- cross-posted to:
- privacy@lemmy.ml
- cross-posted to:
- privacy@lemmy.ml
I’ve been working really hard to research and rank messaging apps by their privacy. The more green boxes the better.
I plan to turn PrivacySpreadsheet.com into a place for privacy data on everything from cars to video games. It’s all open source too on GitHub.
Not trying to advertise, I just put a lot of time into researching all this, and I want to share it since I think others could benefit.
You must log in or register to comment.
Bro put Tinder DMs on the list. Points for being thorough I guess lol.
Jokes aside looks really useful. Good job!
I forgot Grindr DMs, but you already know that ones gonna be red all the way down lmao
Pls share with friends if you find it useful, I dont accept donations or anything, and it’ll never have ads or bullshit.
I’m working on adding more services, but each one takes about 4 hours to research and review.
The is the messenger matrix from the German blog Kukitz-Blog (it is a blog with a strong focus on privacy and is in my opinion well informed). But no worries, the matrix is also available in English.
Maybe you can take some inspiration from the matrix.
The issue with me is ease of use to use with other people. I’ve tried Matrix and Session with other tech minded people and it’s not nearly as seemless as Signal. I’m just waiting for an app that ticks all my boxes, really looking forward to Signal usernames though.
Signal really is that better replacement for WhatsApp since the functionality is identical, others would have to force people to get used to the different ui and the options.
Except Signal UI is… Not good. It feels like using a texting app.
Between the UI and dropping SMS support, I can’t get anyone to use it anymore, and people I had using it have moved on.
Dropping SMS is really frustrating - it was the big selling point I had.
I’m one of those people who thinks SMS has no place in a private messaging app. Signal is the gold standard, and enabling sms merely legitimised this incredibly non private and antiquated messaging protocol.
And gave a constant reminder to people that something better was right there.
And put things in one place.
You’re letting perfect be the enemy of good. At least with SMS support I could get people to switch to “this new texting app”, and we’d then have a proper Signal encrypted chat. And when they texted someone else, Signal would append the “you could have encryption too” signature, generating a conversation about it.
The people who moved off of Signal went back to SMS entirely. How is that better?
Huge bummer. Kind of understood why they did it but they lost a lot of people because of this.
I’ve been using Matrix for years, but now only as a replacement for IRC. The encryption key handling has always been cumbersome and flakey, and too easily broken by users. Not compromised “broken”, but locked out “broken.” It’s been like this for years, and while the UI has improved, it’s still too hard for casual users to confidently use; I’ve given up hope that it’ll ever get to a point where I can recommend it to friends who don’t give a fuck how it works, and who aren’t interested in spending a half hour figuring out how to set things up - they just want it to work. So many encrypted messaging systems have done this correctly, I dispair that Matrix can’t (it’s a common issue with all clients, so I blame the design of the protocol).
Edit oh, I also wanted to say I’d also been disillusioned with Matrix when I realized I couldn’t run my own server. That is, I technically could; I just couldn’t afford to. Synapse is a hot mess of a server, but it also just pounds on the CPU and requires massive amounts of disk space (over time). Matrix is designed such that all content for channels joined by any user is replicated to the user’s home server. It’s a questionable design decison, at best, but a consequence is that regardless of the server software, the storage requirements make running a home server cost prohibative. Compared to, say, running an xmpp server, which could be done effectively on a Pi.
Replicating all chat history + attachments provides a lot of resilience to the network from a node going down, but at the cost preventing to the home lab user from practically hosting a server which just means everything centralizes around Matrix.org, & when anyone on Matrix.org chats with you or your group, that metadata gets synced back to the central hub server once outwardly funded by Israeli intelligence.
