It started, as these things often do, with something almost embarrassingly small.
A loan calculator. A BMI calculator. A text encryption tool. Three single-page browser apps, no backend, no accounts, no nonsense. You'd open them, do the one thing they were built for, and close the tab. I built them because I wanted to build something, and these felt like a good place to not overthink it.
What I didn't expect was that they'd teach me the one principle everything I've made since has quietly obeyed: your data is yours, and it should never have to leave your device just so a tool can do its job.
I didn't set out with that as a mission statement. It just... fell out of the constraints. A calculator that runs entirely in your browser doesn't need a server. It doesn't need to phone home. It doesn't need to know your BMI, your loan amount, or anything else about you at all — and honestly, why would it? The absence of a backend wasn't a philosophical stance at first. It was just the easiest way to ship three tools on a weekend.
But once you've built something that respects the user by default — not because you tried hard to, but because the architecture made it the only option — it's hard to go back to the alternative.
From "one weekend" to "wait, I have eight of these now"
Somewhere between tool number three and tool number four, the itch changed shape. I stopped wanting to build small, useful things and started wanting to build things I'd actually want to use myself, every day. That's a different kind of project. It needs more thought, more design, more "does this actually solve a real problem or am I just procrastinating on my day job."
That's roughly when Craft was born — not as a brand name I sat down and invented, but as a label I gave, after the fact, to a cluster of tools that all shared a kind of DNA: developer-facing, a little opinionated, built by someone who is also the target user. Attic, MirrorHealth, Daily Pulse, AgentWatch — four tools, four very different problems, one shared instinct: build the thing you wish already existed, and don't ask it to phone home while you're at it.
And here's the part I'll admit with a small wince: I built all four for myself first. Selfish, in the most useful sense of the word. If a tool doesn't solve my actual, annoying, Tuesday-afternoon problem, I have no business asking anyone else to trust it with theirs.
Then something shifted again
Somewhere in the middle of building for developers, I noticed I kept sketching ideas that had nothing to do with developers at all. A journaling tool that didn't want to be clever, just quiet and honest. An audio tool for people who just want to make something without fighting a DAW's menu system for twenty minutes first.
That's Wonder — the second half of the family, aimed less at "people who live in a terminal" and more at "people who just want a tool that respects them and gets out of the way." Neiro and PulseJournal live here. Same local-first bones as everything else. Same refusal to make your data someone else's business. Just... a gentler front door.
If Craft is the toolbox in the garage, Wonder is the thing on the kitchen counter that everyone in the house actually uses.
Why any of this matters (a small soapbox moment)
I'll be honest — there's no shortage of tools out there that do roughly what mine do, minus the "your data stays on your device" part, plus a login wall, an onboarding email sequence, and a pricing page with a tier called "Enterprise." I'm not against businesses making money. I just think there's room — plenty of it — for tools that don't need to know anything about you to be useful. That's not a marketing angle. It's closer to a personal allergy.
So that's the shape of things, from three browser apps on a Saturday afternoon to a small, stubbornly local-first family of tools with a two-brand identity I definitely did not plan out on a whiteboard in advance (I promise, it emerged, the way most honest things do).
What's next
I'm planning to write more about the individual tools — starting with Attic, and specifically why I built it, which turned out to be a more interesting story than "I needed a tool like this," even though that's also true. There's a version of that story with a genuinely frustrating debugging session in it, and I think it deserves its own post rather than a footnote here.
If any of this resonates — if you've ever wanted a tool that just does the one thing it says on the tin, without asking for your email address first — you can find the whole lineup at chunk-jp.github.io. And if one of them saves you twenty minutes of your life, a coffee is always appreciated, though never expected.
Thanks for reading this far. I mean that in the least performative way possible.
Top comments (0)