Voice-Log Your Spending in 5 Seconds

I've been tracking every personal transaction since 2020. Roughly 4,600 of them. I'm an accountant so the habit came with the job, and I still almost quit during one stretch in late 2025. Not because I stopped caring about it. I spent 6 months building a different app, neglected my financial diary, and when I finally sat down to catch up it took me literal days, hunting transactions across bank accounts and filling out forms one after the other for every single entry.

That's when it hit me how ridiculous the entry step is.

The 30-Second Tax

Open any budgeting app and watch what happens when you try to log one expense. Tap the "+" to add a transaction. Pick a category from a dropdown. Pick an account. Type the amount. Pick the date (it usually defaults to today but you still check). Write a note so you'll remember what it was. Save.

Six or seven taps. Roughly 30 seconds per transaction. Which feels small until you actually do the math. Five transactions a day for 30 days is 150 a month, which comes out to more than an hour of your month just pushing buttons inside an app. And that's assuming you're on top of it. If you skip a few days and have to backfill, it's worse.

I keep saying this but the reason people quit budgeting apps isn't discipline. It's that logging one transaction takes 30 seconds of form-filling every single time, and of course you stop doing that after a couple of weeks.

I call it the 3-week wall. You're fine for the first couple of weeks. Still motivated, still novel, the app is catching everything. Then life gets busy, you miss a day, then two, then a week. By the time you reopen the app there are 40 transactions to backfill and you just... close it again. That's not really a discipline problem, that's the entry flow doing it.

The Fix Has to Happen at the Input Layer

Here's the part most budgeting apps get backwards. They spend all their engineering effort on the dashboard: charts, graphs, spending trends, category breakdowns. All of that is nice if you have data to look at, but if the entry friction is what's killing the habit, none of it matters because you won't have any data by month 2. The fix has to happen at the input layer. That's the actual bottleneck.

Obviously I'm coming at this from the perspective of someone who built around that idea. The voice parser came before YourDigits existed as an app. I started it as a "what if" experiment: what if you could just say what you spent and it turned into structured financial data, the way talking to a person and having them write it down would? That felt like magic to me. So I built the parser first in its own repo, tested it with sentences that got more and more complicated, and only once it actually worked did I start designing an app around it. Voice entry wasn't a feature we bolted on later, it's the whole reason the project exists.

What 5 Seconds Actually Looks Like

Here's the flow as it works today.

You open the app, hold the voice button, and say something like: "Spent twelve fifty at Subway for lunch."

The parser does a few things at once.

It pulls the amount. $12.50.

It identifies the merchant. Subway. Even if you say "the Subway near work" or mumble the name a little, fuzzy matching handles typos and casual phrasing.

It picks a category. Food & Dining, because Subway is a food merchant and "lunch" confirms it.

It assigns an account. Your default spending account, unless you say otherwise in the sentence.

It gives you a transaction card to review. You can edit anything before saving, but most of the time you don't need to.

That whole thing takes roughly 5 seconds instead of 30.

If you have a few things to log at once, you can speak them all in one sentence. "Spent eight dollars on coffee, forty-five at the grocery store, and twenty on petrol." The parser splits that into three separate transactions and shows you three cards. Review each, adjust anything that's off, save.

If you travelled recently, it handles multi-currency. "Spent 500 pesos on dinner in Manila" will auto-convert to your home currency. I built that because I travel home to the Philippines and I got tired of either manually converting every holiday transaction or just not logging them and leaving a hole in my diary.

Your Audio Stays on Your Phone

This one matters to me personally. The speech recognition runs on-device using Whisper, which means the audio never gets uploaded anywhere. It's processed locally on your phone, turned into text, and the text goes into the parser. Nothing leaves the device.

Most voice features in other apps work the opposite way. You talk, your audio gets sent to a server, a cloud service transcribes it, the text comes back. For a lot of use cases that's fine. For a budgeting app where you're saying your merchant names, your amounts, the places you've been, it felt like that data shouldn't have to take a trip to someone else's server just so a form could be filled in.

On-device speech is also why YourDigits is iOS only for now. The Whisper implementation is tuned for Apple hardware. Android would be a different build.

Why This Actually Changes the Habit

The reason I care so much about the 5-second thing is because I've been on both sides of it. I kept tracking through the friction for 5 years because my accountant brain wouldn't let me stop, and I still hit a 6-month pause in 2025 that showed me how fragile the habit actually is even when you're doing it as a professional habit. If the friction had been lower I probably wouldn't have paused, and the catch-up definitely wouldn't have taken days.

My cousin is the other side of it. He tried tracking twice, using two different tools, and quit within a few weeks each time. And it wasn't because he didn't care about his money. The logging itself just felt like homework to him, and he never got far enough to see what was actually wrong with his finances because the entry barrier stopped him before he could get there.

Voice entry solves the thing that's actually causing people to quit. You speak, it logs, you move on, and the habit survives because it stops being such an effort.

Once the Data Is There, Then What?

Voice entry solves the input problem. It doesn't solve the "what do I do about what I'm seeing" problem, which is a separate one and worth naming.

That's what the rest of YourDigits is for. Once the data is in, the app runs it through a 9-rung system called the Leak Ladder, a priority-ordered system for fixing financial leaks in the order that actually works. You answer 11 questions in the Know Your Digits audit, get a score from 0 to 100, and the app tells you which leaks you have and what to fix first, pay cycle by pay cycle. Voice entry keeps the daily habit going. Without the Leak Ladder layered on top, you'd end up with nice clean data and no idea what to do with it, which was the position I was in for most of my 5 years of tracking.

If you want to see where you stand before you download anything, take the Know Your Digits quiz. 11 questions, no signup, no email gate, roughly 3 minutes.

Try It

YourDigits is live on the App Store as of today. Voice-log is in there from the first launch, along with the audit, the Leak Ladder screen, the pay cycle task system, and everything else the 3 months of building added up to.

If you've been putting off tracking because the last few apps you tried felt like paperwork, this is the part that was meant to fix that. Download YourDigits on the App Store and try logging 5 transactions in under 30 seconds. That's roughly the benchmark I built it around.

If you're not ready to download yet, the Leak Ladder guide explains the philosophy and works as a standalone system. And if you want to see how iPhone voice budgeting compares to what's already out there, I wrote a companion piece on how YourDigits compares to Android voice apps.

Next in this series: a walkthrough of the full system, how the audit, leak detection, pay cycle tasks, and adaptive targets all fit together.

---

Related reading:

• Why Budgeting Apps Fail (And What Should Replace Them): the companion Phase 1 piece on the 3-week wall
• The Leak Ladder: The Complete Guide: the full priority system the app is built around