Rtings is now reviewing which VPN you can trust with your personal data. Spoiler: Not many

[anonymous520]

tldw: dont use hotspot shield

seriously though, cool video, great to see this type of VPN evaluation.

[kissu]

@sp_rtings

If you missed it, it means it might not be clear enough

Haha, as said in my initial message I did not miss it (as shown by the green arrows) but it is an easy thing to actually miss indeed. Thanks for considering an update!

@sp_rtings

I like the idea, but it is not a thing we normally do. VPN are the first software product we test, which is already going a bit off our beaten path. I’ll see if we can have a special, one-time investigation for this which would output an article instead

I understand the limited bandwidth and need to stay focused on a given goal, but thanks for considering the idea!

[obscuracarl (Carl @ Obscura)]

@sp_rtings

Obscura is most likely a VPN we will be testing in our next batch since we keep getting ask about it. I’m not sure when this is happening though. This is a business decision a bit outside of my power. However, the more people ask for it, the faster it will happen!

Feel free to reach out to me directly if y’all have any questions: carl@obscura.net

[anonymous520]

@sp_rtings I see in your Proton review you have “Kill switch works as intended.”

I know the kill switch has been a big issue here as of late.

ProtonVPN Additional note: Killswitch failure and IP Leakage on Linux

Article: ProtonVPN IP Leakage on Linux and Workaround | PrivSec - A practical approach to Privacy and Security Killswitch does not work with all network interfaces. Can you be tested using following command curl --interface <physical interface> https://ipinfo.io Suggestion add an additional note aimed at Linux users in next section https://www.privacyguides.org/en/vpn/#additional-notes

VPN killswitch Linux

Hello, I have a question regarding a VPN killswitch. I’m on linux and I don’t have nftables installed. I have iptables but an empty ruleset. Whenever I activate the killswitch I have no internet without connecting to the VPN. How is such a killswitch possible without firewall rules? The killswitch doesn’t add an extra interface nor does anything special appear whenever i execute “ip rules”. What other ways are there to block internet traffic on linux? I’m new to the linux kernel and therefor…

Remove IVPN

Remove IVPN from list of recommended VPNs for it’s failure to implement a functioning kill switch. When using Apple services on iOS 16+, a VPN connection does not fully protect your privacy against Apple. Even with an active VPN connection and kill switch enabled, traffic from your iOS 16+ device to Apple servers can leak outside the VPN tunnel and expose your local IP address to Apple. For this reason, during the next release we are removing the kill switch feature from the IVPN iOS app. Conne…

Remove ProtonVPN

Why should this tool be removed? ProtonVPN should be removed from recommendations as its kill switch for MacOs is known to not work: When switching to another ProtonVPN server On computer boot up → In both of these instances, the user’s real IP becomes exposed to any online service it is connected to prior to re-connection to new ProtonVPN server. On ProtonVPN’s official website they claim that “[regular kill switch] does protect you while switching servers with Proton VPN” : What is a kill…

Are you guys considering looking more into that feature, especially for Linux and Mac?

[sp_rtings]

Are you guys considering looking more into that feature, especially for Linux and Mac?

Considering, yes! Doing it: …it’s not feasible for us to maintain this test over time!
It’s not technically hard, but there are too many variables for us to manage at the moment unfortunately. For Linux, there are a few different network managers that we would need to test, which multiplies quickly the number of different setups we need to test. For Mac, the network manager works differently than Windows or Linux, so we would probably need to tailor it’s own test.
That said, those tests are still possible to do. The main problem comes from how we do reviews: We maintain the review through time by retesting the products either continuously, or when major updates/events happen. It would not be possible for us to maintain this test over time. However, there is a possibility we do a one-time test for a few selected VPNs for this. I’ll add this to our special projects!

[JG]

I don’t think the world is expecting you to maintain a list of everywhere on Linux you have tested and evaluated VPNs. Even covering just one like Fedora for example and maintaining that is enough. The real point here is to have all OSs covered with VPNs, is all. Not to provide info on many all the time where everything must remain accurate. That of course will be a logistical nightmare to ensure or keep ensuring.

[sp_rtings]

Fair point! Testing one disto, one network manager is still better than none. Our main hesitation was that people would not care much if it wasn’t their own. The more direct feedback I can get on this from users, the better I can convince my peers to move on this!

[JG]

@sp_rtings

Our main hesitation was that people would not care much if it wasn’t their own.

People who own Mac’s are not going to care about Windows or Linux, doesn’t mean there should not be testing for all.

Same logic applies here. I suggest choosing a super popular stable one like Fedora. They keep up with the latest updates and are rock solid stable. Plus they also have immutable variants so that’s a plus. I don’t think DE matters and anyone on another distro on the DE of their choice can also move to Fedora should they want to if VPN work best or know about how it works on Fedora even if not on their preferred.

That’s just my thinking. You can deliberate on this more internally if needed.

