While reviewing Mullvad VPN’s browser extension code for interesting solutions in the product, I noticed that Mullvad recommends using privacy-focused search engines, namely DuckDuckGo and Mullvad Leta. I immediately had the thought: “What? Mullvad Leta? What is it?”, With this thought, I went to study this search engine.
Today I will talk about what I was able to find out about him using their FAQ, as well as other sources.
About Mullvad VPN
Mullvad VPN is a VPN service that began its existence in 2009 and during its work has established itself as one of the best VPN services. Why this is so and what features it has – we will not analyze in detail, I will only say about a couple of points that, in my opinion, make it so:
The absence of any binding to a person (when registering, the service does not require anything from you, your login is 16 digits);
Open source (absolutely all the applications that Mullvad has, open and you can examine them under a microscope at any time);
Country of registration Sweden (here everyone draws conclusions for himself – plus or minus);
Continuous improvement of the product (regular updates, introduction of new functions, development of new products).
So we figured out the developer company (if you weren’t already familiar), and now let’s start studying Mullvad Leta.
Disclaimer: The following will be information from the FAQ section on the Mullvad Leta website, which I translated, as well as some other information that I found on the Internet.
Let’s start by looking at the meaning of the word “Leta”. What does it all mean?
According to the FAQ, the meaning of this word is:
Leta is a Swedish word meaning “to seek”, “to hunt”, “to be on the lookout” or “scout”. It is pronounced “Lea-tah”.
Also, for those who are not familiar with Mullvad VPN, I will say that Mullvad means “Mole” in Swedish (as you can see, it is depicted on the logo of the service).
Is Leta a search engine built from the ground up?
I hasten to disappoint you, but no, this system is a proxy server between you and the Google Search API (yes, that’s it).
Mullvad Leta is a Frontend for Google Search API. The request that you make first goes to the Mullvad Leta server, which then makes the same request to the Google Search API, thereby anonymizing the user.
The search engine itself has this stack:
Here is the situation on the servers:
The Leta servers are running on STBooted RAM only servers, the same as the VPN servers. These servers are running the latest version of Ubuntu LTS with Mullvad VPN’s own kernel, which is configured to remove junk.
Mullvad Leta has 2 search options – Normal and Cached.
Cached lookups work like this – each lookup is stored in a Redis in-memory cache, which is deleted when it reaches more than 30 days of retention.
Cached queries are retrieved from this store, which means that if no user has searched for what you are looking for, then you will get stale results.
What do you need to start using?
This is where some may get frustrated, as you need to have a paid Mullvad VPN account to use this search engine. But why might some be upset? It’s simple, the service does not accept bank cards of the Russian Federation for payment, and in this case, you can only pay with cryptocurrency (by the way, you can get a 10% discount when paying).
If you just try to enter the site, then you will see this picture:
The service also has limits on the number of requests. If you make a request not from the cache, then you have only 50 requests per day, while when you go to the next page in the search results, you also lose 1 request. And if you select the “Search only in the cache” option, which is the default option, then the number of requests will remain the same as it was.
Why is the number of requests limited?
Mullvad writes that each search requires them to make an API call to Google, which costs money and they want to provide customers with a valuable service without the possibility of abuse.
Everything is anonymous…almost!
I consider it important to mention that Mullvad collect some information, namely:
We collect metrics related to the number of cached searches versus direct searches solely to understand the value of our service. In addition, we collect information about the use of the processor, RAM and other similar information to ensure the smooth operation of the service.
For a short time testing this search engine, I personally was very pleased.
The service is made in a minimalist design, there is no visual garbage in it, and the search results are such that it turns out to find what you are looking for the first time.
Unfortunately, the source code for this product is not yet available (if it is available at all, hopefully).