People see a GitHub profile with dozens of repos and assume one of two things. Either it’s a hobby that costs nothing, or there’s some money machine behind it they can’t see.

Neither, at least for me. So I opened the spreadsheet. Everything below comes from my own project page and my repos, counted in mid-October 2026.

the numbers

  • 57 projects across 9 categories: Mantine extensions, React components, macOS apps, CLI tools, WordPress plugins, templates, Raycast extensions, and a few arcade games.
  • 775 GitHub stars, total.
  • 35 releases in the last 30 days. Not a typo. That’s more than one a day.
  • About 20 hours a week, in the evenings after work and at weekends.

where the stars actually are

775 sounds like a lot spread over 57 projects. It isn’t spread at all.

Four projects hold 374 of them, 48%: WP Bones (172), Mantine Split Pane (78), octoscope (69) and Mantine Onboarding Tour (55). The 26 Mantine components together have 411, more than half of everything.

And here’s the part I find funny, in a slightly painful way. My three macOS apps, the ones I spend most of my time on right now, have 35 stars combined. FinderGit 18, Netfox 15, Lancetta 2.

FinderGit alone went through 68 tagged versions between mid-April and early October. Lancetta shipped 32 releases in its first three weeks. That’s where my evenings go. 35 stars.

Stars don’t measure work. They measure how many developers found the thing while browsing GitHub. A Mantine component is found by Mantine developers, who are on GitHub all day. A Mac app is found by Mac users, who mostly aren’t.

the money

Let’s do the income side first, it’s quicker.

  • GitHub Sponsors: one recurring monthly sponsor. Thank you, you know who you are.
  • Donations: nothing so far.
  • The apps themselves: free. All of them.

Now the costs.

  • AI tools: Claude Max on the 20x plan, plus ChatGPT Plus. This is the biggest line by far, and it’s what runs the whole studio: a set of agents, my “employees” as I like to call them, that help me maintain every project. Figuring out how to use them well is an ongoing experiment, and honestly half the fun.
  • Apple Developer Program: the yearly fee. Without it you can’t sign and notarise a Mac app, and an unsigned app is an app nobody installs.
  • Domains: one for each Mac app that has its own site, FinderGit, Netfox and Lancetta.
  • Hosting: a Vercel plan. The project page itself is on GitHub Pages, so that part is free.
  • Writing: including this. Right now Medium costs me more than it pays me.
  • Time: those 20 hours a week. The one cost I never put a number on.

You can do the subtraction yourself. tbh I did it once and then stopped doing it.

so why keep going?

Because the subtraction is the wrong maths. What I get back doesn’t go in that column.

I build what I need first. FinderGit exists because I was tired of tabbing between Finder and a Git client. Lancetta exists because I wanted a gauge while working with agents: hourly and daily token use at a glance, so I know when to slow down and when I can push. If nobody else ever used them, they’d still have paid for themselves in my own week. That someone else finds them useful too is the bonus, and it’s a good one.

The community makes them better than I would alone. The feedback, the bug reports, the “have you thought about…” from people I’ve never met. A lot of what FinderGit and Netfox do today started as a conversation on Product Hunt. Netfox can close the processes holding a port open because someone asked for it there. FinderGit’s sidebar got collapsible sections the same way. That’s the part I’d miss most.

It trains the muscle that matters most: not quitting. Starting a side project is easy. Everyone has a folder of abandoned ones. The hard part is coming back to the same idea for the hundredth evening, when the fun part is long gone and what’s left is edge cases and release notes. Open source taught me to stay.

And that’s how you get the thing I’m proudest of. Built piece by piece, step by step, in “spare time”, but with the same solidity as a commercial product. Signed, notarised, auto-updating, documented. Spare time doesn’t have to mean side-project quality.

One thing I can say for sure: none of these projects would exist if I’d needed them to pay for themselves first.

what I’d tell someone starting today

  • Build where your users already are. My Mantine components get stars because Mantine developers live on GitHub. My Mac apps don’t, because Mac users don’t. If you want stars, build for developers. If you want users, go where users are, and accept that stars won’t follow.
  • A release is not a launch. 35 releases in a month got less attention than one article with a good title. Shipping is half the job. Telling people is the other half, and I’m still learning it.
  • Count it anyway. I’d never added up my stars by project before writing this. The numbers changed what I think I’m building and for whom.

If you maintain open source too, I’d love to hear your numbers. Especially if they look nothing like mine.

And if any of my projects saved you some time, sponsoring on GitHub is the most direct way to keep them going. No pressure. A star helps too, apparently more than I thought.

I’m Giovambattista Fazioli — I run Undolog, an open-source studio where I ship React components, Mantine extensions, macOS apps, WordPress plugins, and CLI tools. Been at it since 1983. You can find everything on GitHub.

Published via ZyVOP — Write once in Markdown, auto-backup to GitHub, and syndicate to Dev.to, Medium & Hashnode in 1 click.