In Search of a Discord Replacement

Hacker News · Feb 18, 2026 · Collected from RSS

Full Article

I've really liked Discord. I'm one of the early adopters from 2016 and enthusiastically convinced my friends and communities to move to it. Since then, I've developed a moderation bot. I helped build multiple communities of 10,000+ users. I've made lifelong friends, helped build FOSS projects, and organized events and grassroots movements. For the past 10 years, Discord has been extremely useful and borderline irreplacable throughout many facets of my life.However, since the start of the pandemic, my friends and I have watched the slow but inevitable degradation of service, unfixed bugs, and watching the company deseperately flail about trying to achieve profitability. Each time we'd encounter some kind of bug, the voice chat I frequent would be flooded with cries of "We've GOT to get off Discord." Each time this happens, I'd look into alterantives, maybe try one out, but never really pull the trigger on migrating to an alternative.On Febuary 9th 2026, Discord announced that it would be universally requiring age verification. This follows a series of laws passed in various jurisdictions around the world requiring age verification for access to potentially sensitive content. This has required most online platforms to partner with third party vendors for verifying users' government IDs... which have been leaked en masse. In their infinite wisdom greed, Discord, decided to enforce this requirement universally instead of just the jursidiction where it's legally required.Many users, myself included, view this as a major breach of trust and the final stage of years of gradual enshittification. It's lit a fire under our ass to finally figure out a migration plan away from Discord. Documented below are my general findings and suitability for the communities I help run.RequirementsMy review and analysis is only for the communities I help run, and while I think the requirements are similar for most communities currently on Discord, they may not hold for everyone. There are three types of communities I participate in and help run:Large centralized online communities - I help run discord.gg/touhou, one of the largest Touhou fan communities on Discord. At time of writing, there are 55,000+ members in the server, with more than 2,000 active weekly active users. Orbiting it are a number of smaller targetted communities focused on different facets of the community, such as Touhou Wiki, Gensokyo Radio, and various other niche communities. These tend to be a central gathering space for those in the community without any one particular central goal in mind, often acting as just an online third place where members of a shared interest community can come together and openly discuss various topics. Many of these make heavy use of Discord's voice channels, video chat, and screensharing. These servers have a hard requirement for adequate moderation tools for dealing with any bad actor willing to join the community.Large productivity oriented communities - This includes the Bevy Discord and various FOSS development related servers. While these also act as a social gathering place, there is also a targetted productivitiy oriented goal associated with these servers. These servers also have a hard requirement for adequate moderation tools for dealing with any bad actor willing to join the community.Small friend group chats - These servers tend to be much smaller, topping off at only a few tens or hundreds of members. Unlike the others, most of the members of these groups already have a decent understanding of the other people in the server, and moderation tools are thus not a hard requirement.The following are what I would consider hard requirements that are non-negotiable:Discord-like Onboarding UX - The greatest issue with bulding community comes from the experience of new users.Multiple text and voice chat channels - The communitiies I help run need to support thousands of users there needs to be a way to support organized threaded chats.Decent Mobile Experience - A mobile app shouldn't be required, and a progressive web app (PWA) might be acceptable in the interim, but allowing users to access the chat from a mobile phone is a strict requirement for community engagement nowadays.Self-hostable - The laws that prompted the need for age verification still apply in target jurisdictions, and relying on large platforms that are easy targets for lawsuits will just result in the community hopping from platform to platform.There is also a short list of "nice to have" features:Free and Open Source (FOSS) - This relates heavily to self-hosting, but having direct source code access helps the community contribute back and build a better experience together.Minimal Vibecoding - In the wake of Discord's announcement, there's been a glut of opportunistic vibecoders trying to steal marketshare. While the vast majority of them can be easily ruled out, there's a small handful of projects that have