Why I Built Stems

Here's something I've never understood: the npm packages that power your app get paid nothing.

A song used in a Spotify playlist generates a fraction of a cent per play for the songwriter. That's a small number, but it's a real number, because there's a market mechanism that moves money from the consumer to the creator.

An npm package used in a billion-dollar SaaS product generates zero. The author maintains it, handles issues, fixes security bugs, and pays their hosting bill out of pocket. This is considered normal. Open source is the gift economy that the rest of the software industry is built on top of, and we've mostly decided to leave it that way.

I think it's an accident of history that we decided software components don't deserve royalties and music does. Both are creative work. Both are used over and over. Only one of them has a mechanism for compensating the people who made them.

Stems is an attempt to build that mechanism for web apps.

The model is simple. You build an app using components from the Stems registry. When you deploy it and pay for hosting, 30% of that net hosting revenue pools. The pool splits between the component authors whose work is in your app, weighted by the type of component they contributed—a payment flow earns more weight than a footer, because it represents more of the value being delivered. They see it accumulate. Eventually they withdraw it.

The weight schedule is platform-set and permanent, not author-declared. That's the integrity constraint. If authors could set their own weights, there'd be an arms race. The hosting-funded pool is the other constraint: the only way to game the system is to pay more in hosting than you'd earn in royalties, which means self-dealing is a guaranteed money-loser by construction.

The honest uncertainty is demand. The math only works at scale—a hundred-app fleet doesn't clear a meaningful bar for most authors. The supply side is fleet-gated: at around a thousand paid apps, representative components start clearing a "worth authoring" threshold. Below that, cash bounties for founding authors bridge the gap.

Whether indie hackers will pay $29 a month to host an app they assembled from components is a question I can only answer by running the experiment. I've done enough of the math to think the model is sound. I haven't done enough of the selling to know if anyone will show up.

I built Stems because I wanted to see if a royalty economy for the web was actually possible. The answer is still "maybe." But I've ruled out "obviously impossible," which is where I started.