Fluxer

Open-source community chat, honestly reviewed. What you actually get when you self-host a Discord alternative that launched mid-controversy.

TL;DR

• What it is: Open-source (AGPL-3.0) instant messaging and VoIP platform — Discord’s feature set, your server, your data [README][1].
• Who it’s for: Communities and friend groups who want to escape Discord’s data practices and Nitro paywalls, and technically-inclined operators willing to self-host. Not yet for non-technical founders who need a push-button deploy [README][1].
• Cost savings: Discord Nitro runs $9.99–$99.99/year per user for premium features. Fluxer self-hosted unlocks everything at VPS cost only — though self-hosting isn’t fully ready yet [README][4].
• Key strength: Cleanest Discord-like experience in the open-source space. Genuinely AGPL-licensed, no telemetry, content not used for AI training or advertising [4][README].
• Key weakness: Built by two full-time employees, launched under emergency conditions, and self-hosting is explicitly not ready yet. Multiple users report constant downtime on the main instance. The main instance subscription model (“Plutonium”) mirrors Discord’s gating strategy — the self-hosted path is the honest one [1][README].

---

What is Fluxer

Fluxer is a Discord-style chat platform — communities, channels, DMs, voice and video — released under the AGPL-3.0 license by a two-person Swedish company, Fluxer Platform AB [4][README]. The pitch is simple: everything Discord does, minus the corporate data practices, plus the option to run it on your own hardware.

The timing matters. The README opens with a candid admission: “I know it’s hard to resist, but please wait a little longer before you dive deep into the current codebase or try to set up self-hosting.” The project went viral earlier than planned, apparently accelerated by Discord’s February 2026 announcement that spooked communities looking for exits [README]. That context shapes everything about the current state of the tool.

As of this review, the project sits at 7,067 GitHub stars. The production instance serves 125,000+ users with a team of two. The GitHub repo is still largely closed — PRs aren’t enabled yet, and self-hosting documentation says “TBD” in the relevant section [README]. A full refactor is in progress and being tested live on the production instance.

The tech stack is not lightweight: TypeScript/Node.js on the backend, Hono as the web framework, Erlang/OTP for the real-time WebSocket gateway, React and Electron for the desktop client, Rust compiled to WebAssembly for performance-critical client code, and a backend data layer involving Cassandra and Redis [README][2]. That’s a serious distributed systems stack — impressive for a two-person team, and a warning sign for anyone expecting a $5/month VPS deployment in the near term.

The optional paid tier is called “Plutonium” — Fluxer’s equivalent of Discord Nitro — and it exists on the main hosted instance. Self-hosted deployments explicitly include no Plutonium traces, and the README states: “nothing is paywalled. You can still configure your own tiers and limits in the admin panel” [README].

---

Why people choose it

The short version from [1]: people want Discord without Discord. That’s the entire value proposition, and it’s resonating.

The AlternativeTo reviews [1] break into two camps with very little middle ground. The believers say things like “It’s like Discord, but without the age verification and data collection. It’s also FOSS!” and “One of the best alternatives to Discord.” The skeptics say things like “it has been made on a rush since Discord started to suck” and flag that the main instance has the exact same problems they were trying to escape — phone number verification for suspected bots, subscription features gated like Discord, and downtime [1].

Both camps are right. The self-hosted path is genuinely different: AGPL-3.0 licensed, full features unlocked, no Plutonium, no phone verification requirements [README][4]. The main hosted instance is a different story — it’s a Discord competitor running on a shoestring team, and the pressure shows.

One reviewer put the current state plainly: “Promising in theory but the way it has been built as a copy of Discord down to the infrastructure plagues it with constant downtime” [1]. Another: “I have a lot of hope for Fluxer, but being as new as it is it’s currently struggling a lot with downtime issues” [1].