I made the mistake of getting my family to switch to Signal. It works great for messaging, but it has other issues—beyond the typical SIM-required complaint. I hate that you have to register with a ‘primary’ device on either iOS or Android fueling that duopoly (SoL if you are on a postmarketOS or KaiOS or Capyloon phone… or just don’t want a internet-capable phone). Notifications are sent thru Google’s FSM (news 1–2 months ago that of course Apple & Google send all the metadata to the feds) & refuse to support UnifiedPush (thank goodness the Molly fork does). They’re also not too happy to support alternative clients meaning you are stuck with the shitty, resource-sucking Electron client while not having a web client or native or TUI client. And the worst cherry on top is shipping those iOS emoji to Android & Linux …eww.
-
Yeah not having it as a default SMS app sucks. Can’t really argue with you there. Perhaps, one could make a fork with it?? Just thought of that now.
-
I seriously doubt any encrypted messenger is going to support OS like KaiOS or non internet capable devices.
-
For unified push, just use molly.
-
iOS emojis…I really don’t care, Signal devs have other things to worry about.
With an FPGA or special CPU instruction set, the encryption algorithms could run on a toaster—which would give access to whatever low-spec handheld you wanted without making it chug to have strong encryption. That also still isn’t covering the future hope of a Linux phone, or someone that just wants to register an account on their laptop.
Using forks puts stress on other teams to keep up with breaking changes, & 90%+ of folks won’t be looking for forks or be willing to trust their unofficial status. I saw the code for UnifiedPush as a Mattermost plugin & it was like 50 lines or something small which is much less than the rest while allowing users to keep control of their metadata which is a big deal if you care about privacy. A fork for SMS support would encounter similar issues, & now you either need to compete with Molly or copy its featureset otherwise users have to choose, SMS or UnifiedPush. That said, I agree with the SMS situation since it was easy to convince relatives to use this new “text app” where encryption magically came to a chunk of their contact list.
Saying emoji was the most important was tongue-in-cheek, but it makes the application feel non-native (& I think Apple’s emoji are particularly ugly). You would think at least the Google set was shipped to Android, or—now hear me out—not ship emoji, don’t override the user experience, let the user’s fontconfig display the one they set. Shipping a whole font (or images) for emoji is why the application size is so bloated for a chat app.
The first two arguments I get. But the emoji argument about not shipping them at all? Yeah if this is going to be a mainstream and easy to use app then that won’t fly. My friends, family, and I all use emojis, gifs, and stickers. I’m sure many people enjoy these things as well. All that bloat.
Are you using a device without an emoji font installed on the system at all? The web works just fine without browsers shipping an emoji font.
-
Is there a way to lock the left cells while scrolling through the other messengers?
Working on it, hard to do well without JavaScript while maintaining the ease of webpage generation
Nice work. Can you add RCS to the table? https://en.wikipedia.org/wiki/Rich_Communication_Services
RCS is a protocol, not a messenger. Google messages is the only client that implemented it.
Unless you know of any other RCS apps
Apple announced to support it : https://www.eff.org/deeplinks/2024/01/what-apples-promise-support-rcs-means-text-messaging
I came here to suggest that as well. I have contacts who are switching from other platforms to RCS, and I have a hard time figuring out how secure that is.
Not that I give a shit, but I can see you potentially catching some flack for listing the USA as an “authoritarian regime” lmfao
Lets be honest, its not much different from China. They both make social media companies censor, and they both track citizens to predict their likliness of committing a crime in the future.
Where’s the lie?
Removed by mod
I wouldn’t say worse than China, but I’d say they’re both equal, in their own way.
It’s the nature of state politics and security. I’d bet even money every government on the planet is equally bad, up to the resources they have at their disposal.
Remember, all governments are collections of individuals, and individuals range in their morality.
Certain types are attracted to certain opportunities…like the power of government.
Would absolutely add Session, I think it’s basically a requirement for this comparison. Great work otherwise
Yes, please add Session. Wire is missing, too.
A version of this with usability features would be nice. Some of these I gave earnest tries, with multiple friends who were willing to indulge my interest, and the tools failed for various reasons: too cumbersome, too confusing, too unreliable, too basic. It’s a subjective metric, but these are social tools, and to be useful, they have to be usable – and many simply aren’t.
