At WWDC25 we launched a new type of Lab event for the developer community - Group Labs. A Group Lab is a panel Q&A designed for a large audience of developers. Group Labs are a unique opportunity for the community to submit questions directly to a panel of Apple engineers and designers. Here are the highlights from the WWDC25 Group Lab for watchOS (part 1). 1. I'm really excited about the new design system on all platforms. Liquid Glass is super cool. What do developers need to keep in mind when building for watchOS 26? To adopt the new design system, start with updating your app for watchOS 10 – If you have done so, your app will be mostly ready for watchOS 26. For more information, see Design and build apps for WatchOS 10. You can then look into Liquid Glass specific APIs to fine tune your app. This topic is covered in Adopting Liquid Glass. If you have SwiftUI views using any custom style, make sure they are still legible and fit with the new design system. 2. Something that really stood out to me were updates to the Smart Stack, with the system prioritizing Widgets when they're most relevant. Tell me more about these new opportunities for apps. Workout apps that record workouts using HealthKit may be automatically suggested on the watch face and appear in the Smart Stack without adding a widget. Relevant widgets are a great way to present information related to a date, location, point-of-interest type, sleep schedule, or fitness condition in the Smart Stack when it is relevant. Relevant widgets don't need to display a empty state view when they are not relevant. They are only shown in the Smart Stack when relevant. The watchOS 26 Design ToolKit in the Apple Design Resources includes a set of templates that you can use to layout your widgets. 3. Is the Wrist Flick gesture available to developers in the same way as Double Tap is? The system uses Wrist Flick to dismiss notifications and incoming calls, silence timers and alarms, or return to the watch face. There is no separate API for the Wrist Flick gesture. Apps that are using XCUIAutomation to make sure their user interface behaves as intended can use the XCUIDeviceHandGesture.flick to automate tests that verify that their app responds appropriately to the Wrist Flick gesture. For apps using automated testing, the XCUIDeviceHandGesture.doubleTap can be also be used to automate testing of the app with the Double Tap gesture. See XCUIDevice.perform(handGesture:) 4. Can HRV measurements be triggered on demand via API in watchOS? Guidelines or processes for enabling energy-intensive biometric sampling on development devices for IRB-approved research? You don’t have direct control on the sampling rate in watchOS. You can use HealthKit (HKQuantityTypeIdentifierHeartRateVariabilitySDNN, to be specific) to query the HRV data, once the system has sampled and persisted the data to HealthKit. If that doesn’t help, we suggest that you file a feedback report with your concrete use case for us to investigate. Specific to IRB-approved research using Apple Watch or its companion iPhone, you might want to look at this FAQ and SensorKit to see if they can be of any help. 5. What is the best advice for someone who is new to making a watchOS app that’s been on iOS and iPadOS? You can start with exploring the system experience features on watchOS, such as notifications, controls, and widgets, and getting familiar with the system spaces, like Smart Stack, watch face, and control center. Knowing the watchOS app design principles and practices is important as well. Design and build apps for WatchOS 10 is a great resource for this topic. SwiftUI is an amazing across-platform framework, and you will use it to create your watchOS app. If you're already using it, great! Keep in mind some watch-only constraints. Comparing to iPhone or iPad, Apple Watch has a limited battery and smaller screen size, which significantly impacts how people use your app and how your app works. 6. Was there any extension this year to the 7 day limit on querying Apple Health data on the watch? There is no change on the limit this year. You can get this official limit at runtime using earliestPermittedSampleDate. There are some exceptions, and so don't be surprised if you see some data types are retained longer. The companion iPhone holds the full set of the health data. If you need to access the health data that has been purged from the Apple Watch, consider doing it with your iOS app, and then passing the result to your watchOS app.

My phone turns or gets connected to Beats, Aipods max 2, and sometimes the Marshall Bluetooth device but it keeps triggering the automation for “When Connected To Bluetooth” - My Car tried deleting and recreating the automation and it’s still buggy.

