breadth without opening six tabs.
- Privacy is architectural, not marketing. Queries go through the instance operator’s server, not Google’s ad stack. SearXNG does not profile users; you can also run it over Tor. The tradeoff: you trust whoever runs the instance you pick.
- Public instances are a convenience, not a guarantee. Heavy shared instances get rate-limited or CAPTCHA-blocked by upstream engines more often. searx.space warns about this openly. For daily driver quality, self-host or use a small private instance.
- Self-host when you care about consistency. Official Docker setup or the install guide gets you a URL only you use. Configuration lives in YAML; you choose which engines stay enabled.
Use searx.space to hop on now
You do not need a repo clone to try SearXNG today.
- Open searx.space — maintained by the SearXNG project, refreshed daily (response times every few hours).
- Scan the online instances table. Useful columns: TLS and CSP grades, country, uptime, search response time. Toggle Hide networks with privacy issue if you want to skip Cloudflare-proxied or analytics-heavy hosts.
- Click an instance URL. You land on a normal search page—type a query and hit enter.
- Bookmark the instance or add it as a custom search engine in Firefox (
Settings → Search → Add). In Chrome:Settings → Search engine → Manage → Add.
Cannot pick one instance? searx.space lists meta-instances that redirect to a random healthy server (e.g. Neocities or Gimmeasearx). Handy for a first look; less ideal as a permanent default because the backend changes.
For programmatic use, searx.space exposes instances.json (format may change—check the site).
Self-host (optional)
If a public instance feels slow or returns empty Google rows, run your own:
git clone https://github.com/searxng/searxng-docker.git
cd searxng-docker
docker compose up -dDefault install docs and engine tuning: docs.searxng.org. Expect to babysit engine blocks and the limiter if your instance is open to the world.
Worth your time if: you want web search without building a Google-shaped dossier of your interests, and you are willing to trade a slightly rougher result page for that. Skip it if you need Google’s one-box answers, signed-in Gmail integration, or flawless image search every time—metasearch will frustrate you on those edges.
Related TMFNK Content
- taken.: See Everything Your Browser Tells Websites About You Search is one leak; taken shows the rest of what your browser broadcasts.
- Best Free Tools That Don’t Require Signups or Show Ads SearXNG fits the same no-account, no-feed pattern.
- Project NOMAD: The Offline Knowledge and AI Server Another stack for keeping research and queries off third-party clouds.
Crepi il lupo! 🐺