If you're interested in funding or helping us find funding for a Discord replacement that's federated and end-to-end encrypted, we're interested in implementing that at @spritely ... we even had been talking about that being our big focus for 2026.
-
@strypey @maj @ekaitz_zarraga @cwebber @spritely yeah but replying to a post that was in large part about how good discord UX is, recommending matrix is not the idea i would have gone with
-
@strypey @ekaitz_zarraga @cwebber @spritely You're preaching to the choir here!
But I'm not going to convince the community to try something new any time soon. -
@cwebber this is going to be way too negative but...
are you really planning on making a "discord replacement" or are you making a spritely demo in the shape of a chat platform? the two are different.
matrix is a demo of their federated state synchronization platform in the shape of a chat platform, not an actual chat platform. you can tell from where their priorities are (e.g. the horrendous state of their moderation tooling. access control that's just Big Number Do More)
discord-style mostly-public communities don't need e2e. they barely need federation to begin with (centralize per community, leave the identity decentralized). this is not something that needs currently unproven technology. sure i have nothing against the tech (after all, you need to actually build it and use it for it to be "proven"), but hyping up what you're building as a "discord replacement", if you're not actually committed to that[1] is disingenuous IMO. there's way too much disappointment in the chat space already and people do not need more.
as i said to someone from roomy.chat a few days before: i would love to be proven wrong (but i doubt it's gonna happen, sorry)
[1]: aren't you also replacing "fedi" as well? sounds like a lot of work for a single org, i'm not convinced y'all can handle the workload of all that combined@kopper @cwebber I'm also curious about this. Building a complete Discord replacement would be a lot of work.
Btw have you seen @soatok 's fediverse E2EE project?
https://soatok.blog/category/technology/open-source/fediverse-e2ee-project/
I'd also be curious how your solution would compare to the many existing attempts.
(Just to name one that I haven't seen others mention yet: cwtch)
I love the OCAP idea but I'm not sure if it covers every attack vector that people actually care about, like preventing metadata leaks, forward secrecy, etc etc.
Also, if you are going for full P2P, how are you planning on tackling the usability issues encountered by projects like SSB? (eg: multi-device state sync) -
@strypey @maj @ekaitz_zarraga @cwebber @spritely yeah but replying to a post that was in large part about how good discord UX is, recommending matrix is not the idea i would have gone with
@tay @strypey @ekaitz_zarraga @cwebber @spritely very diplomatically said
-
-
@kopper @cwebber I'm also curious about this. Building a complete Discord replacement would be a lot of work.
Btw have you seen @soatok 's fediverse E2EE project?
https://soatok.blog/category/technology/open-source/fediverse-e2ee-project/
I'd also be curious how your solution would compare to the many existing attempts.
(Just to name one that I haven't seen others mention yet: cwtch)
I love the OCAP idea but I'm not sure if it covers every attack vector that people actually care about, like preventing metadata leaks, forward secrecy, etc etc.
Also, if you are going for full P2P, how are you planning on tackling the usability issues encountered by projects like SSB? (eg: multi-device state sync)@csepp @kopper @cwebber @soatok there is no silver bullet but ocaps solve problems that are inherent in the ACL security model. all the systems built on ACLs are vulnerable to confused deputy attacks, have no way to safely delegate authority, and centralize their authority with administrators even if the architecture is otherwise decentralized. ocaps are an essential tool, imo, but many other tools are needed to solve the other problems you mention.
regarding p2p, I think it's not really feasible to go "full p2p" given the state of the internet. taking lessons learned from AP and creating a network of federated relays is a practical way forward, I think. true p2p connections can be bootstrapped using the relays, where possible.
-
@dthompson @polyfloyd @cwebber @spritely
I'm curious what the fundamental issues are with Matrix that can't be fixed, and what should've been done differently. It seems like a generally pretty solid project.I like the concept and general idea of "open source Discord" very much, but I'm just not sure why building something from scratch would be better than trying to improve one of the existing Discord-replacements, especially since part of the appeal of Discord was the network-effect driven feeling that "everyone is on it". One mediocre platform that can be improved might actually be better than five great ones.
@Kadin2048
@dthompson @polyfloyd @cwebber @spritelytesting Matrix for my gc, when encryption was used we had constant "could not decrypt message" issues and i had to delete and remake rooms without encryption which meant kicking people one by one so i could be the last one to leave the room. There's also no way i can find to organize rooms into categories or stop them jumping to the top because it was the last room used.
-
@danvolchek @cwebber @spritely I found out about a reverse-engineered open-source clone of Discord just yesterday on a discussion about this on Lemmy. Fluxer. The screenshots make it look exactly like Discord, but I haven’t used it. It appears to be in beta: https://fluxer.app
Edit to add that as a Discord clone it’s probably not federated or end to end encrypted, so my post was more a reply to Dan instead of to Christine.
Edit 2: https://spacebar.chat is the reverse engineered one!
@bigducky
@danvolchek @cwebber @spritelyoh these are very interesting. i gotta try to find out what kind of server specs are required and the state of self hosting. like Stoat is ostensibly self hostable but i know someone who has done it and it was miserable.
