The most honest open source fee runs on goodwill, and comes with a legal way out.
Someone finally built an honest way to make companies pay for open source. Honesty is the whole model, and its single biggest risk.
Open source has one root problem: you cannot charge for something anyone is free to take. The code that runs your bank, your hospital and your tax office sits with people who are rarely paid to keep it working. It still startles me how few hands hold up the tools everyone depends on, and how little of what we take ever reaches them.
A model that closes the gap without breaking the openness deserves a fair hearing. The Open Source Maintenance Fee, introduced in 2025 by the maintainer behind the WiX Toolset, a tool for building Windows installers (not the website builder), does something refreshingly plain. The source code stays completely free, for anyone, forever. Only the finished, ready-to-install version, the one your developers install in a single command, carries a fee, and only once your company is past a modest revenue line. After years of licences that quietly change under everyone's feet, that honesty is the best thing about it.
It is also the thing the whole model rests on. The fee is not a charge for the code. It is a charge for convenience, and the model's own guide tells you how to skip it: take the free source and build the finished version yourself. That door is deliberate. Keeping the source free is what makes it open source at all; close the door and it isn't. So the model asks companies to pay when they have a sanctioned, legal way not to. It runs on goodwill. That is the most honest thing about it, and the most fragile.
And goodwill is not asked evenly, because the cost of skipping is not either. A small company usually finds it cheaper to pay than to build the finished version itself, so it pays. A large company, the kind the fee targets, already runs the engineering to add one build step, so skipping is a trivial, one-time cost against an otherwise monthly bill. The burden ends up inverse to the ability to carry it: lightest on those best able to pay, heaviest on those least able. Nobody designed it that way. It falls out of charging for convenience, the one thing a well-resourced team can build its way around.
Look at who has taken the fee up, and the pattern repeats. The adopters so far are mostly small projects. The large, widely-used projects, the ones whose funding is the real emergency, are not on the list.
OSMF is one answer among a growing number, and that is the hopeful part: the funding gap is finally being treated as a problem to solve, not a fact to live with. The open question is which model can carry the real weight. Not just the famous projects that can draw sponsors, but the thousands of unglamorous ones beneath them, the code quietly holding up everything you run, and without trading away the openness that made them worth funding at all. That balance is the thing no one has struck yet, and this experiment shows how hard it is.
