Frequently Asked Questions

What is Kagi?

Kagi is created with the the goal of a more humane web. We envision the web of creativity and self-expression, where content exists to amplify knowledge, where user privacy is respected and where ethical web principles ensure a sustainable and better web for the generations to come.

Kagi is currently made of Kagi Search, which is a fast, 100% privacy-respecting web search engine and Orion Browser - a fast, zero-telemetry browser. In the future we plan to release an email product as well, offering a completely alternative way to experience the internet in a private, secure and unbiased way.

Kagi Search is a very fast, user-centric, 100% privacy-respecting web search engine with unique results augmented with “non-commercial” indexes and personalized search results.

Kagi Search is designed as a premium (paid) search engine. It is web search built by professionals, for professionals. We want Kagi to be at the intersection of the web and humanity and we want it to be something that people will bring into their lives and use it to amplify their abilities and make them more creative and smarter, not just make them more productive.

Kagi has clean, high performance user interface with only the most relevant results and no telemetry, no ads and no collection of private information.

Is Kagi Search available now?

Kagi Search is currently in invite-only beta. You can sign up by visiting We send new invites every two weeks.

Please feel free to report bugs or share feature suggestions with us. We are completely zero-telemetry and we rely exclusively on our users for product development decisions!

Why use Kagi instead of Google?

We'd like to first acknowledge that Google has indeed changed the world and the way we access information and we value this contribution. We are therefore grateful to be able to use some of Google's search technology and infrastructure in our search product.

This is also where similarities end and our product philosophies start to diverge. We are not trying to build a search product for billions of people, but a refined search experience for sophisticated customers who value high quality results, privacy and speed above everything else.

why use Kagi instead of Google

Why use Kagi instead of DuckDuckGo?

DuckDuckGo has shown the world that a privacy first search engine is possible and we respect this contribution.

But that was more than 10 years ago and DuckDuckGo slowed down with innovations. Its search product is just 'good enough' and has gone for years without any truly new, ground breaking features. Kagi Search already has advanced features like Lenses or personalised results, and some of these crowd-favorites would be problematic for DuckDuckGo to implement as they could limit their ad-based revenue.

The reality of an ad-supported business model is that it will unfortunately force DuckDuckGo to always balance between being user-centric and serving the advertisers, which inevitably hinders innovation.

Because we are depend only on our users for revenue, Kagi is able to offer a much richer search experience. We hope that our relentless focus on the user and their needs already shows throughout our product.

why use Kagi instead of DuckDuckGo

Any search, even with a “free” search provider, has a cost. Your life’s private details pass through your searches and your thoughts and actions get influenced by what you see in your search results. Businesses that monetize your data owe you a lot more than a "free" account for what they get!

Here at Kagi, the arrangement is clear and simple. We are taking a stand against ad-supported business models and designing Kagi Search as a 100% ad-free search experience, focused entirely on the users.

As a Kagi Search user, you get a search experience tailored for you and your needs. Yet you need to be little different to use a paid search product, and you need to think differently to be a Kagi Search user. People that use Kagi are not just using a search engine, they are making a mark on the world, and impact it in small way for the better.

why use Kagi instead of Google

Where are your results coming from?

The short version is we use heuristics and deep learning to understand query intent and select the best information sources, then query them directly. We're using a large number of sources. You can think of Kagi as a "search client" (similar to how you have "email clients") that connects to different indexes and sources of information and does all the hard work in the background to package them into one superior, secure and privacy-respecting search experience for you.

Under the hood, once we understand the query intent, we select the best places to find the answer, and then query them directly in real time via APIs. We then apply a ranking system for the results (which may involve user personalization preferences).

Among the sources we use are traditional web search indexes (including both Google and Bing) where and when we need them. When we do so, it is done in an anonymized way so these engines never know whom is the query coming from. Additional sources of information include Teclis (our own non-commercial index), TinyGem (our news index), Kagi AI (for instant answers), Wikipedia, vertical integration APIs (like DeepL for translation and other APIs for "answer" type widgets like unit conversion and so on).

Teclis and TinyGem are a result of our crawl of about 6M domains, focusing primarily on non-commercial, high quality, high signal-to-noise ratio, content. At the moment our unique results currently serve 5%-10% of all results that you will see in your Kagi searches, helping surface less known, but high quality content from the web.