some level of plausible deniabilty. While vibecoding isn't a hard dealbreaker, it poses a serious question of whether the application can be trusted to be properly maintained into the indefinite future.End-to-end Encryption - Privacy is important! It would be nice to have 1:1 DMs and small group chats support end-to-end encryption. Supporting it for large communities of thousands of users might not be the best idea though.TTL-ed messages - In line with the prior point about E2EE is the support for TTL'ed messages. Some of the servers I'm in are up to 10 years old, and have discussions that should have long been discarded. A TTL that is too short (i.e. 14 days) may be untenable, but a decently long TTL of a year or more might be acceptable.Screensharing and video chat - Some of the communities I help run are heavily reliant on screensharing for shared activities (i.e. group watches, group gaming, raiding, etc.) It would be nice, though not strictly required, to support screensharing and video chat.Non-StartersLet's start off by ruling out some major players in the space that are strict non-starters:Good ol' IRC - The onboarding experience is not well suited for modern communities. Needing to have someone set up an IRC bouncer for non-technical users is already a non-starter.Telegram - Cannot be self hosted. Ties to the Russian government make it a serious security concern. Not FOSS.WhatsApp - Cannnot be self-hosted, does not support multiple text and voice chat in large communities. Not FOSS. Meta's ownership also raises more privacy concerns than it resolves.QQ, WeChat - Cannot be self-hosted. Ties to the Chinese government make it a serious security concern. Not FOSS.LINE - Cannot be self-hosted, does not support multiple text and voice chat in large communities. Not FOSS.Steam Chat - Not self-hostable. Not FOSS. All chat has a very low TTL of only two weeks. It has text and voice chat and screensharing, but does not have video chat.Slack, Microsoft Teams, Google Chat - Cannot be self-hosted. Not well suited for large communities of thousands of users due to per-user monthly costs. Limited message history on free edition. Not FOSS.Rocket.chat - Self-hostable, FOSS, has a rather nice feature set, but voice calls are only available for enterprise customers??? Free version limited to only 50 users. Need to contact support to enable more.Teamspeak - Self-hostable, not FOSS. Has a mobile app. Free license is limited to only 32 users. 1024 users for $500/year.Root - Not open source. Not self-hostable. Does not have a publicly accessible mobile app.Guilded.gg - Not open source. Not self-hostable. Never had a mobile app? Bought out by the pedophile platform, Roblox. No longer exists.Echoed.gg - Not open source. Not self-hostable.A multitude of opportunistic vibecoding asshats trying to stake out a name for themselves.MatrixBarring IRC and XMPP, Matrix is probably been one of the longest lived FOSS chat options out there. Its 2.0 featureset seems to match up with everything we want and more.However, self-hosting Matrix is a genuine pain in the ass. Synapse is a resource guzzler beyond compare. An idle single-user instance uses 4GB of RAM. Its replacement, Dendrite, is simultaneously vaporware and maintainence mode. Dendrite's replacement, Synapse Pro, isn't really self-hostable and is closed source. Arguably made just to grift EU governments out of their money.The true dealbreaker, though, has to be the new user experience. We tried registering multiple users onto a self-hosted instance, and found that all clients but Element's latest and greatest failed to support all of the features described in the protocol. The end-to-end encryption periodically broke and rendered chat history completely inaccessible.We also tried the more resource efficient home server implementations like Conduit and its fork Conduwuit, but that exacerbated the combinatorial feature support woes even more. Instead of just N different clients having problems with one home server, we now had M different server implementations combined with said N clients for a whole N-by-M compatibility matrix of problems. All of this just so a brand new user to get started chatting with their friends. Apparently there's now Tuwunel and Continuwuity, but I'm too burnt out on Matrix as a whole to try to bark up that tree.Matrix has been around for 10+ years and has had major corporate backing, yet its UX still feels worse than any alpha-quality options on this list, including some of the clearly vibecoded projects that have been recently announced. This is truly an abysmal dogshit state of affairs for such a platform. It's almost like the devs at Element and the greater Matrix community wants the experience to remain as broken as it is now.SignalSignal is the E2EE chat app and protocol. It's FOSS software, and can be self-hosted (though with much difficulty). It supports text, voice, and video chat, and the desktop option supports screenshare, though it's closer to Skype's older model where there is a single chan