-
@csepp @kopper @cwebber @soatok there is no silver bullet but ocaps solve problems that are inherent in the ACL security model. all the systems built on ACLs are vulnerable to confused deputy attacks, have no way to safely delegate authority, and centralize their authority with administrators even if the architecture is otherwise decentralized. ocaps are an essential tool, imo, but many other tools are needed to solve the other problems you mention.
regarding p2p, I think it's not really feasible to go "full p2p" given the state of the internet. taking lessons learned from AP and creating a network of federated relays is a practical way forward, I think. true p2p connections can be bootstrapped using the relays, where possible.
@dthompson @soatok @cwebber @csepp in discord-like communities, there is already an inherent centralized authority, in the form of the community moderation staff. i'm not sure what problems ocaps solve in this particular use case -
@strypey @snikket_im @daniel @zash @joinjabber
I'm really just talking about funding open source, I'm sure the issues are being addressed as much as they can be under the circumstances. Crypto libraries in particular need careful high quality attention to be safe and successful.
-
Last year (I guess 2024? When did Discord threaten to IPO/CEO change?), and 5 years ago I launched a Matrix server.
It's good improvements have happened - I saw the Matrix presentation for the new audio and video chats that are coming some day maybe (maybe they released? I think that was 2 months ago), those were cool.
Someone really needs to do a feature parity check nowadays. I mean, I'll build *another* Matrix server, but it's draining.
-
@bigducky
@danvolchek @cwebber @spritelyoh these are very interesting. i gotta try to find out what kind of server specs are required and the state of self hosting. like Stoat is ostensibly self hostable but i know someone who has done it and it was miserable.
@neckspike @bigducky @cwebber @spritely have you seen the docker compose file for Stoat? It's absolutely massive
-
-
We need more institutional support to get to the level where this can be our focus. But in the meanwhile, if you're an individual and you do want to help @spritely advance, you can donate here: https://spritely.institute/donate/
This turned into a huge-ass thread between going to sleep and going to the dentist.
To everyone who's saying "Matrix!" and "XMPP!" those are great ecosystems that should keep moving forward and we should simultaneously support.
I'd like to reframe what we're interested in doing, and in a good position to do. I have to get through some things in my day first tho.
-
@strypey @Kadin2048 @kitkat @polyfloyd @cwebber @spritely I don't follow matrix protocol development very closely so I wasn't familiar with all of these. my main issue is the lack of capability security in the auth layer. that's hard to retrofit and it effects everything on top. I don't want to build on something that has gone all-in on oauth2 with openid connect.
-
@strypey @Kadin2048 @kitkat @polyfloyd @cwebber @spritely I don't follow matrix protocol development very closely so I wasn't familiar with all of these. my main issue is the lack of capability security in the auth layer. that's hard to retrofit and it effects everything on top. I don't want to build on something that has gone all-in on oauth2 with openid connect.
@strypey @Kadin2048 @kitkat @polyfloyd @cwebber @spritely that said, of all the available options out there *right now*, matrix is the best, imo.
-
If you're interested in funding or helping us find funding for a Discord replacement that's federated and end-to-end encrypted, we're interested in implementing that at @spritely ... we even had been talking about that being our big focus for 2026.
We have the skills and the underlying tech to pull this off. What we need right now is resources. Funding for open source nonprofits like ours really fell apart in 2025. If you think you know how to help, feel free to reach out.
-
This turned into a huge-ass thread between going to sleep and going to the dentist.
To everyone who's saying "Matrix!" and "XMPP!" those are great ecosystems that should keep moving forward and we should simultaneously support.
I'd like to reframe what we're interested in doing, and in a good position to do. I have to get through some things in my day first tho.
Okay, here's my followup. What did I mean by "build a replacement for Discord"? That's shorthand, because to be honest, Discord is *a lot of things* because it has a lot of resources behind it. And of course, people have their own preferred directions, especially in XMPP and Matrix, and I think those efforts are worthwhile. But I'm talking about some near-future threats-and-opportunities, so let me explain what we want to, and ought to, build.
-
Okay, here's my followup. What did I mean by "build a replacement for Discord"? That's shorthand, because to be honest, Discord is *a lot of things* because it has a lot of resources behind it. And of course, people have their own preferred directions, especially in XMPP and Matrix, and I think those efforts are worthwhile. But I'm talking about some near-future threats-and-opportunities, so let me explain what we want to, and ought to, build.
So I will detail in this thread, and yes, this will be a Classic Christine Thread (TM), what I mean. The things we would *like* to do, at a high level:
- Get "moderated chatroom with no center" tech in the hands of users
- Which also includes direct file sharing
- Advance Spritely's core tech in the process. There's nothing like a real world use case with real users to push forward your system
- Advancing that tooling also means opening up some things that you can't do anywhere elseWhat does that mean? Read on! Let's go!
-
So I will detail in this thread, and yes, this will be a Classic Christine Thread (TM), what I mean. The things we would *like* to do, at a high level:
- Get "moderated chatroom with no center" tech in the hands of users
- Which also includes direct file sharing
- Advance Spritely's core tech in the process. There's nothing like a real world use case with real users to push forward your system
- Advancing that tooling also means opening up some things that you can't do anywhere elseWhat does that mean? Read on! Let's go!
There's a lot of things that the XMPP and Matrix ecosystems provide. XMPP has probably the widest amount of implementations out there, and the best written documentation and specifications for those specs. This is great! But the end-to-end encryption story isn't good, and there's still a center to each room. But it's also been around for a long time and there are a lot of wonderful things to the XMPP ecosystem, including a variety of lovely clients and servers!
