Started building finance apps in 2014 (https://t.co/PqYWYClq8n). Creative itch, as simple as that.
10+ apps since. Killed half. Rewrote the rest more times than I'd like to admit.
Now building an AI factory to grow these and build what's next.
Posting how it goes.
@framara Still juggling but now all on desktop/mobile. No cli anymore.
Different niches: Claude for new and iterative, cursor for complex technical, grok bot as orchestrator for cloud cursors for “factory-like” process…
@starter_story The pain is sharper in a small niche. Serve it well and you get a chance.
Been serving small specific niches in personal finance for the last 10 years, like savings goals. I've been thanked (and paid) many times for actually caring about the problem.
The bar will also get higher. It's like with home appliances - everyone thought they would free people from housework. What they did instead was raise expectations: cleaner dishes, cleaner house etc.
Seeing the same with my finance apps: people expect flawless design, advanced reporting and modern UX even for something as niche as savings goals.
@BucharaYT Limited freemium has been working well so far. Got to 1M downloads across my finance apps this way. Thousands of customers, but mostly one-time purchases.
Experimenting with trials on some apps now. We'll see.
@jakecastilloooo Yeah... fighting churn in Bills, one of my apps, right now. Preliminary conclusion: it's too late to win back the ones who've already left (zero returns), but not too late to keep the ones who stay happy, so they don't.
@DivyanshT91162 Will look closely. Ended up with the same architecture in SFA Ops, an ops layer for my business: workflows in YAML, agent roles with access rights, a human approval gate. So much fun.
I gave up on sticking to one, tbh. What I found though is that different coding agents are best for different tasks:
- Claude Code desktop: new and complex projects that need quick iterations
- Cursor: complex bug fixing and feature tuning
- Cursor cloud: tasks driven by my Grok bot
- Muse: not sure what niche it'll take yet, we'll see...
@Evan_Swanson_ As someone with a portfolio of tiny apps, this is music to my ears 💯
Take debt payoff: the pain is huge and universal, the tools are still bad. Anything better than average has a real shot.
Google approved Dues: Cancel Subscriptions. It's live on Google Play now.
My feed has been full of horror stories about blocked developer accounts ever since I posted that I'd submitted. Read them all, and now I'm a bit nervous, even though it's probably my 100th submission.
Apple still thinking... They've always been more thorough, or maybe it's just volume? Link👇
Giving my factory a dedicated space - a server, a gateway, an interface.
The key upgrade: deterministic workflows - no more "my bad, that's on me" conversations with agents.
Now there are hard gates, defined workflows, tools and quality bars.
Migrating simple workflows first. Will keep you posted.
I like playing with Opus 5.5 on graphics so much these days. My latest: a new X banner with a "Dues coming soon" mark.
As an engineer and an entrepreneur, I have been waiting for these times to come, when I am finally able to make something beautiful on my own.
@MaxHirsch13 Wow. Never faced any of this in my first 10 years with mobile apps… new times!
Just submitted Dues: Cancel Subscriptions app - it enters into existing niche (which one is not??) with established apps already there.
We’ll see what apple has to say this time.
Just submitted Dues: Cancel Subscriptions to Apple review. Not my first app, but still somehow thrilling.
The idea is simple: fully private, no bank connection. Find, analyze and cancel your subscriptions.
First thing it found for me: £360 a year on LinkedIn. Why???
Yeah, real numbers sell. Testing my new subscription app, the first thing it found in my own data was £360 a year on LinkedIn. Why do I pay this??
A friend who tested it found two subs at £20 a month he'd forgotten about.
Both within the first minute. Not to plug it, just to make the point.