Hi there! My name is Majid. I'm Swift developer 👨🏻‍💻SwiftUI addicted 🚀 Creator of CardioBot, NapBot, FastBot and SugarBot.
How to stream partial outputs from Apple’s Foundation Models in Swift. Use LanguageModelSession.streamResponse to get a ResponseStream (AsyncSequence) of a Generable type, update UI as fields arrive, then collect() for the final result. Note: property order drives emission.
Implement feature flags per build config (Debug, TestFlight, App Store) using compiler conditions (#if APPSTORE/TESTFLIGHT) and a Distribution enum. A FeatureFlags struct drives toggles and is provided via SwiftUI Environment. Remove flags when done; consider remote config for rollouts.
Tutorial on generating structured content using iOS Foundation Models framework. Shows how to use @Generable and @Guide macros to create strongly-typed Swift models that AI can populate. Demonstrates building a recipe generator that takes ingredients and returns structured Recipe objects with nutritional data. Key APIs: LanguageModelSession.respond(generating:), @Generable macro for conforming types, @Guide macro for field descriptions. Makes AI output predictable and developer-friendly by transforming raw text into typed Swift structs.
Introduces Apple's Foundation Models framework for integrating Apple Intelligence into iOS apps. Covers SystemLanguageModel availability checking, LanguageModelSession creation with custom instructions, and response tuning via GenerationOptions. Shows building a healthy lifestyle coach feature with temperature control (0-1 for creativity) and sampling modes (greedy vs random with seeds). Key APIs include isAvailable property, respond() function, and structured prompt engineering for consistent AI-generated recommendations.
Introduces Observations, a new Observation framework type that turns Observable state into AsyncSequence. It fixes withObservationTracking’s single-fire/manual recursion and works with for‑await. Observes touched properties, batches transactional updates, and enables key‑path streams—bridging Combine‑style updates to Swift Concurrency.
Using glassEffect on each view isn’t enough—glass should interact. Wrap related views in GlassEffectContainer so effects reflect each other and render more efficiently. Control morphing with the spacing parameter, or merge distant shapes via glassEffectUnion (same effect/shape/id).
Guide to SwiftUI’s glassEffect for custom views in Liquid Glass: apply the modifier, pick Glass styles (.regular, .clear, .identity), toggle with .identity, tint colors, use .interactive for stronger background response and gesture handling, customize shape. Effect adapts to content behind.
Introduces SwiftUI’s @Animatable macro to animate view/shape stored properties without manual Animatable conformance. Use VectorArithmetic types (e.g., Int → Float), drive with .animation(value:), exclude fields via @AnimatableIgnored. Existing conformance is unchanged.
Explains adopting Liquid Glass toolbars in SwiftUI. Toolbars gain glassy look; use symbol+text buttons. For pre‑Liquid Glass, add a custom LabelStyle with availability checks. Use ToolbarItemPlacement, optional tint/badge, and group with ToolbarItemGroup/ToolbarSpacer.
Adopt WWDC25’s Liquid Glass tabs via SwiftUI’s new Tab API: use TabView + Tab items, bind selection with @SceneStorage, mark .search role, auto-switch to sidebar with .tabViewStyle(.sidebarAdaptable), add .tabViewBottomAccessory, and minimize using .tabBarMinimizeBehavior.
SwiftUI after WWDC25: Liquid Glass design applies when building with Xcode 26. New TabView APIs (e.g., Tab with role .search), glassy toolbars with ToolbarSpacer, GlassButtonStyle and glassEffect. TextEditor gains attributed text, new WebView, macOS List performance boosts, and an Instruments profiling template.
Explores scaling microapps architecture in Swift from single-package to package-per-feature approaches. For small apps, single packages with ~20 modules work well, but large apps with 100+ modules face issues with Package.swift complexity and Xcode performance. Solution: package-per-feature where each feature lives in separate packages with internal modules for layers (domain, UI). Enables selective module dependencies and better separation of concerns while maintaining modularity and continuous delivery at scale.