What Is a Self-Hosted Community Platform (And Who Should Actually Use One)

A self-hosted community platform is community software that runs on infrastructure you control instead of a vendor's multi-tenant cloud. You hold the database. Nobody can suspend it, reprice it, or sunset it out from under you, because there is no landlord in the picture.

That is the definition. The rest of this guide is what the definition costs, where its edges blur, and who should actually want it. I have lived on all three sides of this question, renting platforms, running my own products, and now selling one, so I will try to be the guide I could not find when I started.

The spectrum of owning a community

“Self-hosted” gets used as a binary, and it is not one. There are four bands, and most bad platform decisions come from not knowing which band you are shopping in.

Band one is where most creator communities live. Band three is where most self-hosting guides point. This post spends real time on band four, because it exists precisely for the people who want out of band one and take one look at band three's job description.

Where Seedly Communities honestly fits

Before I pitch anything, the honesty block. I sell Seedly Communities, an owned-source platform. You buy the full source code once, $399, with a commercial license. The reference deployment runs the frontend on Vercel and the backend on Convex, a managed database and functions service, plus your own Stripe account for member payments and optional keys for live video and email. The phrase I use for that is source you own, deployed on infrastructure you control.

What I will not call it is self-hosted in the strictest sense, because the backend runs on a managed cloud service rather than a box you administer. You control the accounts, the data, the domain, and the code. You do not rack the server. If your definition of self-hosted is strictly no-managed-services, the honest recommendations are Discourse or Flarum, and I mean that. This paragraph exists because the category is full of vendors fudging exactly this distinction, and I would rather rank lower than fudge it.

Why people go self-hosted at all

Four drivers show up over and over, each with receipts from the last couple of years.

• -Deplatforming risk. In June 2025 Facebook suspended thousands of groups overnight and called it a technical error. Mighty Networks hosts in the UK found themselves locked out of their own networks during Online Safety Act compliance, without an admin export route first. When the platform is the landlord, eviction is always on the table, even by accident.
• -Fee fatigue. Skool takes 10% on its Hobby plan. Whop takes an effective ~5.7% plus 30 cents. Kajabi added a fee of 0.5% to 5% just for using a third-party payment processor. Percentages scale with your success, and the five-year math is grim.
• -Price-hike whiplash. Bettermode moved legacy customers from $50 to $399 a month. Kajabi restructured in January 2026 with 20 to 25% hikes and no grandfathering. Circle's monthly Professional price went from $89 to $129. Rent is a number someone else controls.
• -Data lock-in. Skool offers no content export at all, members CSV only. Nas.io has no documented export path we could find. The export comparison tests every major exit door, and most are painted on.

What self-hosting really costs

Now the free-as-in-puppy section. Open-source community software is free the way a puppy is free. Here is the real bill for band three, with numbers instead of vibes.

Discourse, the best of the breed, runs on a VPS at roughly $13 to $45 a month once you add mandatory transactional email and backups, call it $160 to $540 a year. Budget a realistic 1 to 3 hours a month of maintenance, plus the occasional multi-hour incident when an upgrade fails mid-rebuild. Email is the wall people hit first. Discourse's own docs warn that email server setup and maintenance is very difficult even for experienced system administrators, which is why their install guide simply tells you to use a service. Flarum wants Composer on a command line to manage extensions, with paid extensions maintained by separate developers. NodeBB brings a heavier Node plus MongoDB or Redis stack and its own history of plugins breaking across upgrades.

And none of the three ships creator monetization. Payments, subscriptions, courses, and events are plugin archaeology or simply absent. If your community is a forum, band three is wonderful. If your community is a business, band three is a second job.

Open source is free-as-in-puppy. SaaS is easy-as-in-rent. Most guides pretend those are your only two options.

What SaaS is genuinely better at

The concession section, because it is true. Rented platforms give you instant start, managed uptime, security patching you never think about, and someone to email at 2am. If your community is brand new and you are still testing whether anyone shows up, rent first. Prove the community, then decide what it should run on. Nobody should take on infrastructure, even friendly infrastructure, for an idea that is three weeks old.

The middle path, owned source

Between free-with-a-sysadmin-job and easy-with-a-landlord there is a gap, and owned source is what fills it. You pay once for product-grade software, deploy it on infrastructure you control, and keep the creator features that open-source forums never ship.

Seedly Communities is my worked example, because it is the one I can vouch for line by line. $399 once, or $150 as an add-on when bought together with Seedly CRM in one checkout. Feed, chat, classroom, events, live streaming, and gamification in one codebase. 0% transaction fees because members pay through your own Stripe. Unlimited members, a claim I can make because no cap exists anywhere in the code. A guided deploy onto your own Vercel and Convex accounts, plus your own domain and branding. And I run my own community on it, live, which is a proof point no amount of copy can fake. The full feature map is on the platform page.

It is not the only occupant of the band. phpFox has sold owned-source community software for years, and FluentCommunity sells lifetime deals in the WordPress world. I name them because a category with only one vendor in it is usually a vendor inventing a category. This one is real, just underpopulated.

A decision framework, six questions

Score yourself honestly. Count your yes answers.

• -Is business-critical revenue running through the community today?
• -Would a surprise ban or lockout materially damage the business?
• -Is a percentage fee material at your revenue level?
• -Can you deploy a modern web app, or follow a written guide with copy-paste steps?
• -Can you live without a branded mobile app on day one?
• -Is the community your product, rather than just your marketing channel?

Zero to two yes answers, rent. Pick Skool or Circle, enjoy the zero setup, revisit in a year. Three or more, you are the person this category exists for, and the fork is technical comfort. If you have real sysadmin capacity and your community is discussion-shaped, run Discourse and never look back. If you want the ownership without the ops career, buy the source. That is the honest sales pitch, all of it.

FAQ

Wherever you land, test the exit before you commit. That is the one-sentence version of the export comparison, and it applies to my product exactly as much as anyone's.