What pulls people in despite this: the Terms of Service are meaningfully different from Discord’s. Fluxer’s ToS explicitly states your content is not used for advertising or AI training, material changes require 30 days’ notice and explicit re-agreement, and enforcement under the EU Digital Services Act includes reasons, evidence, and an appeal route — not the opaque account terminations Discord users report [4]. That’s not marketing copy; that’s legally binding language from a Swedish company subject to EU consumer law.

---

Features

Based on the README and website scrape:

Messaging:

• Real-time messaging with typing indicators, reactions, and threaded replies [README]
• Full Markdown support [homepage]
• Private DMs and group chats [homepage]
• Communities with text and voice channels organized into categories [README]
• File sharing, link previews, image and video attachments [README]
• GIF search via KLIPY integration [README]

Voice and video:

• Voice and video calls in communities and DMs [README]
• Screen sharing [homepage]
• Noise suppression and echo cancellation [homepage]
• Multi-device join, mute/deafen/camera controls [homepage]
• Powered by LiveKit SFU — the same open-source video infrastructure used by other serious projects [README]

Moderation and community management:

• Granular roles and permissions [homepage]
• Moderation actions and audit logs [homepage]
• Webhook support — GitHub, Sentry, Slack-compatible webhooks documented in the API [3]
• Bot support [homepage]

Search and navigation:

• Message history search with filters by user, date, and more [homepage]
• Command palette with keyboard shortcuts [homepage]

Customization:

• Custom emojis and stickers [README]
• Save images, videos, GIFs, and audio [homepage]
• Custom CSS themes [homepage]
• Compact mode [homepage]

Self-hosting infrastructure:

• AGPL-3.0 licensed [README]
• Admin panel for configuring tiers and limits on self-hosted instances [README]
• livekitctl helper for bootstrapping the LiveKit SFU [README]
• Feature flags per instance — voice, Stripe, SMS MFA, manual review mode all configurable [2]
• OAuth2/federation endpoints in the API (federation not yet enabled by default) [2]
• E2E encryption key infrastructure exists in the API spec [2] but not yet user-facing

What’s not there yet:

• Native mobile apps (listed as top priority, not shipped) [README]
• Federation (top priority, not shipped) [README]
• Full E2E encryption (infrastructure present, feature not released) [2]
• Open PRs and community contributions (not yet enabled) [README]

---

Pricing: SaaS vs self-hosted math

Fluxer main instance: The hosted instance uses a “Plutonium” subscription for premium features. Specific pricing is not publicly listed on the website at time of writing — the site directs users to register and join the beta. The Terms reference Plutonium as an “optional paid subscription” [4]. Feature gating mirrors Discord’s approach, which is the main complaint from skeptical reviewers [1].

Fluxer self-hosted:

• Software: $0 (AGPL-3.0) [README]
• All features unlocked, no Plutonium required [README]
• Infrastructure: your own hardware

The AGPL-3.0 license is worth understanding. Unlike MIT, AGPL requires that if you run a modified version as a network service, you must release your modifications. That’s a meaningful copyleft provision — you can’t quietly fork and commercialize it. For most self-hosters this doesn’t matter. For anyone building a commercial product on top, it matters a lot.

Discord for comparison:

• Free tier: basic features, Nitro content paywalled
• Nitro Basic: $3.99/month
• Nitro: $9.99/month ($99.99/year)
• Server Boosts for unlocking community perks: $4.99/month each

For a community that currently pays for multiple Nitro subscriptions and server boosts to access file size limits, custom emojis, and quality settings — self-hosted Fluxer eliminates all of that at VPS cost. The math works. The catch is that self-hosting isn’t ready yet.

---

Deployment reality check

The README is honest to a fault: self-hosting documentation currently reads “TBD” [README]. The refactor enabling easier self-hosting is in progress, being tested live on the 125,000-user production instance [README].

