115 comments

[ 2.4 ms ] story [ 71.5 ms ] thread
When you rank something with numbers, I’d love it to be more like 1 of 5 (1/5) even when you said it before. When you are reading at that line, I had to recalibrate, if that is 4 (out of 10), which is what most people try to rank against.
I'm sad to see XMPP missing from this list. I wonder if the author was simply unaware of it or simply ignored it.

IMO XMPP is technically superior to Matrix. It "only" needs a cross-platform high-quality, branded app àla Element. There's underlying protocol support for all the features: video/audio calls, group calls, threads and reactions. Maybe missing are custom emoji (I think?) and channel grouping (which is still in the works). And of course all these protocol features work fine with federation.

This does miss a major feature of Discord and why, imo, it got such an huge following among gamers at first: voice and video chat.

I've really had a hard time finding a Discord alternative that has the same kind of first-class voice and video chat support that Discord does. Back to Ventrilo and Mumble I guess /s

Great writeup! Looks like this is going to be relevant very soon.

> Tools do not make a culture; the people engaging on it do

Absolutely, but it's also important to keep in mind that the tool has a big impact on culture by virtue of what behaviors it encourages and what limitations it has. "The medium is the message" is very true here, so think carefully about which tool you hop onto.

I know I'm in the small minority of Discord users who mostly uses it as a voice chat room while gaming with my friends, but for that use case the best alternative I've found seems to be Mumble.

I recently set up a Mumble server on my home server and it seems great so far, was able to get my friends connected pretty easily. We'll see how the voice quality and latency compare to Discord.

Ironically, Signal actually ranks a -1 for privacy in this use. Presumably you're already using Signal and getting mainstream contacts to start using it too. You probably have a basic profile that at least includes your real name, and might also have your picture. Maybe you're even one of the 7 people in the world that use the Stories feature in it. Well good news, now all of that is also unconditionally available to anyone in any group you ever join, including any future changes you ever make to that info, unrevocably forever into the future.

Signal has a fun dark pattern where it unrevocably grants permissions for anyone you allow to contact you to see everything in your profile for the rest of time. It has only a single trust level with contacts effectively: full trust. This is unacceptable in any tool you use for online community, unless you exclusively use it for online community and can decline to provide any info in this full-trust level. Unfortunately Signal also makes very sure you can't have a second account, by tying your account to a phone number, and only allowing one Signal instance per mobile device.

Is Signal good? Yes, but only exclusively for communication with people you already trust.

EDIT: typos

Matrix is the only one that offers the killer feature of Discord, which is being able to join many communities from a single login.

Sadly Matrix has never had a good UX for me. IMO they spent too many complexity tokens on e2ee and there are simply not enough left.

Most Discord alternatives fail not on tech, but on polish.

Signal → private but bad for communities

Matrix → flexible but rough UX

XMPP → powerful but fragmented

Discord → centralized but frictionless

Users pick frictionless every time. We probably don’t need new apps or protocols we need a client that works well.

There is zero chance that the target users for Discord is going to try anything more complicated than Discord, so basically all the entries in this list. I recoil in horror thinking about me explaining Matrix even to the most tech savvy friends I use Discord with and I really really hope people would stop recommending it.
Zulip actually looks pretty good if they made it a little sexier.
can someone please try running the experiment of "but what if just forking&spinning up an OSS clone, scaling up to take in the migrants, acquire network effects, collect roughly same subscription revenue, but run on just, like, 10 people?"

Discord has a financially and politically vulnerable posture that is downstream of having to operate a very large team, raise funding, be exposed to investor market pressure. However, it is also one of the rare instances of successful consumer freemium subscription monetization. A clone does not have to pay the tuition of "what makes this specific space compelling, and want-to-pay-for"; it just have to _exists_, passively soaking up migrants from each platform shift.

ITT WTB 3rd place for my frens.

Sounds like you're proposing Element (EMS)
Maybe there's a niche for Valve/Steam to step into here. They already have your data, and many people use Discord for gaming related chat.
CurseForge (Overwolf?) tried with Curse Voice before Discord even existed, but it never made it big. Probably because Discord started taking over when Skype fell off (i.e. Discord had good timing, Curse Voice did not).
I'm looking at modern browser APIs and wondering why no one else is trying certain things.

getDeviceMedia and getUserMedia are very powerful these days. I haven't actually tested it but I believe a chromium browser would have no issue capturing the hw accelerated output from a game. You can pipe these media streams directly to WebRTC peers for playback on the other side. A server with a simple selective forwarding unit could enable larger scale meetings (100s of participants). All of this can happen in <1000 lines of JS and server code. Most of the heavy lifting happens in the browser engine. Concerns like automating browser permissions, global hot keys, etc. can be handled via electron or platform specific options like WebView2.

Mobile clients are a bit cursed right now. The best solution is to maintain a standard client in the app stores. Forcing everyone to sign their own mobile apps is way too much friction. And you do need native for this on mobile. Browser only / PWA has no chance in hell of providing a smooth UX on iOS or Android.

> Anyway this thing [Revolt/Stoat] is so far from being ready for prime time, I only include it here to call out the project. I wish them the best and hope for good things, especially since you can self-host the server. But a lack of stability and features prevent this from being useful for anything beyond experimentation. Maybe someday.

Curious what prompted this verdict. My only experience with Revolt/Stoat has been with the Handmade Cities instance, but said experience hasn't been anywhere near as bad as this writeup seems to suggest.

I read in another post that HMC forked from before the Stoat rebrand and are maintaining their own fork until the rebrand settles down, so that probably reduces the instability.
The title is about "Discord alternatives", the major core metric is:

> Functionality: can it do everything required of a platform for building, organizing, and sustaining a community?

Feels like these are two different things.

What I expect from a Discord alternative is text messaging, voice and video call with screensharing, both possible on community spaces and with personal contacts in a way that is extremely easy to setup.

I would love an updated KDX/Hotline server running an a RasPi or similar at home. This was a solved problem 20 years ago. The migration to online based platforms will always lead to network effects and enshittification, for very little gain.
One of the main problems in suggesting Discord alternatives is that Discord itself is an amalgamation of several apps.

For some a Discord alternative needs to be a voicecall, for others it‘s game streaming, and for others it‘s just a chat, a bulletinboard or newsgroup, while they never used the Voice features.

What most people here seem to forget, is once a social platform gains traction and especially attention from the main masses, it undoubtedly require checks and balances.

Predators, racism, gore, pedophilia, harassment, stalking and so on..

No matter how high you value security, these are matters that hurt real people today. If you attract the mainstream, you must deal with it.

Tradeoffs are made based on the likelihood and frequency of risks vs the perceived reward.

For example - we let people drive cars despite the frequency of deaths that happen in car accidents.

> signal for secrecy

What kind of secret system uses a phone number tied to your ID as a user name?

I still can't wrap my head at all as to why a chatting app needs my phone number. It's totally ridiculous.
What's the point of the licensing on these chat servers charging per-user when self hosting?