Expert iOS App Development Services | Swift, SwiftUI, Liquid Glass
iOS app development is the design, build and maintenance of native applications for iPhone, iPad, Apple Watch, Apple TV and Vision Pro, written in Swift and SwiftUI and distributed through App Store Connect. As of August 2026 it is no longer a one-off build followed by quiet years. Since 28 April 2026, every upload to App Store Connect must be compiled with Xcode 26 or later against the iOS 26 SDK, which turns the annual toolchain bump into a contractual requirement rather than a recommendation.
Mixcore Studio has been shipping production software for more than 8 years with a team of around 20 engineers and 320+ delivered projects. On iOS specifically, most of the value we add in 2026 is not writing new screens. It is getting existing apps through the current submission gates, adopting Apple's new material system deliberately instead of half-way, and making an honest call about which parts of an app should be native at all.
What changed in 2026, and what now blocks an update
Three dates matter more than anything else in the current Apple release cycle, and all three stop an app from shipping rather than merely nagging about it.
- 28 April 2026 — App Store Connect rejects any binary not built with Xcode 26+ against the iOS 26, iPadOS 26, tvOS 26, visionOS 26 or watchOS 26 SDK. Older Xcode versions cannot be used for submission at all.
- September 2026 — Apple's new social-media capability questions become a required part of the age-rating questionnaire, with a matching Social Media content descriptor on the product page.
- Late 2026 — iOS 27, iPadOS 27, macOS 27 (Golden Gate), watchOS 27, visionOS 27 and tvOS 27, announced at WWDC26 on 8 June 2026. iOS 27 requires an A13 Bionic or newer; Apple Intelligence features require A17 Pro or newer.
The current stable toolchain is Xcode 26.6 (build 17F113, released 25 June 2026), with Xcode 27 in beta since 8 June 2026 (beta 5 landed on 10 August 2026), adding coding agents that run on the model of your choice, a Device Hub for managing every test device in one place, and deeper workflow customisation. On the language side, Swift 6.3 shipped 24 March 2026 (current toolchain 6.3.3), bringing the @c attribute for exposing Swift to C, module-name selectors, @specialize and @inline(always), and the first official Swift SDK for Android. Swift is not a compromise on performance either — Apple reported that rewriting its TrueType hinting interpreter in Swift produced an implementation 13% faster than the C code it replaced (Swift.org, 12 June 2026).
The practical consequence for budgets is that an iOS app now carries a mandatory yearly cost even if no feature work is planned. We usually recommend a named maintenance retainer covering the SDK bump, deprecation sweeps and store-policy changes, rather than pricing each forced upgrade as a surprise project.
Liquid Glass is applied whether you planned for it or not
Apple's Liquid Glass redesign stopped being an opt-in style in 2026. Recompiling against the iOS 26 SDK automatically applies the new material to system controls: navigation bars, tab bars, sheets, alerts and standard buttons. Custom components written for the previous visual language do not convert. An app that rebuilds without a design pass therefore ships a visibly mixed-era interface, with system chrome in the new material sitting next to bespoke views in the old one.
That is the real scope question, and it is worth answering before quoting. An app built almost entirely from standard UIKit or SwiftUI controls may need only a review and small corrections. An app with a heavily custom design system needs its components re-specified against the new material, including how they behave over scrolling content and against varying backgrounds. Apple published updated Liquid Glass design kits for Figma and Sketch covering iOS, iPadOS and macOS 27 on 23 June 2026, which is the reference we work from with design teams.
- Audit first — inventory every custom control and classify it as converts automatically, needs restyling, or needs rebuilding.
- Contrast and legibility — translucent materials change how text and icons read over live content, which is where most accessibility regressions appear.
- Cross-platform limits — frameworks that draw their own widgets do not inherit the new system material, so a Flutter or React Native surface will approximate it rather than render it.
On-device and cloud AI through the Foundation Models framework
The Foundation Models framework makes Apple's own models a callable Swift API rather than a demo. It gives an app an on-device model for local work and larger server-based models running on Apple silicon in Private Cloud Compute for anything heavier, so choosing between them becomes an architecture decision rather than a research project.
