How long does it take to build an app on your own?

Every honest answer to how long does it take to build an app starts with a question back, which is why the answers you find are so useless. Six weeks and eighteen months are both true, for different apps, by different people, meaning different things by the word finished. So instead of a range, here is where the time actually went on the two apps I have built as one person, and which parts surprised me.

I run a one-person studio. No team, no contractors, and for a good stretch of it a full-time job alongside. That shapes everything below, and it is probably closer to your situation than a case study from a fifteen person agency.

The finish line keeps moving, which is most of the problem

When people estimate an app they are almost always estimating a working prototype. That version, the one that does the core thing on your own phone, comes together fast now. For my journaling app I had recording, transcription and a list of entries running in something like three weeks of evenings. It felt close to done. It was maybe fifteen percent of the work.

The distance between a prototype and something a stranger can install without you standing next to them is where the months live. Everything in that gap is unglamorous and none of it is optional.

Where the time actually goes

The edge cases outnumber the feature

The happy path took three weeks. Then: what happens when the recording is interrupted by a phone call, when storage is full, when the app is killed mid-save, when someone rotates the phone at the wrong moment, when permissions are denied and then granted later, when the device language is not English. Each one is small. There are dozens, and they are the difference between a demo and something you can leave with people.

Other people's devices

Android is the honest teacher here. Your app works on your phone. Then it turns out several manufacturers aggressively kill background processes to save battery, so a scheduled task that fires reliably on one handset silently never runs on another. I lost close to two weeks on exactly that for the alarm app, and the fix was not clever code, it was finding out that the behaviour existed at all.

Getting into the store

This is the one that catches first-time solo builders hardest. If you are publishing on Google Play with a personal developer account, you need a closed test with a minimum number of real testers running for fourteen continuous days before you can even apply for production access. That is not a queue you can pay to skip, and finding a dozen people who will actually install the thing and keep it installed takes longer than the fourteen days suggests.

Around that sits a pile of small mandatory work. A privacy policy hosted at a stable URL, a data safety declaration that has to match what the app really does, an icon at several sizes, a feature graphic, screenshots, a store description, a content rating questionnaire. None of it is hard. All of it is a day or two you did not plan for.

Anything involving money

Adding a paid tier is not a feature, it is a subsystem. Products configured in the store console, purchase flows, restoring purchases on a new device, what a lapsed subscription does to existing data, and testing all of it in a sandbox that behaves differently from production. Budget several weeks and expect to be annoyed.

The prototype took three weeks and it was about fifteen percent of the work.

My actual numbers

Luna Journal, my voice journaling app, ran to roughly seven months from first commit to something in real people's hands, working evenings and weekends around a job. Call it fifteen to twenty hours a week on a good week and nothing at all on a bad one. Full-time and knowing what I know now, I think that is a three month build.

Meridian Clock, the timezone alarm app, was faster to a first release because I had the toolchain, the store process and my own patterns sorted. The app itself is arguably harder, since alarms have to fire exactly, survive a reboot and behave when you cross a border. But the second app benefits enormously from the first, and that compounding is real. The store setup that took me a fortnight the first time took an afternoon.

I was also learning Flutter through the first one, which I would guess added a couple of months. If you already know your framework, take a chunk off.

A way to estimate that is less wrong

Estimating the whole app is hopeless. What has worked better for me is estimating only the core interaction, the single loop the app exists for, then applying a multiplier.

• Estimate the core loop honestly. Record, transcribe, save. Set an alarm, ring, dismiss.
• Multiply by four for the edge cases, real devices and the general gap between works and works for strangers.
• Add a fixed four to six weeks for store setup, assets, policy and the testing window, whatever the app is.
• Add a third on top if it takes payments.
• If you are learning the framework as you go, add half again to the whole thing.

That is not a formula so much as a way of not lying to yourself in week one. My estimates are still wrong. They are wrong by a third now rather than by a factor of five, which is enough to plan around.

What actually makes it faster

Shipping something small. Not as a philosophy, as a practical matter. The first version of Luna did far less than what is in it now, and every week it sat unreleased was a week of building on guesses. Ship the core loop, put it in front of people, then decide.

The other one is finishing an app before starting the next. I have three or four ideas at any point and the pull towards the fresh one is strong, especially in month five when the current thing is all edge cases and store paperwork. Every founder I know who is still shipping does one at a time. Every one who is not has four repos at eighty percent.

So the useful answer is somewhere between three and nine months for a focused first app, longer around a job, shorter for the second one. The coding is the small part and it is the only part anyone estimates.

If you want one number to plan with, take your honest estimate and double it. I have not once regretted doing that.

Benjamin

Founder & sole developer, Novaire Digital