from jerrimu@lemmy.world to selfhosted@lemmy.world on 29 May 21:33
https://lemmy.world/post/30477060
Hey everyone.
I make Peersuite, an opensource free communication platform.
It’s private by default, there’s no sign-in or email collection.
It’s peer-to-peer, there’s no server, after discovery you are connected directly to your friends my AES-GCN encrypted WebRTC channels. It forms a mesh and identifies superpeers. Because there is no server, in order to save your data between sessions, you can download your workspace into a password encrypted file. Happy to answer any questions.
FEATURES: chat with images, PMs, channels, and file send group audio/video calling screensharing kanban board whiteboard for diagrams/flowchartswith PNG export collaborative document editing with formatted PDF export
The best way for self hosting is docker, its on dockerhub as openconstruct/peersuite. You can also download desktop versions from the github or use on the web at peersuite.space
github - github.com/openconstruct/Peersuite
threaded - newest
That sounds pretty cool! You have listed a lot of great privacy features. I’m sure you know they will likely restrict adoption, because Grandma probably won’t be able to figure it out.
It does look neat and I will give it a try.
I tried to make it super easy, but I’m not a UI guru or anything.
Hopefully gramma will figure it out, I’m a grandparent and I made it lol.
A couple of questions. If I was trying to keep a consistent workspace to build a community around, would it be persistent after the host logs off, and are their tools to protect it from trolls etc who discover it a workspace?
They would have to guess the roomcode and password, it’s pretty difficult to brute-force.
I’m planning some sort of auth system, but not sure how I wanna do it yet.
It would not be persistent. You can download a workspace to an encrypted file. I have plans to make a node.js server for workspace permanence.
Thank you! I’ll be watching with great interest! Lots of potential :)
I have been exploring self-hosted Discord alternatives and had been looking at Rocket Chat, so I am wondering what is the pitch for this versus something like that? I am very early in my exploration, of course.
P2p
yes peer to peer
Rocket chat needs a server, and doesn’t e2e encrypt by default are this biggest differences.
Ok, understood. So if you’re not online, you pretty much lose messages, or are they cached and the next time the sender is online you get them?
My use case is a kid using a minecraft server and wants to talk to his friends, and we’re using mumble now, but they want “discord” and they want things like plugins that allow mgmt from the discord channels, which I would be willing to try to develop, but the model pretty much requires a server to be online.
In general, I’m trying to make a small internet for my kids and their friends to have “normal” internet experiences without being on the wider internet. No youtube, but pinchflat -> jellyfin. No discord, but mumble. No google drive, but nextcloud.
That’as a noble endeavour, IDK if peersuite is the best app for that at the moment.
This space definitely needs competition
I like Matrix, but I do run into issues, like messages not being decrypted even though I verified my session. The average user is not ready for it. Or rather, it’s not ready for the average user
Agreed. Also Discord users complain matrix isnt really a replacement because voice UX
EVERYTHING is encrypted in peersuite, it’s mandatory. I tried to make the UI intuitive and simple, but IDK if I’m great at it.
Simples is best, hands down. Maybe some tooltips wouldn’t go amiss on your demo page?
But as someone else pointed out, a link to the GitHub either in the header or footer of your demo page would be ace too. It was the first thing I looked for this morning seeing your posts and had to come back to the comments to find it 🙈
Keep up the great work, and I shared between our nerd group and they have eyeballs on it. Seems a no brainer for teams collaborating online where we’ve got used of shitty Slack 🤢 Dicksword whatever.
ninjaedit: if you want some help Android app wise, give us a ping. I have a few bored devs lurking looking for app ideas ;)
I think I’m going to go the route of buildind a PWA first then using bubblewrap to generate an APK. I tried capacitor, but audio/video didn’t work, it has meh WebRTC support.
.
it uses WebRTC over https for anything, should just work
Looks interesting. Definitely needs a mobile client to become a viable alternative.
