Private search and browsing: DuckDuckGo and beyond

01

What does private actually mean?

Start with what you want to keep private, and from whom. A search company, an advertising network, another person using your laptop and the operator of your internet connection can each see different parts of a browsing session.

• Search privacy: Whether your search provider stores queries, links them to you or builds a profile from them.
• Browsing privacy: What websites and the third parties they load can learn as you move between pages.
• Local privacy: What remains in your browser history, cookies, downloads or synced account.
• Network privacy: What your connection reveals, including the IP address websites receive.

A private search engine can improve the first layer. Browser protections help with the second. Private windows address parts of the third. Tools such as Tor address the fourth as well. They work together, but each has its own limits.

Read past the slogan: No personal search history, no advertising profile and no stored data are different claims. A service may keep anonymous counts, preferences or temporary technical logs without maintaining a personal history. A query can also identify you by its contents if you type a name, email address or other identifying detail.

02

Seven ways to search with more privacy

You can try these in your existing browser. You do not have to install the Brave browser to use Brave Search, or the DuckDuckGo browser to use DuckDuckGo Search.

DuckDuckGo: an easy everyday starting point

DuckDuckGo is our first suggestion for someone who wants a straightforward change. Its privacy policy says it does not save or share your search history or save IP addresses and unique identifiers alongside searches. It temporarily uses connection information to serve requests and deal with abuse, and uses anonymous methods to improve its products.

It also supports preferences stored on your device, and some optional services require information of their own. The useful promise is a search experience without a personal search history at DuckDuckGo. Your ordinary browser may still keep the results-page address in its own history.

Keep in mind: DuckDuckGo Search and DuckDuckGo's browser are separate products. Changing your search engine alone does not add tracker blocking to every page you visit.

Startpage: a search intermediary with an extra browsing option

Startpage strips identifying metadata before sending a query to its search and content providers. Its policy says it does not log your searches or record IP addresses for ordinary search. It describes exceptions for abuse prevention, preference and security-check cookies, and separate account or subscription data.

Its Anonymous View feature can open a result through a proxy, so the destination receives a connection from Startpage instead of directly from your device. Use that specific option when you want it; an ordinary result click does not automatically proxy your browsing.

Our take: A useful alternative to try if you want an intermediary for search and occasional proxy browsing. Some pages work poorly through proxies, and the proxy operator becomes part of the trust decision.

Brave Search: an independent index, with a clear data caveat

Brave Search uses its own web index. Its privacy notice says it stores de-identified search queries, including query text and contextual information such as country, language and date. It says these records are not linked to an individual, device or IP address, and are not grouped into a user's history.

This distinction matters if your main requirement is that the search words themselves must never be stored. Brave's approach aims to avoid personal profiling; it is not a promise of zero query storage. The notice also describes brief IP processing for service integrity and separate handling for its conversational AI feature.

Our take: Try it if an independent search index appeals to you. Review anonymous usage metrics in its settings; Brave documents a switch to disable them.

Swisscows: family filtering, with advertising to examine

Swisscows offers a family-friendly search environment that filters sexual and violent content. That can be useful for a household, though filtering also changes what you can find.

Its current privacy policy says it does not build personal profiles, but describes anonymized analytics and advertising-related data sharing. More significantly, it says embedded advertising from Microsoft Bing and Google can automatically transmit personal data to those providers when loaded, even without an ad click. Account features also collect personal information.

Our take: Consider it for its filtering, but examine the free service's advertising carefully. Swisscows offers an ad-free Pro option. We would not describe the free version as a search engine that collects or shares absolutely nothing.

MetaGer: a nonprofit service funded through search credit

MetaGer is run by the nonprofit SUMA-EV and combines results from several search engines. Its current access model uses a generated key topped up with paid credit, without requiring an email-based search account. Check its site for current payment options and prices.

The privacy policy says IP addresses and user-agent identifiers are not saved or shared for web search. Queries are sent to partners to obtain results, and results including the search term are held for a few hours for display. It also describes additional storage if optional address-bar suggestions are enabled. Payment data depends on the payment method.

Our take: Interesting if you value nonprofit operation and an ad-free model. Search credit is a practical difference from free alternatives. Its anonymous browsing feature is an additional service, rather than a property of every link opened in your usual browser.

Kagi: paying for search, with account and log details

Kagi is a paid search service. An account and subscription change the trade-off: you fund the search service directly, while account and payment information still need handling.

Read its current policy rather than stopping at a no-history claim. It says search queries can be logged temporarily for debugging without being linked to an account. The retention table lists seven days for load-balancer and virtual-machine logs and 90 days for sampled server-error records. Kagi also keeps usage volume for billing and uses external payment processors.

Our take: Worth trying if you are comfortable paying for search. Treat its account, billing and temporary debugging records as part of the decision. Kagi also documents Tor and Privacy Pass for people who want stronger separation between authentication and searching.

SearXNG: open-source software, and a choice of operator

SearXNG aggregates results from other search services. It is software people can host, rather than one company running one universal search website. The project is designed without user tracking or profiling; public instances are listed at searx.space.

On a public instance, the operator sits between you and the upstream engines. You still need to consider that operator's policies, configuration and server logs. Open-source code makes inspection possible, but cannot prove what every public server runs.

Our take: A good option for people who enjoy choosing or running their tools. Self-hosting gives more control, along with maintenance responsibilities. It does not automatically make you anonymous: a small private instance can send queries from a server address closely associated with you.