Finally, a lot of queries are answered by our instant question answering systems, for example when you ask how far is the sun). Significant number of queries like for example 10 usd in eur are answered using dozens of different sources and APIs, all connected for you in a fast, privacy-respecting and transparent way.

How does Kagi achieve superior search quality?

Kagi is able to surface unique, high quality, non-commercial content in its search results while penalizing results heavy in ads and tracking.

For example, here are the three types of search results that are new and unique to Kagi (for the query steve jobs).

search quality example
Actual Kagi search results

These highly relevant, high quality results, can not be found on any other search engine in the world currently for the same query and are 100% unique to Kagi Search. These results are characterized by being non-commercial and highly authentic authentic. We even crawl the Web Archive to dig up great content that does not exist on the original website anymore, but does still live on the web thanks to Web Archive. When using Kagi, we show you a percent of search results that are unique to Kagi and do not come from the index of other search engines.

Kagi also allows users to set preferred domains in search results, and completely mute results from others. This allows for an unprecedented level of personalization and search result refinement.

search quality example

Finally Kagi also features "lenses" which are special user-definable filters that allow the user to narrow the scope of search to only a couple of websites for example. Make your own "Programming" or "Recipes" Kagi lens and enjoy results only from sites you trust.

Does Kagi do anything to actually protect the freedom of speech, and does it censor results?

Kagi does not censor anything results based on political or any such affiliation. Since we get some results from Google they would be subject to their biases but on the other hand we also get results from Bing and our own index, which helps reduce those biases. A good example of anti-bias product thinking is our "News 360" lens which includes news from respectable media outlets from all over the world, all at once.

The only signal that influences our ranking is presence of ads/trackers on a website. Those websites that are heavily bloated will be penalized no matter what their agenda is.

How does Kagi rank results?

Our algorithm takes into account perceived relevancy of sources and gives (currently small) priority to non-commercial sources. For example let's take a search result coming from a very heavy tracking/ad-monetized site. Whether this will cause Kagi to push it down in results depends on where its organic position would have been in the results prior to our ad/tracker penalty. If it was highly relevant to your query, it wont affect it much if at all but if it is competing with other results (say, around rank 10 or lower) then it will be punished much more in favor of others.

On top of that user personalization are applied, such as "muted" domains get pushed down and "prefered" domains get pushed up in the results.

The way that ad-monetized search engines work is there are so called 'organic results' which are the best results that a search engine has. Then on top of these results for example Google adds 'paid' results from Adwords and DuckDuckGo adds paid results from Bing Ads. These results usually have something like 'ad' added to it and these are the results that your adblocker removes. When Kagi gets results from Google or Bing index through their paid API, it gets only the organic results. So results in Kagi are always best organic results from these engines.

How does Kagi's model reduce my 'privacy threat surface'?

Let's consider two search engine business models:

- One is free for the user, but the real customer is advertisers who buy their services (for example Google or DuckDuckGo)
- The other one is paid, where user is also the customer (such as Kagi)

In the first model there are only two ways to make more money - get more users or sell more ads per user. Once an ad-monetized search engine hits a ceiling of number of users interested in its product the only other way to make more money is to sell more ads per user. In order to do that, it has to increasingly invade user privacy in order to mine more data so it can sell ads more effectively. For example Google was doing only contextual ads at some point in time, only to give in to the pressure of investors and start adding more and more ad-tech which ultimately led us to a privacy issues on the web we have today.

On the other hand, in a model like Kagi, we only make money when the users pay us, and they only pay us when our product delivers on its promise. And our main selling points are result quality, features, speed and of course privacy. If we compromise on either on these, we lose users. So our business model incentivizes us to become better and better (which is also great for our users!) including having the best possible privacy.

The structure of these core business incentives is why, when all other things equal, a paid business model for a search engine reduces privacy threat surface where an ad-monetized inherently increases it. A paid business model is the only one to guarantees the alignment of incentives between a search engine and the user.

We have designed our search product in a way that we do not need to monetize your data, time or attention in any way, ever. On the contrary, our number one goal is reducing your time to getting an answer and leave you to your other business.

What data does Kagi collect?

Only bare necessities to run the service. Please see our privacy policy for more information.

How fast is Kagi?

