Introducing Fluux Messenger: A Modern XMPP Client Born from a Holiday Coding Session
19 points by adamcstephens
19 points by adamcstephens
I’ve been a big advocate of XMPP nearly 20 years ago (maintaining a public server with 2000 active accounts, spending time in the XMPP dev room at Fosdem).
I dropped the ball.
But, for some reason, those last months saw a lot of XMPP-related news/project. It’s like a kind of renaissance and I’m very glad it happens.
So thanks to all the people active in the XMPP world who, unlike me, didn’t abandon the ship!
Up until a few years ago, my only experience with Jabber was Google Talk (as I'm sure was the case for a lot of casual users at that time). I knew it was an open protocol, and I was happy to be able to use it on Linux just as I was transitioning to Linux as my main platform. Then of course Google Talk went away and people thought Jabber was dead even though it wasn't. I went back to IRC.
I'm very glad I got connected to the Jabber network again because there has been a lot going on that makes me very excited for a present and future built on XMPP. I have been very unhappy with all of the proprietary messengers, many of which require things like smart phones and phone numbers, that have become so popular. Here we have a standard that's been growing for over 25 years with a very active global community of users and developers building servers, clients, and services. It's the best possible time to jump back in!
Google Talk had a strong influence in killing the first wave of XMPP popularity:
https://ploum.net/2023-06-23-how-to-kill-decentralised-networks.html
It took XMPP close to 20 years to recover…
This accusation is often made but as someone who lived through the era is seems a bit overblown. The only thing that went away when Google talk shut down was Google talk users. Everything else seemed to continue as it had been before.
As a fellow XMPP-Nerd, I also disagree with that. XMPP was open to that disruption because the rest of the ecosystem was so weak that it was hard to recommend.
It had a proliferation of clients where many were half-finished or... weird. It was hard to recommend people one to use. OTP was regularly half-broken. Mobile didn't make things easier. Clients like Cisco Jabber being forced on people also didn't help (worst chat client, ever).
Larger servers like the CCC being actively hostile to their users also didn't help (e.g. they responded to user issues with "we can delete your account, it's federated, go somewhere else" - they framed this as wanting to reduce users on their server to help distribution along... but let's say you need a lot of context to understand that).
XMPP is dead because it could never find a way to properly serve general users.
I don't think it is anything like dead. However I agree with you in the main. Relying entireln hobby clients and volunteer run servers is not likely to lead to "mass market"
Okay, dead is a bit overstated.
I don’t fully agree with the hobby problem though. Adium was a really good client and by hobbyists.
I’m more about a certain trend in community cohesion. Like, it felt like no one would join an XMPP client project. They always started their own, only built for them.
That’s obviously reductionist, but the feeling I walked away with.
One of the major fail points for XMPP has been XEP for every feature, individual clients should support messages and pubsub (only a basic subset) but not all features of a protocol. Clients should serve as a brand for users to flock to like Whatsapp or Signal.
That's less of a problem now. There are profiles that tell you what XEPs your client needs to support to be a modern chat app on desktop or mobile.
In general the XEPs don't matter. As you say an app should build what features make sense for its use case. Of course comforming to any relevant spec as needed, but you don't start from the spec you start from the feature IMHO.
Fellow lobster ploum, I'll be at the XSF stand at FOSDEM Saturday, I'm secretly punk inside, and have a (hopefully) functional Signal/XMPP bridge to demo, come say hi! ;-)
I'm really glad to see more happening in this space to compliment things like Borogove SDK. The more open source software gets built on open standards like XMPP, the better.
What do you call that feeling where...Either you had a deep thought about a topic, or maybe even lightly dabbled with said topic...but then, step away for a little while....and then for some unknown reason, you start to see that same topic appear and bubble up in many places through your web surfing? (Yes, i said web surfing, I'm old, leave me alone! :-) Anyway, i played with xmpp (prosody) earlier in 2025 (which i had not played with xmpp since like 2015 or so?)....And then got lonely, didn't have bandwidth to play with more use-cases beyond simple chat, and then just stepped away, killing my xmpp server instance. And, now!?!? I'm seeing all manner of blog posts seemingly almost EVERYWHERE about xmpp. So, I'm getting that itch ever so slightly nudging me to play with xmpp again...But, I'll resist for now.
Still, it does feel like xmpp is being talked about a bit more than normal in the recent, say, 2 or 3 months. I'm sure its me, and that phenomenon that i can;t recall what its called....right? Or, is it that there is a legit, marked increase and interest in xmpp these days?
I know what you're talking about, and I can't think of the name either. It's like when you buy a blue Honda and you start seeing blue Hondas everywhere. They were always there, just like XMPP was always there ;-).
But it's probably a mix. I think some of the increase in awareness of XMPP is that people aren't happy with centralized options like Signal or heavier options like Matrix, but people are hungry for something, anything that is open and can be an alternative to services run by trillion-dollar companies. We're seeing the damage being done by the Apples and the Metas and we want out while still being able to talk to our family and friends.
Glad to see this XMPP has been needing a client with this kind of architecture for some time