Why I Built a Budgeting App That Tells You What to Do

I spent 5 years tracking every personal transaction I made, and for most of those 5 years I wasn't actually fixing anything.

The tracking habit came from the day job (I work as an accountant in a private company in Australia). I'd logged roughly 4,600 transactions since 2020. I call it my financial diary, a record of my life written in transactions instead of words. The data was all there. I was just ignoring what it was telling me.

I had bad habits I could see clearly. I knew I was spending on things I shouldn't. I could point to specific categories that were leaking money. But I had no system for deciding what to tackle first, so I just kept tracking and feeling vaguely guilty, month after month, for years.

That gap between "tracking" and "fixing" is the reason YourDigits exists.

The Streak

I started tracking in 2016, back in the Philippines, when I got my first accounting job. I used Bluecoins and I liked it at first because it worked like the double-entry systems I was learning at work. Balance sheets, P&L, the completeness of it all felt right to me.

But there was a lot of the app I never actually used. Graphs, charts, trend reports, forecasts. All the stuff that made it feel professional was the stuff I never opened, because even with my training, none of it helped me in my personal finance. I was logging for the sake of logging. I kept that up until 2019, when I quit my job to prepare for my master's studies in Australia. Then I lost my account because Bluecoins at the time didn't have cloud backup, and three years of data disappeared in one afternoon.

When I moved here in 2020, I started a fresh account. And this time I didn't stop. 4,600 transactions and counting. Same accountant habit, same financial diary, but now with better reasons to care about it because I was living in a new country and figuring out my money from scratch.

What Tracking Actually Showed Me (and What It Didn't)

One month I realised I'd spent $250 on Uber. Two hundred and fifty dollars. I wasn't taking Uber for fun. I was taking it because I kept missing my bus to work or clocking out too late to catch the one home. The money wasn't really the problem. The problem was time management. I would never have seen that without the tracking.

So the diary worked. It was doing its job as a record. What it wasn't doing was making me any better with money. I had other patterns just like the Uber one: things I could see clearly in the data and still wasn't fixing (I wrote about a few of them in What Your Spending Reveals About the Rest of Your Life). Not knowing where to start is how years go by.

The People Who Couldn't Even Get Started

I had a cousin who tried tracking twice. Once with Bluecoins, once with a spreadsheet. He lasted a couple of weeks on each. Not because he didn't care. Because the logging itself was too much work. Every transaction was a form he had to fill out. Category, amount, date, note, save. And then he had to do it again. And again. And by the third week he'd just stop opening the app. A friend tried the same thing and quit the same way.

I felt the same friction but pushed through it out of professional habit. For most people the input method is the wall they hit, and the entry barrier stops them before the data can build up. That's the part I wanted to fix before anything else, because I could see the pattern in everyone around me.

The Pause (and the Backlog That Broke Me)

In June 2025, I got busy building something else. My first ever coding project was a web puzzle app, and it took me about 5 months. I had zero coding background before that. I didn't know what frontend meant. I didn't know what backend meant. Claude Code held my hand through the whole thing, and even then the output was inconsistent because I didn't know how to write proper rules or documentation. The codebase was messy, the migration history was a mess, and I was learning security concepts I still don't know how to execute myself (RLS, qualifying objects, pinning search paths) just well enough to tell the AI to check for them.

While I was head-down on that project, I stopped logging my transactions. For about 6 months. No diary.

When I came back in November 2025 to catch up, it took me literal days. I was hunting transactions across every one of my bank accounts, one at a time, trying to piece together what I'd spent where. And in the middle of that catch-up I had this thought that wouldn't leave me alone. I've just spent 5 months building a web app from nothing. What if I built a personal finance app next? One that would solve every complaint I had, because I could implement any feature I wanted.

By December I was working on what became YourDigits.

Why Voice Came First

The project didn't actually start with the app. It started with the parser.

I was sitting with that backlog and imagining the opposite of what I'd just been through. What if you could just talk and have it become financial data? You wouldn't fill out a form, you wouldn't pick a category from a dropdown, you'd just say "spent twelve fifty at Subway for lunch" out loud and a transaction would appear. That felt like magic to me.

So I built the parser first, in its own repo. It's called speech2txn and it predates the YourDigits app by weeks. I started simple, one sentence at a time, then expanded to multiple transactions per sentence and fuzzy merchant matching so typos and mumbling wouldn't break it.