What the tech stack tells you about the self-hosting footprint: Cassandra (distributed database, not trivial to operate), Redis, and Erlang/OTP in the mix means this is not a “throw it on a $6 Hetzner VPS” deployment. Cassandra alone typically wants 4GB+ RAM per node and runs as a cluster for real reliability. The features.canonical list includes cassandra, docker, docker_compose, redis, and sqlite — which suggests multiple services in play.

The README explicitly says: “I’m aware the current stack isn’t very lightweight. I’m working on making self-hosting as straightforward as possible.” That’s the developer acknowledging the gap themselves.

What you can do today:

• Register on the main hosted instance (invite codes required during beta) [homepage]
• Join the Fluxer HQ community to follow development [homepage]
• Monitor the self-hosting documentation at docs.fluxer.app for when it becomes available [README]

What you cannot do yet:

• Deploy a production-ready self-hosted instance with documented steps
• Contribute code via pull requests (PRs not yet enabled)
• Use native mobile apps
• Enable federation between instances

The livekitctl deployment helper for voice/video is available in the repo, which suggests the voice infrastructure is ahead of the main deployment docs [README].

If self-hosting today is a hard requirement, Fluxer is not the right choice yet. Check back in a few months. The developer has been transparent about the timeline and the reasons for the delay.

---

Pros and cons

Pros

• Genuinely AGPL-3.0. Not “source available,” not fair-code — proper copyleft open source. The codebase is in the open and self-hosted instances can configure everything including premium tier features themselves [README][4].
• Honest terms of service. No content used for AI training or advertising, EU DSA-compliant enforcement with appeal rights, 30-day notice on material changes [4]. These are structural protections, not just promises.
• Discord-familiar UX. The learning curve for users migrating from Discord is essentially zero. Communities, channels, DMs, reactions, custom emojis — it’s all there and it works the way you expect [README][homepage].
• Serious voice infrastructure. LiveKit-powered voice and video with screen sharing, noise suppression, and echo cancellation is not a half-implemented feature [README].
• Webhooks that work. GitHub, Sentry, Slack-compatible webhook formats are documented and functional — real bot and integration support [3].
• Federation and E2E encryption on the roadmap with infrastructure already started. The API spec already includes federation endpoints and E2E key infrastructure, which means these aren’t vaporware [2].
• Developer is communicative. The README is unusually candid about what’s wrong and what’s coming. Multiple AlternativeTo reviewers contrast this positively with other Discord alternatives [1].

Cons

• Self-hosting is not ready. The documentation says “TBD.” This is the single biggest blocker for the target audience of this site [README].
• Heavy tech stack. Cassandra + Redis + Erlang/OTP + LiveKit means operator complexity. Not a weekend project [README].
• Two full-time employees. The entire production service for 125,000 users runs on two people. That’s admirable, but it’s also a risk concentration that shows in the downtime reports [1][README].
• Constant downtime on main instance. Multiple AlternativeTo reviewers flag this explicitly: “constant downtime”, “pretty slow”, “struggling a lot with downtime issues” [1]. Some of this is growth pain, some is the refactor running in production, all of it is real.
• Main instance subscription mirrors Discord’s problems. Phone number verification for suspected bots, Plutonium gating — the hosted instance hasn’t escaped the patterns it’s supposedly competing against [1].
• No mobile apps. Listed as top priority. Not shipped. A Discord alternative without mobile is missing most of where Discord usage happens [README].
• No federation yet. Planned, infrastructure exists in the API, but not yet enabled [2][README].
• PRs not yet open. You can’t contribute code yet. The community can’t accelerate development [README].
• Potential “vibecoded” concern. One AlternativeTo reviewer [1] flagged that the speed of development and some architectural choices suggest AI-assisted coding at pace. The README’s framing (“Holy smokes, what a ride”) doesn’t contradict this read. Whether this matters depends on your risk tolerance for production software.