In 2026 the framework gained multimodal prompts that let an app reason about images alongside text, Dynamic Profiles for swapping models, tools and instructions inside a single continuous session, and support for any model provider shipped as a Swift package conforming to the Language Model protocol. Google supplies exactly that for Gemini through the Firebase Apple SDK and Firebase AI Logic. In practice this is the most useful architectural change of the year, because one LanguageModelSession call site can fall back from local to cloud without rewriting the feature.
Two honest caveats belong in any AI scope, and we raise them before the estimate rather than after. First, the on-device model is tuned for summarisation, extraction, rewriting and classification. It is not a world-knowledge or hard-reasoning engine, so anything requiring current facts or long chains of logic needs a paid cloud model and a real inference budget. Second, device eligibility is a genuine constraint: Apple Intelligence features require A17 Pro or newer, so a share of your existing install base will never see the feature you are paying for. We size that share from your own analytics before committing to the design.
There is also a compliance consequence. App Review Guideline 5.1.2(i) now requires explicit user permission and clear disclosure before personal data is shared with third parties, expressly including third-party AI. Adding an LLM feature is therefore a consent-flow, privacy-policy and vendor-review exercise as much as an engineering one.
Age signals, privacy manifests and the current review rules
Age verification became an engineering requirement in 2026, not a policy checkbox. The Texas App Store Accountability Act took effect on 1 January 2026, and on 6 July 2026 the US Supreme Court allowed Texas to enforce its age-verification and parental-consent requirements (SCOTUSblog, 6 July 2026). App stores must verify the age of prospective users and pass age and parental-consent signals to developers, which means apps have to consume those signals rather than self-attest.
- Declared Age Range API — Apple's interface for reading an age band, including re-validation when parental consent changes and handling revocation, which is the case most implementations forget.
- Age-rating questionnaire — updated responses were due 31 January 2026, and the social-media capability questions become required in September 2026.
- Regional divergence — Australia uses its own 16+ and R 18+ bands rather than the standard international ones, and Korea allows GRAC Rating Classification Number overrides for Games and Entertainment apps from 12 August 2026. Ratings have to be checked storefront by storefront rather than set once globally.
- Privacy manifests — required, with valid signatures, for a fixed list of around 80 commonly used SDKs including the Firebase family, Lottie, OneSignal, RxSwift, SDWebImage, hermes, Flutter, Capacitor, Cordova and UnityFramework.
The awkward part for product teams is conversion, not code. A verification wall in front of onboarding costs sign-ups, so the design question is how late in the funnel the age signal can be requested while still meeting the obligation in the markets you actually serve.
Distribution and store fees now differ by country
Store exclusivity broke open on several fronts at once, and the correct answer is now regional rather than global.
- United States — apps may include buttons, external links and calls to action pointing to outside purchase mechanisms, with no entitlement required.
- European Union — external purchase links still require the StoreKit External Purchase Link Entitlement. Apple's 5% Core Technology Commission replaced the EUR 0.50 per-install Core Technology Fee as a single unified business model from 1 January 2026, applying across the App Store, Web Distribution and alternative marketplaces (Apple Developer, "DMA and apps in the EU"). Notarisation remains required on every distribution path.
- Japan — the Mobile Software Competition Act is now in force, enabling alternative marketplaces, third-party payment processing and non-WebKit browser engines. The permitted distribution paths are not the same set as the EU's, so the model has to be confirmed against Apple's current Japan terms before it is committed to.
- Brazil — Apple enabled alternative distribution and payment processing on 18 June 2026.
Our advice is usually less exciting than the headlines suggest. The question is net revenue after engineering, payment processing, tax handling, refunds and support, not the headline commission. For a large share of apps the external flow costs more to build and operate than it saves, and staying on StoreKit remains the rational choice. Where volumes justify it, we build it and instrument it so the comparison is measured rather than assumed.
One more regional gap worth costing early: Siri AI, the rebuilt assistant Apple announced at WWDC26, will not initially be available in the EU on iOS, iPadOS and watchOS. If your roadmap assumes assistant integration, the European build genuinely differs.