I don’t know if it’s humorous, but one unexpected thing I discovered was that Wire’s and Session’s embedded animated GIF finder+inserter is so hugely desireable with my friends, it became an almost minimum requirement. Funny GIFs are immensely popular.
Session, Wire, and Element are done and will be added later today
I just saw Session - thanks!
But now I’m confused. Maybe you could add notes about what some of the rows mean. For example:
- Upon what is based the “recommended for private comnunication?” Recommended by whom? Under what criteria?
- Why is Session’s voice/video “n/a” when it supports encrypted voice and video calls?
- Why is running a private server, rated as higher security than distributed, tor-like onion networks? (can self host), and why is Session listed as “no” when anyone can self host routing nodes in the network? This preference for centralized servers over distributed onion networks is particularly baffling for a privacy-focused table.
This is a huge labor. Thanks again for attempting it.
Based
This is worthy of a more usable interface than this spreadsheet widget.
It took me a fair bit of scrolling to identify which attributes each of the six purple “N/A” values for SimpleX are, but now that I have I agree they’re accurate (though I think there is an argument to be made for just writing a green “no” for each of them).
It is noteworthy that SimpleX is currently the only one of these (currently 34) messengers to not have a single red or yellow cell in its column. well done, @epoberezkin@lemmy.ml! 😀
edit: istm that SimpleX (along with several other things) getting a “no” in the “can hand IP address to the police” row is not really accurate. SimpleX does better than many things here in that they don’t have a lot of other info to give to the police along with the IP, but, if Bob has their phone seized (or remotely compromised) and then the police reading Alice and Bob’s messages from Bob’s phone want to know Alice’s IP address… they can compel a server operator to give it to them. (And it is the same for a user who posts a SimpleX contact link publicly.)
Nice work so far! It’s a big task, really.
Smart idea hosting on git. Gives it a chance to be maintained and have a history.
Any way to download as a csv/excel file? (I can just copy/paste from the web, but that’s imperfect)
I’m working on it, and an Excel file will be available later today under the “datasets” directory in GitHub
I noticed that some of these are apps and some are protocols. It makes sense to list the app if the protocol is proprietary, but it’s confusing that there can be multiple apps for an open protocol and not all of those apps could feature the same level of privacy.
Please submit a GitHub issue so I can track the suggestions and problems, thanks
who has analyzed the code to determine how sweet new comer SimpleX really is?
Well, Trail of Bits did more than a year ago
simplex.chat/blog/20221108-simplex-chat-v4.2-security-audit-new-website.html
This is awesome! Is there a way to freeze the first column? Just so you can scroll to the right and see the categories
Working on it
deleted by creator
You got some errors for XMPP e2ee: the popular mobile clients all enable it by default, it has perfect forward secrecy and a/v calls are usually also e2ee and of course data is encrypted in transit.
Yep. Really need to compare the best-practice XMPP clients (e.g. Conversations, Siskin), not half-developed clients more suited to the XMPP landscape of 20 years ago. – Just as Matrix’s ranking in the table is high because only the state-of-the-art clients are considered – there are plenty of Matrix clients which don’t support e2ee, for example.
This list of mistakes isn’t exhaustive, but extending from poVoq’s mentions, here are some things XMPP(conversations) does actually have positive findings for:
- End to end encrypted by default [OMEMO]
- End to end encryption is available [OMEMO]
- Voice/video calls are end to end encrypted [“calls are always end-to-end encrypted with DTLS-SRTP”]
- Utilizes Perfect Forward Secrecy [OMEMO]
- Data is encrypted in transit [TLS and OMEMO]
- You can verify contacts out of band [https://gultsch.de/trust.html]
- There has been a third party code audit [2016]
- Provider can scan for illegal content [If you send content unencrypted, otherwise no different to Matrix/Signal]
I’m not sure there’s much differentiation between any apps when it comes to “What can the apps hand to police?”; if the police have physical access to your device and app, they have access to everything you do on that device/app.
This makes me feel things. Incredible.
