Mobile appsMethodTimelinesCase study
Building an iOS app: from idea to the App Store
9 min read
Building an iOS app means five pieces of work, in order: pin down what the app is for, build it (in Swift or Flutter), test it on real iPhones with TestFlight, prepare the App Store listing, and pass Apple's review. The review is the shortest part. What takes time is learning the rules before you write the first line.
We took this path for our own app, Tubissime, on sale on the App Store since 7 November 2026. The dates and pitfalls below come from that release.
The whole journey at a glance
| Step | What happens | For Tubissime |
|---|---|---|
| Developer account | Enrol in Apple's programme, as the company | LPR Studio's own account |
| Build | Screens, server calls, in-app purchases | First commit on 28 October 2026 |
| Testing | Test builds on real phones (TestFlight) | Full order flow tested before submission |
| Store listing | Name, subtitle, description, keywords, screenshots | 18 localised listings |
| App Review | The app and its in-app purchases are reviewed | Submitted 5 November, approved overnight 6 to 7 November |
| Release | Goes on sale, automatically or manually | On sale the evening of 7 November 2026 |
Ten days from first commit to launch is fast, but the app sat on a server that was already in production: lyrics, orders and delivery already existed for the website. An app that starts from nothing, server included, usually takes a few weeks to a few months.
Before any code: the developer account
Publishing on the App Store requires the Apple Developer Program, at 99 USD a year. Enrol as an organisation, not as an individual: the company name is what appears as the seller under the app, and the account does not leave with an employee or a contractor. Apple then asks for a D-U-N-S number, the Dun & Bradstreet company identifier; if you need a new one, allow a few days.
On Android, the Google Play Console charges a one-off 25 USD fee. Set both accounts up while the app is being built, never the day before launch.
Swift or Flutter: the choice that shapes everything after
There are two serious ways to build an iPhone app.
- Swift, Apple's own language, gives the most direct access to the system: widgets, the Watch, features that ship with each new iOS. It only produces the iPhone app, though; Android has to be written separately.
- Flutter, Google's cross-platform toolkit, produces the iPhone and the Android app from the same code. It is the right fit when the app is made of screens, forms and calls to a server.
For Tubissime, Flutter was the obvious choice: one codebase for iPhone and Android, and the heavy lifting (writing lyrics, composing music, delivery) stayed on the server. The app is the interface. Before any of this, our comparison of web apps, PWAs and native apps helps check whether an installed app is really needed.
What Apple checks, and what to plan for from day one
The App Store Review Guidelines are worth reading before you design a single screen, not after a rejection. Four of them change the design:
- Payments. Digital content unlocked in the app is sold through Apple in-app purchase (guideline 3.1.1), commission included: 15% under the Small Business Program, 30% otherwise. A promo code that unlocks content outside that payment is not allowed. In Tubissime, the coupon field used for testing only exists in internal builds, never in the App Store version.
- Accounts. Apple does not accept mandatory sign-up without a good reason, and an app that lets people create an account must let them delete it from inside the app. Tubissime launched with no account at all: an email address is enough to order, just as on the website.
- Privacy. The listing declares what data is collected and why, and links to a privacy policy. Every analytics or payment SDK you add becomes part of that declaration.
- Reviewer access. If part of the app sits behind a login, Apple needs a demo account. Without one, the review stops there.
In-app purchases, and the first-purchase trap
An in-app purchase is created in App Store Connect (product ID, price, name and description in each language, a screenshot for the reviewer) and then sold from the app. The most important part happens on the server: the purchase receipt sent by the phone must be verified there before anything is delivered. In Tubissime, a song only starts once that check has passed.
The trap is timing. Apple spells it out in its documentation: the first in-app purchase of each type (consumable, subscription and so on) must be submitted with a new app version; once that first one is approved, further ones can be submitted without a new version. So plan all your purchases into the first version's submission, and check that they are actually attached before you send it.
In Tubissime's case, with three packages for sale, App Store Connect would only attach one purchase to the first version: the other two were refused until the first had been approved. So we chose a manual release, rather than put an app on sale with two of its three packages not yet purchasable. The two purchases were submitted on the morning of 7 November, approved the same day, and the app was released that evening.
TestFlight: test before you submit
TestFlight, Apple's testing service, installs work-in-progress builds on real iPhones: up to 100 internal testers, and up to 10,000 external testers after a short review of the build. In-app purchases can be tested there without real charges. This is where you run the whole journey, purchase included, on an ordinary phone rather than the developer's simulator.
The App Store listing: written to be found
App Store search relies on three short fields: the name (30 characters), the subtitle (30 characters) and the keywords (100 characters, hidden from users). On top of those come the description, screenshots and the promotional text, which can change without a new version.
Each language has its own listing, and each country reads its own. Tubissime has 18: one per app language, plus variants for the same language read in several countries. Keywords are researched, not translated: the word a German types when looking for a birthday present is not the translation of the French one.
Review and release
According to Apple, half of all submissions are reviewed in under 24 hours and 90% in under 48 hours. Tubissime went into review on 5 November; we replaced the submitted build with a corrected one on the 6th, which restarts the review, and approval came overnight from the 6th to the 7th.
Once approved, the app either goes live automatically or waits for your go. A manual release lets you pick the moment: the end of an approval, an announcement, or simply a quieter day.
After launch
The first version is only a start. Every update goes through review like the first one, and the listing improves as real search data comes in. Tubissime's next version, submitted on the evening of the launch itself, mostly carried new keywords in all 18 listings. The Android version, built from the same code, has since followed on Google Play; the Android-specific steps are in How to publish an Android app on Google Play.
For the full budget, store fees and maintenance included, see our guide to mobile app development cost, and to brief a provider, our mobile app requirements document template.
Where to start
Start from how the app will be used, not from the technology: who opens it, to do what, and what it sells. At LPR Studio, the visual prototype is free and presented on a video call within 5 working days: you see your app's screens on a phone before you commit to anything, and we then take the release all the way to the App Store.