Ahhh... speed. "I feel the need... the need for speed!" is a quote from "Top Gun" that we live by. We are obsessed with increasing the speed and lowering the latency of our search product.

kagi speed user feedback

There are three ways Kagi achieves this. First, we optimized our technology stack to increase code execution speed and decrease HTTP(S) connection latency.

Then we have minimized the amount of data transferred between Kagi and the browser. Kagi Search is lightweight and transfers as much as 20x less data for the same search results page compared to some of our competitors. This not only considerably increases your speed but also has a neat side-effect on reducing CO2 emissions (as every byte transferred over the web impacts it). Using Kagi Search will not only benefit you but also the environment.

Product Search Result Page Size CO2 Load Time
Kagi 0.76 MB 0.43 g 0.4 s
Ecosia 1.55 MB 0.89 g 1.2 s
Bing 1.94 MB 1.11 g 0.8 s
Google 2.43 MB 1.39 g 2.4 s
DuckDuckGo 2.48 MB 1.42 g 2.1 s
Data from

Finally, our infrastructure is global and you will always be automatically connected with the Kagi node closest to your location.

Can you share a few pro tips for using Kagi?


We take pride in our own index featuring high quality, non-commercial results. To get a feel of it try a query like 'steve jobs'. You will see a number of organic results you will not find in any other search engine. There is also 'Digs' section for results we think may be particularly interesting and 'Blast from the Past' section for results we thought are high quality but are not on the traditional web anymore, rather they exist saved by the internet archive project.

why use Kagi instead of Google

Kagi also features instant answers for your questions using our advanced AI capabilities. If you ask Kagi a question, in many instances you will get a direct answer. Kagi also features a 'computer' for use as an advanced calculator, unit conversion or translator. Just try something like '5*123', '10 ft in meters' or 'translate hello to spanish'.

In Kagi you can use search operators that you are used to as well as all search bangs popularized by DuckDuckGo search engine. For example, if you want to search for TVs on Amazon, you can enter '!a tv' and will automatically take you to the right search results page on Amazon.

If you have the skills to work with CSS, Kagi allows you to specify a custom CSS modification, so you can practically make the search results page look like anything you want. This is just one example of the many power-user settings that we have, with many more on the horizon.

Finally, try Kagi Lenses. Define a "lens" by specifying what sites you want to see in the results. This is a pretty powerful feature in the hands of a power-user and we provide a couple of lenses to get you started (for example a lens to search only online discussions and forums).

How do I set Kagi as my default search engine?

Please find instructions for your browser below.

Instructions for Chrome based browsers (Chrome, Edge, Brave, Opera, Vivaldi...)

  1. Open Menu -> Chrome -> Preferences -> Search engines -> Manage search engines.
  2. In "Other search engines", find Kagi and click the three dots to make it default.
  3. If Kagi is not in the list, click "Add". Type in "Kagi", keyword "k" and for URL Make sure to click to make it the default search engine.

Instructions for Firefox based browsers

  1. Log into Kagi, then right click browser address/search bar and select "Add Kagi Search".
  2. Go to about:preferences#search and change the default search engine to Kagi Search.

Instructions for Safari

  1. Install the unofficial Kagi Search for Safari extension from one of our community members.
  2. Activate the extensions in Safari->Preferences->Extensions.
  3. Enjoy Kagi search from your Safari address bar!

You can also try Orion browser (Kagi's WebKit-based browser for Mac), which has built-in integration with Kagi Search.

Generic use
For search URL use:{searchTerms}

and for autosuggestions use:{searchTerms}

What is the name of Kagi mascot?

It does not have an official name yet. But we like to refer to him as 'doggo'.

Are you hiring?

It depends. Do you have advanced Crystal (backend) and/or vanilla Javascript (frontend) knowledge? Can you solve difficult problems and code exceptionally well? (We’ll send a test project.) Send your CV to

Security Issues

If you discover potential security issues on our platform, please disclose them to our security contact.


Kagi's founder is Vladimir Prelovac: "With Kagi and Orion, we’re striving to build better ways to experience the web, ways that are safe yet fun for everyone, including my own kids." Kagi was created in 2018, is bootstrapped by the founder and is being built by a world-wide passionate team.

We are also building Orion, an insanely fast, privacy-respecting browser, currently in private beta for Mac.

If you'd like to know more or get involved, please feel welcome to reach out to

We hope that our work impacts people's lives in some small way for the better. Thank you very much for being a part of this.