Every open source project that becomes a business eventually has a difficult week. Funding runs thin, or growth stalls, and somebody proposes moving one small thing behind a paywall. It is always something reasonable. It is rarely the last one.
The way to avoid that week is not good intentions. It is to design the business so the tempting move does not exist — and to say publicly where the line is while the project is still small enough that saying it costs nothing.
What is free, permanently
The Birkly core is free and open source at release, and the core is not a starter tier. It includes:
Collections, entries, fields and sub-collections. The templating language and engine. Server-side rendering from the project folder and the client script for existing sites. The admin. Users, roles and permissions. Media handling. Public form submissions and the approval queue. MCP — full read and approval-gated write. The Agent Guide.
That list is deliberate. Everything required to build, run and maintain a real site, and everything that would create lock-in if it were paid, is in the free core. You can run Birkly forever, on your own server, without sending us a euro or a request.
The test we use: if a feature being paid would make it harder for you to leave, that feature cannot be paid.
The Agent Guide is the clearest application of that test. It is the most distinctive thing we have built and the most obvious candidate for a premium tier. It is core, because a structured claims format that only paying customers can publish is a format nobody adopts — and because we would rather it become normal than become a revenue line.
What costs money
Hub plugins. Commerce (catalog, cart, Stripe checkout, orders, subscriptions, licence keys) is paid, currently from €29.90 a month. Automation — workflows with triggers, conditions and actions — is available and free. Events, Studio and the site packs are in development.
The logic is that these add capability rather than removing a restriction. A shop needs a payment pipeline that we have to build, maintain, and keep compliant as card rules change. That is ongoing work with an ongoing cost, and charging for it is honest. Nothing in Commerce gates your existing content: your products are entries in a collection, and they stay readable and exportable whether you are paying or not.
Sponsorship and donations. Companies that depend on the core can sponsor it, and individuals can donate. This is the part that funds the free core directly, and we would rather it grow than shrink relative to plugin revenue.
Why not the usual models
Not hosted-only. The obvious business is to run it for you and charge monthly. It is a good business. It also means your content lives on our infrastructure, which is precisely the arrangement Birkly exists to be an alternative to. We may offer hosting eventually as a convenience; it will never be the only way to run it.
Not open-core-with-the-good-parts-out. The common pattern is a genuinely free core that is missing one thing every serious user needs — usually SSO, roles, or audit logging — which converts the free tier into a demo. Roles and permissions are in our free core for exactly this reason.
Not usage-based on the API. Metering reads would make the agent story a metered product, and an Agent Guide you pay per query to serve is not infrastructure. It is a toll booth.
What would have to go wrong
Being honest about the risk is more useful than promising it away.
The failure mode for this model is that Hub revenue does not cover core development, and the core degrades quietly — not paywalled, just under-maintained, which is how most open source projects actually die. If that happens, the visible signal will be release cadence, and you will notice it before we announce it.
What we are committing to is narrower and checkable: the feature list above stays free and open source. If we ever need to move something out of it, we will have failed at the thing this project is for, and we would rather say that now, in writing, while it is easy to say.
Birkly is in alpha. The roadmap is public, the Hub catalog shows what is available and what is coming, and the sponsors page is where the funding side lives. If you want alpha access, join the waitlist.