Every time you search on Google, Google logs your query, your IP address (the number that identifies your internet connection and lets websites infer your location and build a behavioral profile over time), and, if you are signed in, your account identity. That data feeds the ad targeting system. Whoogle is a self-hosted search proxy that inserts itself between you and Google: you search Whoogle, Whoogle fetches the results from Google using its own server address, strips everything Google added for tracking purposes, and returns plain results to your browser. Google sees a request from Whoogle's server. It does not see your IP or your account. The project lives at github.com/benbusby/whoogle-search and has 11,595 stars.
| Fact | Value |
|---|---|
| What it is | Privacy-preserving Google search proxy |
| Built by | Ben Busby and contributors (open-source) |
| License | MIT -- free to use |
| Platforms | Docker, Linux, Heroku, Railway |
| Requires | Docker or a Python web server |
| Install method | Docker or one-click Railway/Heroku deploy |
| Verdict | Install if you want Google search quality without Google logging your searches |
What Whoogle actually is
Whoogle is a proxy for Google Search. A proxy, in this context, is an intermediary server that makes requests on your behalf. Instead of your browser connecting directly to google.com, your browser connects to your Whoogle instance, which then connects to google.com. Google sees the Whoogle server's IP address and network identity, not yours.
This distinction matters because the thing Google logs is the combination of your IP address, your browser fingerprint (the unique signature created by your browser's settings, installed fonts, screen size, and dozens of other details), and your signed-in account, if any. Break that link and Google cannot build a profile of what you searched for today, last week, or across a period of years.
Whoogle is not a search engine. It has no index, no crawler, and no ranking algorithm of its own. When you search for something, Whoogle passes that query to Google and brings back Google's results. The search quality you get is Google's quality. The tracking is not.
What you get after using it
The interface is a plain search box. Results look close to Google's organic results, without the visual clutter of shopping boxes, sponsored results, and knowledge panels that surface Google's own products. Tabs for Images and News work. Wikipedia snippets appear. Clicking any result takes you directly to the destination website.
What disappears: sponsored results (the paid ads that appear above and below organic results), AMP pages (AMP, Accelerated Mobile Pages, is Google's mobile page format that loads content through Google's own servers and keeps your browsing behavior on Google's infrastructure), tracking pixels (small invisible images or scripts embedded in pages that report your activity back to an advertising network), and the JavaScript code Google uses to track which results you click and how long you stay.
What does not disappear: the actual search results. Whoogle strips the monitoring layer while leaving the content intact.
One important limitation: Whoogle is not anonymous from your own internet service provider, which still sees that your device connected to your Whoogle server. It shields you from Google specifically. It also depends on Google remaining accessible from wherever your Whoogle server is hosted. If Google detects unusual traffic patterns from a server sending many automated-looking queries, it may present CAPTCHAs (the "prove you're human" challenges) that interrupt results. Public Whoogle instances that many people share hit this more often than a private instance used by one household.
The install experience
Docker is the fastest install path. Docker is a tool that runs software in a self-contained environment called a container, so you do not need to install Whoogle's underlying dependencies manually. A single command starts Whoogle:
docker run --publish 5000:5000 benbusby/whoogle-search
After that, Whoogle is running at http://localhost:5000 on your machine. "Localhost" means "this machine" -- it is a shorthand address that only works from the same computer the server is running on. Set that address as your browser's default search engine (browsers let you add a custom search engine with a URL pattern; for Whoogle that is http://localhost:5000/search?q=%s, where %s is a placeholder your browser replaces with whatever you type) and you are done. Every search goes through Whoogle from that point forward.
If you want Whoogle accessible from any device on your network, run it on a machine that stays on (a home server, a Raspberry Pi, or a cheap VPS -- a Virtual Private Server, meaning a rented cloud machine you control) rather than your laptop, and use that machine's local IP address instead of localhost. A Raspberry Pi 3 or 4 handles Whoogle comfortably alongside other services.
Railway and Heroku both offer one-click deploys, which means you click a button on the Whoogle GitHub page, authorize Railway or Heroku to create a project from the repository, and get a hosted URL in a few minutes. This option costs nothing on Railway's free tier for light personal use and requires zero command-line work. The tradeoff is that your Whoogle instance now lives on a third-party server rather than hardware you control, though it still shields you from Google.
There is no database to configure, no account to create, and no configuration file required for a basic install. The setup time with Docker is under five minutes including the time to verify it works.
Where it fits and what to compare it to
DuckDuckGo is the most common starting point for people who want Google-quality results without Google tracking. It does not require self-hosting, has its own index supplemented by Bing results, and has a respectable privacy policy. Its results are good for most queries and excellent for informational searches. The gap opens up on niche technical searches and local queries, where Google's index depth shows. If DuckDuckGo already satisfies you, Whoogle adds no clear benefit. If you have returned to Google repeatedly because DuckDuckGo could not find what you needed, Whoogle gives you Google's results without Google's tracking.
Brave Search uses an independent index with no Google results. It is the choice for people who want to avoid both Google's tracking and Google's index. Search quality is improving but not yet at Google's level across all query types.
Startpage is a commercial service that functions similarly to Whoogle: it fetches Google results on your behalf and strips tracking. It is hosted, free to use, and requires no technical setup. The difference from Whoogle is that Startpage is a company with a privacy policy and business model you have to trust, whereas Whoogle is open-source software you run yourself. If self-hosting is not appealing, Startpage is the hosted equivalent.
SearXNG queries multiple search engines simultaneously (Google, Bing, DuckDuckGo, and others) and aggregates the results. It is more privacy-focused and more configurable than Whoogle, but also more complex to set up and maintain. A public instance like search.sethforprivacy.com lets you try it without installing anything.
The decision breaks into two questions. Do you want Google's results specifically, or just good results? And are you willing to run a server, or do you want a hosted option? Whoogle answers the first question yes and requires a yes to the second. Startpage answers the first yes and does not require the second. SearXNG answers the first with "multiple engines" and requires a technical setup similar to Whoogle.
Verdict
Install Whoogle if you rely on Google's search quality for technical or specialized queries and want to stop feeding Google a searchable record of your interests. The Docker install is genuinely simple, the results are Google's results, and you retain the search habits and shortcuts you already have.
Hold off if you search mainly for informational topics where DuckDuckGo or Brave Search are adequate. Those options require zero setup and zero maintenance. The overhead of running a server is only worth taking on if the Google result quality gap is real for you in practice.
If you want the privacy properties of Whoogle without running any infrastructure, try Startpage first. You get Google results with tracking removed, hosted by a company with a commercial privacy commitment. If you later decide you want full control over the proxy itself, Whoogle is the self-hosted path to the same outcome.
For anyone who has already stood up AdGuard Home, Nextcloud, or another self-hosted service: Whoogle fits naturally into the same setup. It is a lightweight container, it runs reliably on the same hardware, and it plugs a data collection channel that DNS-level blocking cannot address. DNS blocking (intercepting domain lookups at the network level, as tools like AdGuard Home do) stops trackers from loading, but it does not see the content of your search queries. Whoogle does.