Return the old application

[Arxad]

Why was it necessary to delete an old beautiful application? I don't need any authorization.
I just wanted to choose which application to use the wg for. Where is the possibility to import wg.conf ?

Return the option to download the old app!

[brendanosborne]

I'm glad to hear you enjoyed using the original version. I was consistently asked for new features beyond the scope of the original app, and I wanted to make that happen. The user auth is to manage subscriptions to a paid premium version that will offer extra features. It is a beta, so I'm testing the auth system but I'm not sure if it'll be mandatory or not in the final release. For now I'm making sure everything works as it should.

You can still download the old version in the releases
https://github.com/TunnlTo/desktop-app/releases

The import feature is not built yet as mentioned here and some more details here

[Permanently]

I found this new release of TunnlTo upon resetting my Windows setup and starting from scratch, and I've got to agree that this is not it. At all. Signing in with email or OAuth for a VPN application alone is ludicrous, and that's not touching on the now closed-source program. I'd have been happy to chip in if it meant open-source whilst paying for a few extra features, but this is a step too far for me personally. Even if it's all free on beta right now, it still doesn't take away from the first point; a good number of people want to use a VPN without having to put in their details.

For anyone in the same boat, WireSock Secure Connect is a good alternative on surface view. Don't know if it's a 100% equivalent as I've only just started as well, but I'd say it's a good competitor - especially based off what other opinions are of the program online.

[brendanosborne]

There have been a handful of people sharing your sentiment, but overwhelmingly people have not been bothered by the sign-in, probably for the simple reason I have reiterated in every release:

There is a requirement to login on app startup. We are testing the user authentication system in this version. User auth is implemented to handle unlocking paid features. It is undecided if user auth will be mandatory in the final release.

I also mentioned in an above post:

The user auth is to manage subscriptions to a paid premium version that will offer extra features. It is a beta, so I'm testing the auth system but I'm not sure if it'll be mandatory or not in the final release. For now I'm making sure everything works as it should.

As I've stated - it is a beta, so I'm testing things, gathering feedback and the product is evolving. However, now that the majority of issues are resolved I am removing the mandatory auth requirement. In the meantime, the open source version that you were previously using is still available.

and that's not touching on the now closed-source
I'd have been happy to chip in if it meant open-source
WireSock Secure Connect is a good alternative

Are you aware that Wiresock (what TunnlTo uses under the hood) is closed source? I made sure to highlight that in the original TunnlTo open source version so there was no confusion, but as you're working in the cybersec space I'm sure you would have looked into this yourself. Either way, I'm curious about your thread model where its critical to have an open source GUI to build a config (the result of which you can open and inspect btw) for a closed-source kernel level networking driver that manipulates your network traffic. Funnily enough, you're now using and have recommended Wiresock Secure Connect which is also closed source (previously open source too - the Wiresock developer did the same thing as TunnlTo) and also uses the closed source Wiresock driver under the hood.

Can you elaborate on the requirement for an open source frontend to a closed source kernel network driver? I like to understand my users different use cases. Is it a business need or something else that I'm not aware of?

[Permanently]

I didn't pride on saying Wiresock is open source. I'd rather have closed-source and no authentication to use a VPN, rather than closed-source and required authentication. All I have to do is look at Portmaster to check it's making the network calls I expect it to. Wiresock being closed-source isn't news, but what isn't good at all about TunnlTo is the authentication layer for what some people use for anonymity - which I really don't think I need to keep banging on about.

If you're going to remove authentication, then fine - I'd much rather prefer FOSS over anything, but what we're left with is closed-source vs. closed-source and entering an identifier. Sure, I can't seem to find an open-source Wireguard client that lets me split-tunnel, but that doesn't automatically mean I'm going to write off closed-source programs, because sometimes that's the only choice. If you really want to do a paid version, which sure, totally appreciate, then suggest to maybe do something like an open-source free version and a closed-source premium version - keeps it transparent.

But hey, your project.

[brendanosborne]

I'd rather have closed-source and no authentication to use a VPN, rather than closed-source and required authentication.

I spent yesterday removing the auth requirement and hope to have a release next week. You will still need to authenticate if you want to pay for extra features though. Maybe in future this will change. A lot of users are gamers or stay at home workers where anonymity is not their priority. They're more concerned about geo-blocks, latency and bandwidth.

All I have to do is look at Portmaster to check it's making the network calls I expect it to.

There was a user on here the other day asking why Wireshark was not representing split tunneling correctly because the capture was happening before the Wiresock redirect. I don't know your threat model, but my guess is Portmaster uses WFP and with "competing" kernel network drivers I wonder if that is enough. I would consider monitoring it at the router level depending on the circumstances, but if you're at that point you probably wouldn't want to be on Windows anyway.

FOSS is good but the right incentives and business model need to be in place. I'm skeptical of the notion that open source automatically equals secure. imo it can lead to a false sense of security. The code needs to be properly vetted either way but in the meantime the app is signed by a EV code signing certificate issued by Microsoft.

Thanks for the feedback.