03

Which one should you try first?

For a small, everyday change, start with DuckDuckGo or Startpage. Try Brave Search alongside them if you want to compare an independent index. The other choices make more sense when a particular feature or funding model matters to you.

Try a few ordinary searches: a local place, a technical question and something obscure you already know well. Notice whether you can find useful results. A service you can comfortably use every day is easier to keep as your default.

04

Your privacy changes when you open a result

Imagine searching for a walking route and opening a travel website. Your search provider's promise covers its service. The travel website has its own servers, policies, advertising and embedded content. Startpage explicitly describes this boundary in its policy.

The destination can normally receive your IP address and browser details. If you sign in, it can identify your account. Advertising scripts or embedded videos may create additional connections to other companies. A private search engine does not control all of that.

Cookies are small pieces of site data in your browser. Some keep you signed in or remember preferences; others help track activity. Fingerprinting uses a combination of device and browser characteristics to recognise a browser, even without an ordinary tracking cookie. Mozilla explains how its tracking protection handles several of these techniques.

Reject optional cookies where you can, use browser tracking protection, and grant camera, microphone, notification and location permissions only when the page needs them. Logging out can reduce an obvious connection to your account, but does not by itself remove cookies or other ways to recognise your browser.

A separate question: A result can respect your privacy and still lead to a scam or an unsafe download. Check the destination address and keep your browser's security protection enabled.

05

Give your browser a say in what gets collected

Your browser is where the next stage happens. Pick a browser you are happy to use, keep it updated, and look through its privacy settings. Here are three practical options.

Firefox: familiar browsing with tracking controls

Firefox includes Enhanced Tracking Protection and cookie isolation. On desktop, open Settings, then Privacy & Security, and review the protection level. Strict blocks more tracking, but can affect site features; use a narrow exception when a trusted site needs it.

Brave: built-in blocking for everyday browsing

Brave includes Shields for blocking trackers, cross-site cookie tracking and other techniques. Check Shields on a page to see what is being blocked. Review optional features and data settings as well; having protections built in does not mean every extra feature has the same data requirements.

DuckDuckGo Browser: search and browsing in one place

The DuckDuckGo browser adds protections beyond the search box. Its documentation describes blocking known third-party tracking scripts and handling several other tracking techniques. Availability and limits vary by platform, so check the documentation for the device you use.

A few well-understood protections are easier to manage than a large pile of overlapping extensions. Review extensions you already have, remove ones you no longer need, and check their permissions before adding more. A privacy-oriented browser can still keep local history, connect to services and handle account data; review its own settings and policy too.

06

Incognito, VPNs and Tor solve different problems

Private windows: fewer traces in the browser

A private or incognito window is useful for a session you do not want added to your normal browser history. Close all its private windows to end that session. As Google's own explanation makes clear, websites and network operators may still observe activity, and downloaded files and saved bookmarks remain.

Signing into an account identifies you to that site even inside a private window. Use this mode for local separation, such as shopping for a surprise on a shared computer. It does not hide your IP address.

VPNs: moving the network trust decision

A VPN routes covered traffic through a provider, so websites generally see its exit address instead of your home connection's address. Your internet provider can see that you connect to the VPN, while the VPN provider becomes a party you must trust with the connection. Cookies, account logins and browser fingerprints still matter.

Check what a VPN covers: a browser-only feature may leave other apps outside the tunnel. Mozilla's Firefox privacy notice explicitly distinguishes its browser proxy from protection for other applications and says the feature alone does not make browsing fully anonymous.

HTTPS already encrypts traffic between your browser and an HTTPS website. That protects page contents in transit; it does not stop the destination reading your requests or necessarily conceal which service you connect to. The EFF's HTTPS and Tor explanation shows how those layers differ.

Tor Browser: a stronger separation from your IP address

Tor Browser routes web traffic through the Tor network and includes protections designed to make browsers harder to distinguish. The Tor Project explains how this goes beyond incognito mode.

It can be slower and some sites challenge or block Tor connections. Follow the project's best practices, including avoiding extra browser extensions. If you sign into a personal account or provide identifying information, that website can still know who you are, even though Tor hides your original IP address.

Choose for the job: A private search engine and tracking protection are a manageable everyday start. Use private windows when local history matters, and consider network tools when hiding your connection's address is part of your goal.

07

A ten-minute privacy reset

Put your coffee somewhere safe, open your browser settings and make a few deliberate choices.

1. Choose one search engine. Try DuckDuckGo or Startpage with a few searches you would normally make.
2. Set it as your default. Find Search or Search engine in your browser's settings. Changing the homepage alone may leave address-bar searches with your old provider.
3. Review search suggestions. Address-bar suggestions can send text before you press Enter. Check which provider receives them, or turn them off if you prefer.
4. Review tracking protection. Check your browser's privacy controls, optional cookies and site permissions. Keep security updates and malicious-site protection enabled.
5. Check local and synced history. Decide whether you want history on the device, in a synced account, or neither. Deleting local history does not necessarily delete cloud records or data already held by a website.
6. Separate the occasional private session. Use a private window when useful, and avoid personal logins if your goal is to keep that activity separate from an account.

Try the setup for a week. If an engine misses an important result, use another deliberately, then return to your chosen default. You can improve everyday privacy while keeping the web useful.

When you find tools worth keeping, give them a home in a small bookmark system. Open the sites directly when you need them, rather than searching for the same destination every time.