Working on it, capacitor doesnt work because it’s webview doesnt work with WebRTC
I'm gonna try this when I get time to tinker around since my friend and I are becoming exceedingly weary of Discord for communication. I also hate how bulky and rigid it is in its design.
Edit: Tested it on LAN between my PC and phone; it's easy to get running and works great. Some buttons seem invisible on mobile, like the channels list, so I had a hard time joining the "general" channel upon connecting to the workspace. I want to test it with my friend to see how it works across networks but it seems really promising so far.
What kind of phone? I do testing during development and it works fine on my tablet.
I have a Samsung Galaxy S21 FE 5G. It seems like anything near the bottom of the screen gets cut off and I'm not able to scroll down to see it.
I have a galaxy I test on, what can’t you see? Could you take some screenshots?
I just tested it again and now #general shows up right after I join on mobile. I think it's possible that I was having network/browser issues or something at the exact moment I tried it before. :p
I am really into this, it was such a breeze to set up. Are there plans to offer more customization options like custom CSS, an editable configuration file, or something like that? I have a Podman container set up and I'm thinking of creating a volume mount to add some very basic styling but official support would be great.
No current plans, but that would be a great idea! If you just want to edit the CSS, it’s all in index.html
Ah, no worries for now. I took out the inline CSS and replaced it with an external file, so I'll work on that.
I made some minor adjustments already, like removing the transform on hover and having the whiteboard to expand to fit its container, so that's nice. Have you thought about implementing draw.io or excalidraw?
Someone mentioned excalidraw on github also. If it would blend into the rest of the UI I wouldn’t mind, I’m not in love with the one I made lol. Could you make a PR with your changes on github? So far nobody has contributed any CSS changes, and I know I’m not great at design.
I think implementing a dedicated tool will help polish the experience since draw.io and excalidraw both support drawing diagrams and flow charts and that's great for organizing visual data, making mind maps, and brainstorming.
I will look into making a PR next week but I've barely ever used Git and I'm not a developer so expect it to be kinda crap spaghetti code, lol. My CSS experience begins and ends with user layouts and themes on websites like MySpace and Tumblr, unfortunately, but I like making things look nice.
Galois/Counter Node?
Also, why didn’t you drop a repository link? github.com/openconstruct/Peersuite - is it the one?
Yes it is, and I will add it, somehow I thought I did.
Sounds great! Will mobile clients be impossible due to the p2p nature of it?
I’m working on getting an android version up, don’t have a mac so I can’t make an IOS version. It works great in mobile browsers, I have been testing on mbile during development.
Can I tick a box in settings to force it to only connect to Tor Onion Services for enhanced privacy?
Not currently, would love to have it working under tor though.
Does it use double ratchet?
No, just AES-GCN encrypted WebRTC
How quickly does it load when you’re connected to 1,000 chats, each with 100,000 messages?
You know, I’m not sure
@jerrimu @jagged_circle lol!
I read this as a very diplomatic way of saying, "Why.. would you do that? Don't do that." 😏
Yes lol.
So basically you’re repeating mistakes made by matrix
There is literally no way to do performant e2ee at large scale. e2ee works by encrypting every message for every recipient, on the users device.
At 1000 users, that’s basically a public room.
@moonpiedumplings @jagged_circle I read your initial question as 1,000 active chat *rooms* (with some large number of users for each), which.. seems excessive. That's what I was referring to.
1,000 individual private 1-on-1 chats (or group chats with 2-3 users), if that's what you meant (and especially over a long period of time, with lots of inactive chats), seems like a more common scenario*. If that was your question, I apologize.
Lol I misread it too.
@moonpiedumplings @jagged_circle
* I can't speak on behalf of the author, but I could imagine handling it by simply not decrypting _everything_ on startup, and only decrypting an older chat if you click on it or attempt to run a search on everything. Although for a search, I would expect some kind of hashed (and of course encrypted) database that allows a quick search of all prior messages.
