
Issue 752
29th May 2026
Written by Zachary Powell
Comment
I feel like I should kick things off this week by introducing myself.
Hi!
I’m Zachary Powell, Head of Developer Community here at next.app devCon. I’ve spent much of my professional career as a mobile developer. While I’ve hung out in Android spaces more than iOS spaces, I’ve always been interested in mobile development as a whole, and I’ve suffered through my fair share of Objective-C based projects.
This week’s iOS Dev Weekly marks the first one written by me, so please be kind (although, of course, feedback is welcome!). If you missed last week’s announcement, make sure to check that out.
But enough of all that, on with the newsletter!
We are officially in the pre-WWDC lull, but the quiet will not last for long. With WWDC26 kicking off on June 8, the developer community is buzzing with anticipation for iOS 27 and the next generation of Apple software. Apple took the rare step of explicitly signposting AI advancements and new developer tools in their event invitations, so we can expect a highly technical and packed keynote.
As someone who has always been passionate about developer communities, it is amazing to see Apple recognizing community builders from around the world who truly help educate, shape, and grow the community we are all part of. Seeing this highlighted front and center on the developer site is awesome. Good job Apple, keep it up!
– Zachary Powell
Sponsored Link
Call for Speakers at SwiftCon is open!SwiftCon is coming to Berlin as part of next.app devCon, the largest gathering of mobile app developers. We’re opening the CFP for iOS builders who go beyond tutorials: SwiftUI in production, architecture decisions, performance wins, and hard trade-offs. If it shipped, it belongs here. Submit your talk today.
News
Introducing the 2026 Apple Design Award finalistsIn case you missed it, Apple has announced the 2026 finalists for the Apple Design Awards. It is great to see such a wide range of apps and games highlighted here. Of course, I am a particular fan of PowerWash Simulator, so seeing it under “Delight and Fun” definitely brought a smile to my face.
Tools
AgentKitten: Provider-Agnostic AI Agents in SwiftWith the broader tech ecosystem flooded with AI agent frameworks like LangChain and the Vercel AI SDK, the Swift and Apple platform space has felt surprisingly barren. Ferran has stepped up to fill that gap with AgentKitten, a lightweight, open-source Swift package designed to build provider-agnostic AI agents. Rather than replacing native application features wholesale, the framework focuses on practical on-device augmentations like semantic search, structured automation flows, and intelligent app actions.
Code
Taming Row Height and Spacing Jumps in SwiftUI ListsDynamic content changes inside a SwiftUI List row often lead to jarring layout snaps or awkward lingering spacing. In this excellent deep dive, Fatbobman walks through why the native layout engine struggles with frame-by-frame height interpolation for dynamic list rows. He then details a pure SwiftUI solution using state decoupling, a custom layout container that clips content rather than squeezing it, and an animatable stack that handles row spacing on the exact same timeline.
Building a Custom Data Store in SwiftDataWhile SwiftData excels at abstracting away the complexities of Core Data for local storage, there are times when you need your data layer to speak to a custom backend or a non-traditional database. Mohammad Azam delivers a highly practical guide on how to break away from the default container setup. He demonstrates how to implement a custom DataStore in SwiftData, allowing you to intercept persistence lifecycles and route your model contexts directly to external web APIs or custom local files.
Most local AI agent setups default to giving Large Language Models access to a traditional bash shell to execute commands and write files. Alejandro Martinez decided to see what happens when you swap out the terminal for a native Swift environment. In this compelling architectural experiment, he documents how he replaced bash inside his custom macOS AI harness with an embeddable Swift interpreter. By parsing and executing Swift abstract syntax trees directly via swift-syntax, the model can generate and test pure Swift logic on the fly without needing to compile static binaries through swiftc. It is a brilliant look at building sandboxed, highly controllable runtime environments specifically tailored for LLMs.
And finally...
I can’t be the only one excited for this year’s edition of WWDC Bingo.
