What I learnt shipping AmethystWindows to 400 stars — and then archiving it
A few years ago I built a tiling window manager for Windows 10, inspired by Amethyst. I called it AmethystWindows. It got 400 stars, 30 releases, and a small but real community of people who used it daily. Then I archived it. This is what I took away from the whole thing.
What I built and why
I wanted Amethyst-style automatic tiling on Windows. Nothing good existed at the time, so I built it myself — mostly to scratch my own itch, partly to see if I could ship something real. The initial goal was modest: a working prototype that handled the basic layouts. It was my first real experience in the desktop apps world and it had a lot of architectural problems. Largely experimental but I shipped it anyway.
What shipping to a real audience taught me
- I learned is that stars don’t produce contributors. Most of your users are silent. They found your tool, it worked, they moved on with their lives — which is exactly what you want. Expecting community in return for a useful free tool is a category error.
- Issues are product signal, not a todo list. The complaints that kept recurring — edge cases around multi-monitor setups, virtual desktop quirks — were telling me something about what users actually needed, not just what was broken. Reading issues as a product person rather than a maintainer changes what you fix first.
- Building on platform internals you don’t control is a long-term liability you need to price in from the start. Win32 and DWM are a moving floor. They’re not hostile to third-party tools, but they’re not built for them either. Every Windows update was a potential regression I hadn’t caused and couldn’t predict. This wasn’t a reason not to build — it was a cost I should have been more explicit with myself about.
The decision to archive
Windows 11 introduced enough changes to the windowing and DWM internals that keeping AmethystWindows working would have meant rewriting most of the core. I looked at that work honestly — a significant rewrite of a side project I no longer used daily — and decided the answer was no.
That’s not failure. That’s a decision. Naming it clearly matters, because the alternative is an unmaintained repo that slowly rots and misleads new users.
A more recent reflection — on the rise of Ai…
Ai allows developers to explore different areas from the one they are daily involved. Are you a react developer who wants to move into native development? Ai empowers you. Access to documentation, boilerplate and code generation has become free. Sure knowing the coding world still helps. But for how long?
What really makes the difference, more than ever, is knowing your users and understand truly their needs. I really believe AmethystWindows’ primary key of success was this. I recognized a niche (original amwthyst power users constrained to work on windows) and satisfied their needs.
What carried forward
Almost everything I now believe about building software for power users came from AmethystWindows. Ship native. Be opinionated — tools that try to please everyone end up with eighteen configuration options and no personality. Make the default experience good enough that most users never need to touch a setting. Respect the platform you’re building on instead of fighting it.
Those instincts are in everything I build under rightright.me now. The surface changes — different platform, different problem — but the underlying model is the same: a small, sharp tool for a specific kind of user, maintained honestly.