---

Who should use this / who shouldn’t

Use Fluxer’s main instance if:

• You want to try a Discord alternative today without running your own server.
• Your community is small enough that Nitro perks aren’t critical.
• You’re willing to tolerate beta-quality uptime in exchange for better data practices.
• You want to get familiar with the platform before self-hosting becomes available.

Wait for self-hosting if:

• You want to run your own instance — the docs aren’t ready and the stack is non-trivial. Come back in a few months.
• You need guaranteed uptime for an active community.

Don’t use Fluxer if:

• Mobile is critical to your community — there are no native apps yet.
• You need federation between instances — it’s not live.
• You’re a non-technical founder who needs push-button deployment — that’s not the current state.
• Your community is risk-averse and needs a stable, proven platform.

Consider a mature alternative if:

• You need self-hosting today with documented steps and a stable stack.
• You’re running a business-critical internal communication tool.

---

Alternatives worth considering

• Revolt.chat — the closest comparison. AGPL-3.0 licensed, self-hostable, Discord-like, further along on self-hosting documentation, smaller community, less polished voice. Worth comparing directly.
• Matrix/Element — federated by design, genuinely decentralized, much steeper learning curve, bridges to other protocols, the right choice if federation is a non-negotiable requirement.
• Rocket.Chat — mature, well-documented self-hosting, business-oriented, not Discord-like in UX, better for internal team chat than community.
• Mattermost — team messaging, not community chat. Stable, open-source core, good self-hosting docs. Wrong use case for friend groups or gaming communities.
• Zulip — threaded conversations, open source, solid self-hosting, different mental model than Discord. Excellent for async technical teams, wrong for casual communities.
• Discord — the incumbent. Best apps, most integrations, most polished, closed source, data practices that are driving people to look at this list in the first place.

For communities specifically leaving Discord over data concerns or Nitro costs, the realistic shortlist is Fluxer vs Revolt.chat vs Matrix. Fluxer wins on UX familiarity and has the most momentum right now, but Revolt is more deploy-ready today. Matrix wins if you care about federation as a guarantee rather than a roadmap item.

---

Bottom line

Fluxer is the most promising Discord alternative in the open-source space right now, and also the most honestly unfinished. The developer hasn’t overpromised — the README is direct about what’s broken, what’s coming, and why the timing went sideways. The AGPL license is real, the terms are meaningfully better than Discord’s, and the feature set for messaging and voice is already solid.

The problem is that the thing most people on this site want — self-hosting — isn’t ready. The stack is heavy, the docs say TBD, and the refactor that enables community contributions is still in progress on a live production system run by two people. For a community ready to move off Discord today, Revolt.chat or Matrix is the pragmatic choice. For communities willing to wait six to twelve months and watch a well-intentioned project mature into something deployable, Fluxer deserves a spot on your list.

If and when self-hosting drops with clean documentation, the math becomes obvious: your own hardware, AGPL code, all features unlocked, no Plutonium. That’s the version worth running. Just not yet.

---

Sources

1. AlternativeTo — Fluxer: Open-source instant messaging, VoIP, group chat, self-hosting (8 reviews, 3.8 rating). https://alternativeto.net/software/fluxer/about/
2. Fluxer Documentation — Instance API Reference (WellKnownFluxerResponse, feature flags, federation endpoints). https://docs.fluxer.app/resources/instance
3. Fluxer Documentation — Webhooks Reference (GitHub, Sentry, Slack webhook formats). https://docs.fluxer.app/resources/webhooks
4. Fluxer — Terms of Service (Effective date: 2026-04-02, Fluxer Platform AB). https://fluxer.app/terms

Primary sources:

• GitHub repository and README: https://github.com/fluxerapp/fluxer (7,067 stars, AGPL-3.0 license)
• Official website: https://fluxer.app
• Documentation: https://docs.fluxer.app
• Launch blog post: https://blog.fluxer.app/how-i-built-fluxer-a-discord-like-chat-app/
• Roadmap: https://blog.fluxer.app/roadmap-2026/