One app, all my focus in August: update #5

I didn't expect to write another update so soon, but here we are. It's Monday, July 27, and I'm freakin' hyped about this one. But it needs some context first, because the story is a bit backwards.

Ignoring my own flagship product

Somewhere along the way, I got so absorbed in building that I forgot about the thing that ignited the whole vibe-coding spark in the first place: TWO Docs, my writing and docs app.

Not so long ago, I was deep in design week, auditing and overhauling most of my projects with better layouts, and it was during that overhaul that I realized I'd been quietly ignoring my own flagship product.

That's on me. Not because I stopped liking what I built, but because I hit a massive concrete wall with the app I couldn't solve at the time. I was frustrated, and honestly, just limited by my own capacity.

Back then I was still working with just Claude AI and the GitHub web editor, manually updating every file one at a time. That process was daunting. Fixing anything big at scale was hard.

Now that I've moved to Claude Code with Cursor, plus the Claude chat agent for writing more detailed prompts, execution is roughly ten times faster, and I'm getting zero deployment errors on Vercel.

That shift gave me the confidence this week to go after the most critical part of my first-born SaaS app.

Read also: I build software like a toddler learning to walk

How it got fixed

Three things were on the table. Eliminating the Liveblocks integration, since fewer dependencies are always better.

Fixing the syncing across shared workspaces. And fixing live editing, so the app actually feels instant when someone else is typing.

A huge part of the app hinges on that feature, right alongside split view, the two-docs-side-by-side layout. I wasn't in the mood to promote the app or go all-in on it while that piece was still broken.

But after a few genuinely great sessions this week, followed by some deeply frustrated ones, a few dozen iterations, and a fair amount of yelling at Claude AI, I got there.

I fixed the most critical part of the app, and it's holding up after real testing, trial and error, and some serious hairpulling, even though I still don't understand ninety percent of what the fuck is happening under the hood.

Why single focus, and why now?

So here's the decision: single focus on TWO Docs. Eighty percent of my time goes to polishing this app before I move on to building the Mac and iPad versions. (which I am terrified to even start)

I built this app specifically for people like me. Daily iPad users who like to write there, and it should be the flagship piece of everything I market going forward.

There's a principle underneath this decision that I want to say out loud, because I think it applies beyond just this app: when you're close to the finish line on something, it's better to give it one long stretch of real focus than to keep splitting your attention across five projects at once.

I've been building more than I can chew for months. The one app I'm most proud of is exactly the one that needs the most attention right now, and spreading myself thin any longer would only keep it half-finished.

Read also: I did more in 30 days with Claude than I did in a year

August is going to be about polishing the MVP, redesign and details included. Auditing the entire app again, doing the sanity checks I keep putting off.

Mapping and wireframing it properly so the UX and UI match everywhere, instead of leaving pieces untouched. A massive overhaul of the website, which is honestly too embarrassing to show right now, though I'm sharing it below anyway.

Building a real knowledge base and FAQ. Expanding the templates section (because I have been neglecting that). And redoing the pricing strategy, including optimizing the free tier.

Where this leaves me

Am I setting myself up for disappointment if nobody buys or uses the app? Probably, a little. But that's okay.

I also think I won't see real sales for a long time, not until there's a dedicated iPad app and the docs app can shine in its own place. I know the competition is there, and I've made peace with that. But who knows. Maybe I end up like Coda, acquired by Superhuman one day.

This focus makes sense given my limits and my skills right now. Building a polished version helps me refine my approach before I even start on the dedicated iPad and Mac apps. (I am fucking terrified doing this by the way)

I thought about documenting every single step here on the blog, but that would be mentally draining to keep up with. I do plan to update regularly though, especially the failures, the moments I got properly frustrated, and the reasoning behind the approach and the pricing as it comes together.

For now, I'm back at the drawing board, mapping out the first steps and making sure it's carefully planned so I don't hit my weekly Claude limit too early.