Only once the parser was actually working did I open Canva and start sketching mock screens for the app. Even those original mockups are nothing like the app I ended up with. I scrapped the whole design partway through and rebuilt it. The voice entry wasn't something I added to an existing app. It was the whole starting point, and the app grew around what the parser could already do.

The Sheets Budget I Already Had

The other thing I was quietly working on in the background, through all of this, was a budget flow I'd perfected in Google Sheets.

In January 2025, I finally sat down and worked out a pay cycle budget that actually made sense for how I got paid. I'm paid fortnightly. Monthly budgets never fit properly and trying to run two halves of a month always felt like I was forcing the wrong shape. I got the Sheets version working and I used it for the rest of the year, even through the 6-month tracking pause. The budget template kept working even when the diary wasn't.

So when I started building YourDigits, I already had the flow. Income at the top. Committed bills. Planned categories. A running total that had to hit zero before the cycle started. I wasn't inventing a budgeting method, I was porting one I'd already lived with.

That's the thing I needed most from an app. I was tired of rebuilding the budget manually in Sheets every single pay cycle.

Where All Three Pieces Finally Met

I had the voice entry. I had the Sheets budget flow. I had 5 years of tracking data. But I kept bumping into the same problem I'd been bumping into the whole time. Even with all of that, I still didn't know what to fix first.

And then I thought about my cousin and my friend for the hundredth time. They were stuck at the entry barrier. I was stuck at the action barrier. Same frustration at two different stages. They couldn't sustain the logging long enough to see the patterns. I could see the patterns and couldn't sustain the fixing long enough to make them go away.

Voice entry solved their problem. It made logging easy enough that the habit could survive past week three. But what was going to solve mine?

The answer was a system that ordered the problems for me. Something that looked at my finances, identified the specific leaks, put them in the priority order that actually works, and told me "this one first, this amount, this pay cycle." Not a dashboard. Not a chart. Tasks. Specific dollar amounts. Automated from whatever the real state of my money was.

That's what became the Leak Ladder. 9 leaks, in a priority order, fixed one rung at a time. Rungs above your current one get paused until the foundations are solid, so you don't end up investing while you still have high-interest debt (debt at or above 7% APR, the threshold where paying it off beats any realistic market return). The whole system is there so you don't have to decide what's next. The app decides for you and you just do it.

YourDigits is both pieces. Voice entry solves the input barrier. The Leak Ladder solves the "what do I fix first" barrier. Together they're the thing I needed and the thing my cousin needed, wrapped into one app.

Three Months, Start to Finish

I started in December 2025. I finished the core app by mid-March 2026. About 3 months, compared to the 5 months the web puzzle app took me. And YourDigits is a lot more sophisticated than the puzzle app was, which I'm still kinda in disbelief about.

Claude Code hadn't levelled up that much in three months. I'd just learned how to write instructions better. My first project had no rules, no documentation, no clear structure. This time I had actual SOPs for the work, proper rules for how the code should be organised, a clean migration changelog so I could trace every change. When Claude gave me something wrong I'd update the docs so it wouldn't happen again, and over 3 months the output got noticeably better because the context it was working from kept getting better.

Who This App Is Actually For

It's for the cousin who couldn't sustain tracking, the friend who quit the spreadsheet after three weeks, the person who's tried three apps already and feels like budgeting just isn't something they're capable of. And it's for me, the version of me who had a beautiful five-year diary of decisions that weren't being made. Tracking is necessary but it's not enough. You need the tracking to be easy and a system on top that turns the data into actions. My hope is that YourDigits makes a 5-year streak possible for someone who doesn't have accounting in their blood, and that the streak actually changes something this time.

Try It

YourDigits is live on the App Store. Voice-log is in there, the in-app audit is in there, the Leak Ladder with all 9 rungs is in there, the pay cycle task generator, the adaptive targets, the Plan Health card. Everything the 3 months of building added up to.

Download YourDigits on the App Store. If you want to see where you stand before you download anything, take the Know Your Digits quiz first. no signup, roughly 3 minutes. It'll show you which of the 9 leaks are active for you and what your Leak Ladder position is, which is exactly what the app uses to generate your first pay cycle tasks when you sign in.

If you want the philosophy side before the product side, The Leak Ladder: The Complete Guide explains the full 9-rung system and works as a standalone read.

---

Related reading:

• You Don't Have a Savings Problem. You Have a Leak.: the foundation article for the whole Leak Ladder philosophy
• Voice-Log Your Spending in 5 Seconds: the feature that came first and the reason the project exists
• The Leak Ladder: The Complete Guide: the full priority system the app is built around