Whoogle

Whoogle is *Google's results without Google watching you* — a self-hostable open-source search proxy that fetches Google results, strips the ads, tracking, AMP redirects, and JavaScript, and serves you the clean answer without ever tying the query to you.

Background. Created in 2020 (by Ben Busby) and open-source on GitHub, Whoogle is a lightweight *Flask app* you run yourself — on a $4 VPS, a home server, or a local Docker container. It proxies Google's index, so you *inherit Google's ranking quality* while interposing your own server between you and Google. That self-host-first design plus an open codebase is why it earns an A in /search: the privacy comes from architecture you control, not a promise from a middleman.

What you trust. Ideally, *yourself*. Because Whoogle is *open source and self-hostable*, the privacy-maximal configuration is to run your own instance: then *your* server makes the request to Google, so Google sees the instance's IP and no per-user profile, and Whoogle itself logs nothing about you because you operate it. It *strips tracking, ads, AMP, and JavaScript* from results, removing the surveillance and manipulation layers Google bundles in. Public instances exist for convenience, but the threat model explicitly *favors self-hosting* — on a public instance you're trusting that operator (and sharing an IP with other users, which is its own mixed bag). The trust, done right, collapses to "your own box."

Operational specs. A small *Flask* application, deployable via *Docker* or directly, that you point your browser at (and can set as your default search engine / via the OpenSearch hook). It returns Google-quality results with the cruft removed, supports configuration of result regions and filters, and runs comfortably on minimal hardware (a $4 VPS is plenty). *Public instances* are listed for those who won't self-host, but the recommended posture is your own private instance. Free and open-source; your only cost is the small server you run it on.

Philosophy. Search quality and search privacy are usually presented as a tradeoff: use Google and be profiled, or use a privacy engine with weaker results. Whoogle's premise rejects the dichotomy — Google's *ranking* is genuinely good, so the move is to *keep the results and remove the surveillance* by self-hosting a proxy. Open-sourcing it and making it trivially self-hostable is the whole point: a privacy tool you run yourself, whose code you can read, leaves no operator to trust. It's the SearXNG-adjacent philosophy applied specifically to keeping Google's index usable without Google's tracking.

Grade rationale. A in /search. The grade reflects open-source code, genuine self-hostability (the privacy-maximal config), removal of tracking/ads/AMP/JS, Google-grade result quality, and a tiny resource footprint. It sits with the other privacy-search approaches (Mullvad Leta as a managed proxy, Brave as an independent index, SearXNG as a meta-search) — Whoogle's niche is *self-hosted Google passthrough*. The caveats are about the self-host requirement and Google-dependency, not code integrity.

Useful when. Use Whoogle when you specifically want *Google's result quality* but refuse to feed Google your queries and IP — and you're willing to *self-host* (a few minutes with Docker on any cheap box). It's the right pick for the privacy-minded who find independent indexes' results lacking on long-tail queries and want the real thing, de-surveilled. Set your self-hosted instance as your browser default and you have private Google for daily use.

Caveats. The privacy is *only real if you self-host* — on a *public instance* you're trusting an unknown operator with your queries and sharing its IP, which can be rate-limited or blocked by Google, so don't treat random public instances as private. It *depends on Google's interface*: scraping/proxying can break when Google changes things or rate-limits an instance's IP, so it needs occasional maintenance and isn't as set-and-forget as a hosted product. Running it well is a (small) sysadmin task. And it inherits Google's *ranking biases* — it removes the tracking, not the editorial slant of the index. None of these undercut the A: for self-hosted, de-tracked Google-quality search, Whoogle is the canonical tool — the asterisk is simply "run your own instance."