In Search of a Discord Replacement

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.

Requirements

My 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-Starters

Let'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.

Matrix

Barring 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.

EDIT: Not even 24 hours after this post was published, a scathing report came out detailing multiple side-channel vulnerabilities in Matrix.org's Vodozemac library powering Matrix's E2EE feature. I'm no cryptography expert and cannot evaluate the finer technical claims of the article, but the way Matrix.org handled this report on a person-to-person basis is deeply unprofessional. It's really disappointing to see, and is pushing me even further away from it as a platform.

Signal

Signal 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 channel of discussion and a call will notify everyone in the chat. Signal also comes with a high quality and accessible mobile app, and is arguably developed with a mobile-first approach.

Signal is stunningly solid option for small groups and one-to-one chats, but I would caution against building any particularly large community on the platform. Signal, by default, supports up to 1,000 users in one group chat, but anyone using Discord right now knows how unruly that can be.

Fluxer.app

Fluxer.app is a rather promising option, directly aiming to replace Discord. It's clearly developed with the intention of being a direct replacement for Discord's service, and has monetization built in from the get go. It seemingly has all of the features we expect out of Discord out of the box, and is free and open source software under AGPLv3, but currently not accepting any PRs due to an ongoing refactor.

The project is really new, only being announced and open sourced within the last month. The entire project is 1.5 million lines of code and there is no git history prior to open sourcing. The author claims that they did not vibecode it and they did the "normal" practice of squashing history when making an initial release, but I feel like that is just a plausibly deniable excuse, and will need to wait and see if they actually follow through with it. They're currently not accepting that many PRs due to a "ongoing codebase-wide refactor", and, for the same reasons, self hosting instructions are currently missing as a whole.

One potential concern is that Fluxer.app is licensed under AGPLv3 with a CLA. This isn't an uncommon combination to see used in FOSS rugpulls where the owner of a project takes it closed source, typically forcing a community fork in the process. This isn't usually a hard dealbreaker, but I would rather not risk the headache of a secondary migration from losing trust yet again.

EDIT 2026/02/18: The author has published the refactor, which is in the open, and may soon be accepting PRs. They also removed the CLA requirement. This is great to hear!

Sharkord

Sharkord is a relatively new FOSS project with the aim to be a direct Discord replacement. It supports basic text chat, voice chat, video chat, and no-audio screensharing at this current time. It was relatively easy to self host.

However, due to how new it is, it's currently riddled with bugs, and currently does not have a mobile app or PWA support. It's unclear whether this is the result of vibecoding or just general developer incompetence. Given the author's comments on use of AI for the interface and their clearly AI generated profile pic on GitHub, there's a high likelihood that a non-insignificant amount of the app was vibecoded and calls into question if they can be trusted to actually maintain the app long term.

Stoat (formerly Revolt)

Stoat was the option I tried long before any of this chaos, and honestly the one I'm rooting for the most. It aims to be a direct replacement for Discord. It's FOSS and is self-hostable. It supports text chat, and previously supported voice chat until they decided to rewrite the backend for audio. It's never supported video chat or screensharing, and does not have a usable mobile app yet.

It's also set up such that self-hosting includes infrastructure for all of Discord. You can have the equivalent of multiple Discord servers on the same self-hosted instance. This poses some extra headache for the admins, as users can hide discussions that are out of reach of moderators or system administrators.

Unfortunately, the development pace has not kept up and many of the missing fundamental features are not currently present.

Spacebar (formerly FOSScord)

Spacebar is a project trying to create a 1:1 version of of Discord down to the API calls being made. They have a custom client and server that can work with each other or be used with Discord's official API. It's all FOSS and self-hostable.

However, due to its nature in attempting to replicate Discord 1:1, it lacks a mobile app. At time of writing, it seems like Spacebar does not support anything tied to a voice channel in Discord.

XMPP (with Fluux Messenger)

Fluux Messenger from ProcessOne is a XMPP frontend that looks awfully like Discord. Any backing XMPP server can be used with it, such as ProcessOne's ejabberd. This should cover text chat like Discord fairly well, but will require additional effort to support an integration with Jitsi for voice, video, and screenshare.

Fluux Messenger also has the same licensing issues as Fluxer.app. It's also AGPLv3 with a CLA, but thankfully XMPP is an open protocol so any adequate frontend can be used as a substitute.

Mumble

Mumble is a FOSS, self-hostable, VoIP software and is made to be close to Teamspeak in UX. It's created with a desktop use case in mind but there are compatible third party mobile apps. Its voice channels are structured close to that of Discord's, where users can hop in and out at any time, and the channels exist independently from any single given call. It's a voice chat-first application with some level of support for text chat for those in a voice channel. Joining a server requires actively joining a voice channel. Unlike Discord, the text chat does not have long term text chat history. I think this is appropriate for a text chat that is purely supplemental to a voice chat, but that It does not support video chat or screensharing.

Zulip (w/ Jitsi)

Zulip is a fully open source FOSS chat application that can be self hosted. What differentiates itself is the high number of integrations it has with other services. Most notably, it has a frontend integration with Jitsi, which support voice, video and screenshare. It has a usable mobile app.

As a text chat, it's probably the most complete and solid experience out of all of the options in this list. Zulip's threading model, or "topics" as they're called in Zulip vernacular, take getting used to, as every conversation is in a topic, and each channel does not have a top level discussion that can be branched from.

The Jitsi integration as it currently stands does not act like Discord's voice channels, where users can hop in and out at any time, but are tied to established calls tied to messages sent in a text channel. This might be viable in a productivity oriented community like FOSS development groups, but this may not be a good fit for large general purposes communities.

Conclusion

Overall, I'm not too happy with the state of the viability of many of these alternatives. Many options have been around for 10+ years, and still do not come close to the convinence provided by Discord.

Thus far, out of all of the options I've reviewed, Zulip, with Jitsi, might be the best immediate fit for the requirements outlined in this post. It seems like it ticks all of the boxes for a usable UX, solid self-hosting support, has a usable set of cross-platform clients, and seems to support all of the immediate fundamental features most Discord communities are used to. Its model isn't a great 1:1 fit with what users expect out of a Discord server and its interface is definitely oriented for use as a business oriented chat, but, given enough time, I think most users could become accustomed to its use.

Fluxer.app might be a viable alternative in the future, as its feature set aims to directly compete with that of Discord's. However, the intent on monetization and lack of FOSS engagement, and poor self-hosting support. and questionable licensing decisions casts doubt on its long term viability. In the case of a FOSS rugpull (a la Emby vs Jellyfin), I'm sure the community will be able to pick up the slack eventually, but that may be way too far out to consider for now, and I would rather not do a secondary migration in the case of such an event. EDIT 2026-02-18: They've dropped the CLA!

That's it for now. As I try more alternatives, I'll be sure to update this post with my observations. I don't think it's practical to migrate very large communities at the current moment, but keeping an alternative around in case shit really hits the fan may be a viable migration path forward.

Acknowledgements

This blog post was researched and edited by the following members of the discord.gg/touhou community: Shio, ReUKing, Great Sage, Jordy, Mami, Photographicrat, and Ember.

Appendix: Summary Table

Tags: Discord, Open Source, Privacy, Self-hosted
RSS Feed | Permalink