[kissu]

@JG

People who own Mac’s are not going to care about Windows or Linux, doesn’t mean there should not be testing for all.

When you do have a business, allocating resources is important and putting 50% of extra time into a 3% usage OS might not be the right call.
Rtings are doing great on that regard and want to do the effort but most people don’t want to invest crazy effort for a super niche community that won’t bring them any views/money/etc (this is a generic statement as of why people do not care that much about Linux in general).

[JG]

@kissu

extra time into a 3% usage OS might not be the right call.

Sure. That’s why Proton lacks on Linux so much and hence the reason for the majority of the gripe against it on the VPN and Drive fronts to say the least.

But if you’re considering desktop Linux, it also means considering the 3%. Otherwise Linux will continue to be disparaged and not be taken seriously even in 2026.

[sp_rtings]

When you do have a business, allocating resources is important and putting 50% of extra time into a 3% usage OS might not be the right call.

This is the main reason why we did not test the killswitch on other OS in the first place and focused on Windows.
The new argument now, is if we are testing VPN for a privacy recommendation, it makes little sense to test the killswitch on Windows only…. so the arguments changed since the initial discussion!
The more traction we get on this new privacy test, the more I can convince the need to test the killswitch on Linux internally!

[JG]

@sp_rtings

The more traction we get on this new privacy test, the more I can convince the need to test the killswitch on Linux internally!

There already is traction. I just can’t prove it to you but I assure you, anyone even new to Linux (along with those using it for years) are going to want this. It’s valuable info!

Also think of it in another way - you’ll be one of the only entities performing a truly holistic testing of VPNs and will be continuing and building on the great work you have been doing and did on this video thus far. The critical thinking passionate tech users are going to recognize your quality work. This community already does.

[anonymous520]

@sp_rtings

Windows only

Or at the very least, it should be obvious to users that the test is for Windows only. I am not sure “Kill switch works as intended.” provides that clarity unless you feel its safe to assume most readers don’t just look at the pros / cons summary and leave.

[sp_rtings]

it should be obvious to users that the test is for Windows only.

We noticed that today as well when taking a new look at it! We’ll make it clearer.

[kissu]

@JG

Sure. That’s why Proton lacks on Linux so much and hence the reason for the majority of the gripe against it on the VPN and Drive fronts to say the least.

Proton definitely has no excuse indeed, especially given their size + employees.
At the same time, it doesn’t look like they are hiring/interested into getting people to code for desktop/Linux environment, they stick to Web/mobile from the offer I last checked.

@JG

But if you’re considering desktop Linux, it also means considering the 3%. Otherwise Linux will continue to be disparaged and not be taken seriously even in 2026.

This is why FOSS is so appreciated into some communities, because people coming from Open Source usually consider the following order: Linux → MacOS → Windows because of ecosystem/friction/OS BS.^[1]

@sp_rtings

The more traction we get on this new privacy test, the more I can convince the need to test the killswitch on Linux internally!

You can always submit a new thread with a topic to check the interest of people willing to have a Linux test.
The more people interested, the more reason to do it for Linux.
The more Linux stuff available, the more people will be interested.
If it picks up well enough, could be a very virtuous loop overall.

---

Side note but if you make the test on Linux, you can also share it on OpenWrt because their community will also be very happy to learn about it.^[2]
Same for quite a few Linux news/websites that would gladly share a cool group effort.
Windows is more popular and the go-to basic OS, but Linux is more of a passionate-driven community-centric hardcore believers that might be worth fighting alongside if you want an audience that keeps coming back and support the Rtings team!

---

1. or at least through my Webdev own lens + experience, might be very different for a Java/Kotlin dev ↩︎

2. and probably other places where Linux is used, hence A LOT of other places like routers etc ↩︎

[Blackbird]

@Blackbird

The news made me buy IVPN. I turned all the settings to maximum just to test it, and everything seems to be working so far. So maybe this is the one!

After testing it properly, I’ve started to like it more. I did have to install Proton back on my other phone because I couldn’t confirm payments, and I really tried every possible method several times. This isn’t a problem because I plan to keep my Unlimited subscription. If anything, it shows the opposite of a problem. This also happened with Mullvad, but on my PC as well, despite all the tricks I tried.

On my PC, I can handle my banking when I use V2Ray. One time, a CDN probably blocked the traffic, but luckily you can change the port very easily and continue. Switching between single-hop and multi-hop has been made so easy that it only takes a few seconds.

Changing the protocol when needed is also simple. And the fact that you can automate WireGuard key changes is something I didn’t even know to wish for, as I had never seen anything like it before. Thanks @sp_rtings for the article that made me give IVPN a chance!

[Blackbird]

I’ve been a complete and utter moron. It’s only now that I finally grasped the idea of installing the Proton VPN on the profile I use for handling payments. It eliminates all the unnecessary fiddling, and at least I haven’t managed to produce any leaks in any of the tests I’ve run. Now I can just keep it on multi-hop permanently.