Before I put serious time into this, I wanted to get a quick sanity check from people here. Not trying to pitch, start drama, or add another slop project to the ecosystem, and I know contributing to an existing project is often smarter than starting fresh. I would rather find out that something already covers this or that the trade-offs make it impractical before writing code.
What if you took the open federation and mature OMEMO foundation from XMPP, Dino, and Gajim, paired it with the community/space feel and self-hosting ergonomics of things like Movim, Mony (Harmony), or Fluxer, and aimed for the snappy native UX and solid Linux media polish people expect from Discord’s upcoming Rust client?
Right now each piece leaves clear gaps around drop-in voice rooms, screen sharing with system audio on Linux, granular per-channel permissions, and consistent desktop/web parity. Does combining those pieces make architectural sense to you, or does mixing XMPP and E2EE with Discord-style features inevitably create a splintered ecosystem or compromise one side too much? Where do you see the biggest pain points?
Edit: To clarify, the goal isn’t another chat messenger like Signal or WhatsApp or SimpleX or Cwtch, etc. Nor is an attempt to create some Matrix based client like Element. This is specifically aimed at a real, federated alternative to Discord for communities, voice rooms, and screen sharing.


Fair point on Signal and Matrix for 1-on-1 or family messaging, and I agree nobody wants to migrate relatives onto a small project.
To clarify, the goal here isn’t a WhatsApp or Signal replacement. It’s specifically aimed at the Discord gap: persistent voice rooms, screen sharing with audio, spaces, and community permissions, built on top of existing XMPP infrastructure so it isn’t locked to a single app or server going down.
Instead of focusing on one or more features, why not design something that can be expanded to do anything and then focus on building the features you want and other people can build other features onto it?
Start with a solid base that provides both anonymity and privacy. I’ve looked at all the chat apps I could find and I think Cwtch or Briar are probably the best bases to start on. Check whether those projects are capable of being expanded to support the features you want. If not, perhaps then build a better base.
Personally, I’m not interested in federation (eg. email). I would focus on something peer-to-peer without servers and uses onion routing to hide IPs. Cwtch and Briar are like that.