Snapp Technology
Hi folks,
What a plethora of hardware announcements from Apple this week! Doing well on that end is only half of the picture for Apple. The Swift user group reported a healthy chunk of announcements in What’s new in Swift: February 2026 Edition. Xcode 26.3 finally saw its official release, and the platforms just got updated to 26.3.1
This week we’re continuing our explorations of AI, looking at some awesome reverse engineering around NSVisualEffectView, refreshing our memory on adopting Coordinators for great UIKit to SwiftUI communication and more.
Have fun!
Articles
Swift
Announcing Swift System Metrics 1.0: Process-Level Monitoring
The toolkit around Swift gained another useful package. It collects process-level system metrics like CPU utilization time and memory usage and runs on both Linux and macOS, providing a common API across platforms. Setting it up is pretty straightforward, so profiling our apps at runtime is now just a few lines of code away.
UI/UX
Reverse Engineering NSVisualEffectView
Although this article is dissecting NSVisualEffectView in particular, it’s also quite an inspiration. You know, it usually starts with an idea, then there comes the realization that our mental model may not match reality, which leads to looking for alternatives, learning a lot on the way, and ending up with something that’s worth sharing. Great, great stuff!
How to use SwiftUI Coordinators to communicate with UIKit
There are still many developers out there that need to bridge UIKit delegates back to SwiftUI on occasion. Coordinators act as the essential link between UIViewControllerRepresentable and your SwiftUI state, so here’s (a refresher on) how to utilize them.
Differentiate Without Color
Relying on color alone to convey information? Some users may have color vision deficiencies, so knowing a bit (or two) about SwiftUI’s accessibilityDifferentiateWithoutColor environment variable may help you build truly accessible SwiftUI apps.
AI/ML
Building Wispr with Kiro: A Spec-First Approach to Swift Development
Like it or not, AI-powered coding is here. There are cases where one’s approach is applicable to multiple models, and that’s exactly what we are sharing here. The author used AWS’s Kiro, but you can easily transfer the development lifecycle to other coding agents. Definitely worth reading.
Core Data Agent Skill: Now available open-source
Working on an app that utilizes CoreData? You may find this open-source AI skill useful, as it guides your agents to follow some pretty well-established patterns.
Talks
How to build apps with AI: Xcode, Claude, Codex, and more!
If you have to watch just one tutorial on getting around the basics of what building apps with AI looks like, let it be this one. Paul is a proven tutor with a lot of books in his portfolio, well known in the community, and his streams are packed with great stuff - see for yourself!
Utils
Sosumi.ai - Making Apple docs AI-readable
In their own words:
This service translates Apple Developer documentation, Human Interface Guidelines, WWDC sessions, and external Swift-DocC sites into AI-friendly Markdown.
Access it in your browser, over MCP, from the command line, as an AI skill, or with an unofficial Chrome extension.
We have tried both the MCP approach and linking the plain URLs in some AGENTS.md / SKILLS.md, and we can report that both work just great. All we had to do was to follow the instructions :)