Jarvis, six months in

Personal · 2026-08-25

Personal · 21 April 2026 · 4 min read

Jarvis, six months in

Right, so I've been putting off writing this one, mostly because I knew I'd have to be honest about it. And honest is annoying when the thing you're writing about is also the thing you've been building in your spare time, evenings, weekends, the odd stolen hour on the 26 bus.

But that's the whole point of this blog, isn't it. Unfiltered. So here we are.

Jarvis started as a solution to a problem I actually had, which sounds obvious but you'd be surprised how many products don't start that way. I needed something that could sit across my scattered notes, half finished tasks, todo items I'd write at 11pm and forget about by morning, and actually make sense of them. Not organise them into pretty categories. Just tell me what mattered today.

Six months on, here's where it actually stands.

What's working

The core loop works. You give Jarvis a messy brain dump, it pulls out the actionable bits, and it doesn't try to be clever about it. No fake personality, no "great question!" nonsense. Just: here's what you said, here's what looks like a task, here's what doesn't.

That sounds small. It isn't. Most of the tools I tried before building this wanted to turn everything into a project management exercise, tags and boards and colour coded priorities I'd set up once and never touch again. Jarvis doesn't ask for that. It just reads what you give it and gets on with it.

The other thing that's genuinely working, and I didn't expect this, is the todo item numbering. Item 8, item 9, whatever, referencing them directly instead of vague descriptions has changed how I actually plan my week. I know exactly what I'm looking at. No ambiguity. It's a small thing but it's the difference between a tool you use once and a tool you build a habit around.

What isn't there yet

Right, the honest bit.

Roadmap items one and two, the ones I keep telling myself are "coming soon", aren't done. One's a proper calendar sync, the other's a way to handle recurring tasks without me manually re-adding them every week like an idiot. Both are half built. Both keep getting bumped because something more urgent shows up, which is, I think, just what building something solo actually looks like. You don't get the luxury of a team picking up the slack while you focus on the fun bit.

Theoretically Jarvis should already do a bunch of things it doesn't yet do well. It should handle context switching between projects better. It should learn your patterns over time rather than treating every input like it's brand new. It doesn't, not really, not yet. I could sit here and tell you it's "in progress" and leave it vague, but that's the kind of vague I've come to distrust in other products, so I won't do it here.

The problem it actually solves

Here's the bit I'm most confident about. Jarvis solves the problem of task paralysis, that moment where you've got forty things floating around your head and no idea which one to grab first. It doesn't solve productivity in some grand sense. It's not going to make you a different person. It just clears the fog long enough for you to start.

What it doesn't solve, and I want to be upfront about this, is discipline. If you don't actually do the tasks it surfaces, nothing changes. It's not magic. It's a mirror, mostly, one that's slightly better organised than the mess you handed it.

Where this goes

I built Jarvis because I needed it, not because I thought there was a market sitting there waiting. Whether there's product-market fit, real product-market fit, is still an open question. What I do know is it's solved my own problem enough that I keep opening it every day without thinking twice, and that's a decent signal even if it's not a spreadsheet full of user metrics.

Anyway, that's the honest state of things. Roadmap items one and two are next. I'll write about them properly once they exist rather than promising them into being.

For now, Jarvis does the small thing well. Sometimes that's enough to build from.