Docs editor SaaS refinements: it's more work than I thought

Man, am I fucking glad I paused Mac app development. My entire flagship SaaS was full of holes and inconsistencies.

When I started using TWO Docs every day and went through the notes I'd gathered over the weekend, which I highlighted in my Monday post, it became clear I wasn't even close to ready to ship a Mac app, let alone market it aggressively.

Read also: I underused my own SaaS app, until now

This isn't really about code issues, though there are a few of those too. It's about how fucking inconsistent the whole app was in UI and UX, and the closer I looked at each component, the more obvious that became.

What I found

A few examples:

• There was no way to create subfolders or move a note into a different one.

• No way to move a subfolder back out to the top-level navigation.

• Notes was missing an entire top navigation, along with split view inside a note.

• You couldn't even create subcategories in notes. And that's just what I noticed first.

The list didn't shrink. It kept growing the more I looked.

I wanted to write this down mid-update so I could process what's going on, and maybe catch another hole or flaw while I'm at it.

The good news is my current approach is working. One thing at a time, vet it, repeat. I'm not trying to one-shot the entire fucking thing when I don't have a clear picture of what's happening, let alone hold every change in my head at once.

Plus, if I were to even try to one-shot it in a prompt, Claude will tell me it's all rainbows and sunshine, but we all know that's never gonna happen.

Read also: Claude said we fixed the SaaS MVP bugs, but I don't fucking know

The better news is I'm now far more confident in how the app operates and holds up for the public. At least the docs and notes part.

What I haven't touched

I haven't gone near the rest of it yet: planner, activity, library, studio. I'm sure that's going to hand me a new low, maybe a few disasters, once I start dissecting it.

But I figured if I get docs, folders, and notes solid first, I can at least say the editor works, and the rest can go to beta if I end up screwing something up. I don't think it will, though.

Claude did fucking Claude stuff

A few things went wrong, and I barely know how to explain it. We were fixing a bug, and I asked Claude multiple times whether it would touch the core editor or any basic functions. It kept guaranteeing me it wouldn't.

Turns out Claude screwed me over completely. I couldn't save documents, and on refresh, the text and title would just disappear.

That was so damn frustrating that at some point I was yelling so much that Claude ended the conversation on its own. I'll own my part in it too.

I copy-pasted the same three sentences a hundred times because Claude got dumb again, and I just spiraled into frustration.

That was my cue to step away and do something else, so I recorded some kids' stories for Story Valley and wrote this instead.

Disaster avoided. But I'd call this the frustration moment of the week.

Reaf also: Claude AI Is both my greatest tool and my biggest frustration

What's next?

I can't really complain. A few sessions, hitting my five-hour limits, about two days in, and I fixed most of the critical UX, UI, and functions for the core of the app. You can also see what's been deployed in Vercel in the image below.

Based on the current pace, I might finish the single workspace web app this week. If not, I'll need until next Tuesday.

Planner, activity, library, and studio are more work because they run on different logic: how they retrieve docs, how they're built, where everything lands.

Now that I think of it, I don't remember half of the shit that I previously deployed, and that's my mistake.

I don't think I overbuilt the app. TWO Docs is exactly what's sitting there right now. There won't be more menu items or extra complexity piled on. Because the brand name carries the SaaS.

Read also: I turned down $8,000 for my domain, and built TWO Docs instead

I think I just didn't think through the ongoing logic, UX, and UI carefully enough, and that was already flawed across my three main menu items.

Since this is different logic, the user journey might end up different too. I'm no wireframe guy, but I'll take a shot without them first. If that doesn't work, I might still design wireframes, though I don't even know where to start.

All in all, this stretch had its highs and lows. But it leaves me far more comfortable handling whatever comes up next. Not because I understand the code a hundred percent, but because I'm starting to know this app inside and out.

Let's see what Friday brings. I'll write another update then.