When native iOS is the wrong choice
Native Swift is not automatically correct, and we would rather say so at proposal stage than three months into a build. If your app is largely forms, lists and content served from an API, with no deep platform integration, two separately written native apps mostly buy you duplicated logic and two sets of bugs.
The pattern that has held up in 2026 is shared logic with native UI, and Kotlin Multiplatform is its usual expression. JetBrains' published case studies report that Bitkey by Block shares 95% of its mobile codebase with KMP, and that Forbes shares over 80% of its logic across iOS and Android and now rolls features out on both platforms simultaneously. Compose Multiplatform 1.11.0 added native UIView-backed iOS text input with system context menus, Autofill and Translate, and turned on concurrent iOS rendering by default.
Fully cross-platform toolkits remain viable and are moving in the same direction on iOS plumbing. Flutter 3.47 (12 August 2026) raised the minimum iOS version from 13 to 15 and made UIScene lifecycle adoption mandatory for iOS 27, and 92 of the top 100 iOS Flutter plugins have migrated to Swift Package Manager with CocoaPods moving to maintenance mode. React Native 0.87 (11 August 2026) added experimental Swift Package Manager support, and Capacitor 8.5.0 (31 July 2026) added iOS UIScene support with a migrator. What none of them do well is render Liquid Glass convincingly or reach the newest Apple APIs on day one.
Choose native iOS when the platform is the product: Liquid Glass fidelity, Foundation Models on device, widgets and Live Activities, watchOS and visionOS surfaces, background processing, or tight hardware and sensor access. Choose shared logic with native UI when feature parity across platforms matters more than platform-specific polish. We will tell you which case you are in, and the answer is sometimes that your existing app needs compliance work and a Liquid Glass pass rather than a rebuild at all.
Our expertise
- Swift 6.3 and SwiftUI
- Liquid Glass adoption
- Foundation Models on device
- App Store review readiness
- Privacy manifests and age signals
- iPadOS, watchOS and visionOS
Frequently asked questions
Can I still submit an app update built with an older version of Xcode?
No. Since 28 April 2026, every binary uploaded to App Store Connect must be built with Xcode 26 or later against the iOS 26, iPadOS 26, tvOS 26, visionOS 26 or watchOS 26 SDK. This is a hard gate, not a warning, so an app that has not been rebuilt since then cannot ship any update at all until the toolchain is upgraded.
Does adopting Liquid Glass mean redesigning the whole app?
It depends entirely on how much custom UI you have. Rebuilding against the iOS 26 SDK converts standard system controls automatically, so an app made mostly of stock UIKit or SwiftUI components may need only a review. Custom components do not convert and will look a generation behind next to the converted chrome, so a heavily bespoke design system needs its components re-specified against Apple's updated design kits.
Will Apple Intelligence features work on all of my users' devices?
No. Apple Intelligence features require an A17 Pro chip or newer, and iOS 27 itself requires at least an A13 Bionic. Any AI feature therefore needs a defined fallback for ineligible devices. We size the eligible share from your own install base analytics first, because on a long-tail user base it can be a minority.
Should we build natively in Swift or use Kotlin Multiplatform or Flutter?
Build natively when the platform is the product, such as Liquid Glass fidelity, on-device Foundation Models, widgets and Live Activities, watchOS or visionOS. Use shared logic with native UI when parity across platforms matters more. JetBrains' published case studies report that Bitkey by Block shares 95% of its mobile codebase with Kotlin Multiplatform and that Forbes shares over 80% of its logic across iOS and Android, which is the strongest current case for that middle path.
How much commission does Apple take now, and can we use our own payments?
It varies by storefront. In the United States apps may link out to external purchases with no entitlement required. In the EU an entitlement is still required and Apple's 5% Core Technology Commission replaced the per-install Core Technology Fee on 1 January 2026. Japan and Brazil each have their own variants. The decision should be made on net revenue after engineering, payment processing, tax and support, which for many apps still favours StoreKit.
Contacts
We are always happy to talk with you.
Feel free to contact us in any suitable way
Request a quote
Let's discuss your project!
Please, provide us with a brief description of what you
already have and what you are going to achieve.
Mail us contact@brainiacminds.com