Honestly I would just copy what matrix did in 2.0
Everything is decrypted on exiting transit. I use WebRTC for all the data.
I didn’t say 1,000 users.
I haven’t had a chance to test the limits of Peersuite, but since it’s a mesh network, I don’t think it would run well with even 100 people.
Chat is Peersuite is limited to workspaces. You can only talk to people inside your workspace. I didn’t want to add accounts, and sign-ins and all those things that collect data. So when you sign into peersuite, you are only talking to the people in the workspace with you. It works entirely different and your questions don’t really apply to how it works.
@jerrimu A usability suggestion, having just tried it out - save the username and room password in the export file to make it more like a traditional chat experience. So when you import the chat file, the username and password are pre-populated along with the room name.
That’s a really good idea! Now in the roadmap.
I feel like saving the password in the export is a bad idea if security is your thing
The export is password encrypted already. I’ll see if there’s a good way to do it.
Fair enough, it lowers the risk. Are you doing key stretching? Ie. X rounds of pbkdf or whatever it’s called?
@martinb @jerrimu I wrote the initial comment with the idea of saving just the username, but then figured "why not?" for the password. If the password is saved in browser memory (and based how I *think* the app functions, it would have to be), then it wouldn't be much different than saving a password in firefox's password manager (for example). Assuming reasonable crypto usage by the app, of course.
Do you have voice calls/group voice calls and streaming already in or in the roadmap?
Group voice and video are working fine. I’ve only tested screen share with 3 computers but if works fine also.
I just realized it’s on the post. Sorry for that. This looks like a solid thing, for sure. Since we’re talking, do you ever worry about people making use of the privacy features for nefarious purposes?
I guess that could be an issue. I don’t think privacy makes you a criminal, but sure they want it. But so do tons of people who follow the rules. A Scottish Indepence group cloned it on github. That’s really coo to me, I love the idea of political dissidents using it, but the idea of it being used to hurt people does bother me.
Well hell I may stand this up tonight. My only question is does the voice chat support push-to-talk?
Edit: Ok, gave it a spin. It does not support push-to-talk but being fully browser based I don’t think that’s a trivial thing to implement anyway.
That said, this is pretty sweet though certainly still rudimentary. I was really looking forward to the screen sharing but my friend on the other end said the quality and framerate were pretty bad. Not sure what flexibility there is as far as adjustable bit rate and framerate with what you’re doing but I’ll definitely be keeping my eye on this project.
Push to talk is in testing now. WebRTC changes quality automatically based on bandwidth. Its usually really good. It runs at full quality in testing ( i have like 6 devices I hook up and test). If you guys normally have a good connection, try again?
Perhaps talking about bitrate wasn’t correct of me. After looking at this again image quality itself is actually pretty good but the framerate is a different story.
To provide context, I used it to share the video game I was playing as my friends that use discord tell me they primarily stick to it for its screen sharing capability which they use when gaming.
I’m not sure how to best test this and provide metrics to you if this is improvable or even something you care about.
To attempt to take the connection factor out of the equation I opened two browser windows and viewed my own screen share from a different username and even then the framerate is not great.
How are you using it browser or electron? What OS? It looks great in my testing, wil ltry and reproduce the effect you’re getting,
ptt is live on git & site now.
Will you package it to flathub?
I could package the electron version to flathub, I hadn’t considered that. I will read up on it.
feel-ford-bean-base
pass: poop
workspace code ^^^ for anyone that wants to test it out with me. I don’t have friends to do so :(
My group recently switched to Matrix and so this would be a tough sell, but it seems interesting. I haven’t been a fan of Matrix and miss the ease of UI in discord, but was happy to leave with it’s direction. How would you sell it with a small group that has small, but mounting usability issues with Matrix?
Get everybody in to at least try it. I tried to make everything as intuitive as possible.
Does it have voice rooms yet ?
Voice is workspace-wide currently, there’s no way to separate off as it stands now