My use case is a Repeat Timer app. The user can configure a repeating timer, say: 8 minutes with 3 sets So I would like to configure either 3 alarms (8mins, 16mins, 24mins) or a repeating alarm (8mins every 3 mins) BUT... I would like the first and second alarms to break through but only for 5 seconds for example, and then stop (so the user doesn't need to tap the screen to silence the alarm). I don't think this is possible after reading the API docs, but am asking the question anyway. Thanks for any advice or guidance here, and happy WWDC!

My app uses App Intents to create App Shortcuts. When I build and run my app in Xcode, the App Shortcuts Preview tool (under Product menu) shows the following message: No Flexible Matching Assets This target is for a platform which is not supported by Flexible Matching or does not have Flexible Matching enabled. All of my project's targets are iPhone only with a minimum deployment of 18.0. In the build settings for this project, Enable App Shortcuts Flexible Matching is set to Yes. (build settings reference) Any guidance on how to troubleshoot this? Thank you!

After a user has launched an App Clip Experience from Safari and is done with the activity that the App Clip provides – is there a way for the App Clip to programatically close itself so the user returns to Safari?

Hello, We are using NearbyInteraction to get the distance between an iPhone and Apple Watch. Overall it works pretty well, but we are noticing that it sometimes produces duplicate (or near duplicate) timestamps. Here is an example of the data produced (time is in seconds since scan start): 0: 1.882862 1: 1.88296 2: 1.981459 3: 1.981495 4: 2.080235 5: 2.080397 As you can see there's a pattern where two updates have very similar timestamps. This seems to only happen when using 2 U2 chips together. Is this an expected behavior? If so, why? The timestamps are calculated using: let bootTimeInterval = bootTime() func bootTime() -> TimeInterval { let uptime = ProcessInfo.processInfo.systemUptime let nowTimeIntervalSince1970 = Date().timeIntervalSince1970 return nowTimeIntervalSince1970 - uptime } // Then later on each NISession update... let timestamp = bootTimeInterval + ProcessInfo.processInfo.systemUptime The app produces unix timestamps but I have converted them to be relative to the start of the scan to be easier to read. Note that bootTimeInterval is calculated once at the start of the app, so all timestamps produced on a UWB update should be using the same clock. Things I've confirmed: The NISession is only able to connect to one device, so it is not coming from a different device. Any feedback is appreciated, thank you.

I'm encountering what appears to be a specific precedence behavior with ManagedSettingsStore.shield and would appreciate some further clarification. My current understanding is that category-level shields take precedence over individual app allowances. My test involved... Using FamilyActivityPicker to select a single target application (e.g., "Calculator," which falls under the "Utilities" category). Using FamilyActivityPicker again to select the category of that target application. I applied shields using ManagedSettingsStore (named .individual): store.shield.applicationCategories = .specific(Set([utilitiesCategoryToken])) store.shield.applications = Set([calculatorApplicationToken]) Result: The calculator app remains shielded, suggesting that the category-level shield on Utilities overrides the attempt to allow the individual app. I also tried this using a single picker, but received only the category token instead of all application tokens in that category. Is this observed precedence (where store.shield.applicationCategories effectively overrides store.shield.applications for apps within the shielded category) the intended behavior? If so, are there any mechanisms available within the main app's capabilities (potentially using a Device Activity Report Extension or Shield Extension) to allow a specific ApplicationToken if its corresponding ActivityCategoryToken is part of the store.shield.applicationCategories set? Essentially, can store.shield.applications be used to create "allow exceptions" for individual apps that fall into an otherwise shielded category? Additionally, I mentioned that selecting an entire category in the picker only returns the opaque category token, not any application tokens. Is there any way in which I could return both the category and all application tokens by just selecting the category? Any insights or pointers would be greatly appreciated!

Is there a way to increase the frequency of UWB background ranging?

When fetching Apple Savings transactions via FinanceKit, the data is missing key context about where the deposit originated from. Here’s an example transaction I retrieved: Transaction( id: 77371A0C-7122-42C7-BEBC-85BDF654AD2B, accountID: 68D9FE9D-6DA6-4A27-BB9D-19209CD29A56, transactionAmount: 1.46 USD, creditDebitIndicator: .credit, transactionDescription: "Deposit", originalTransactionDescription: "", merchantCategoryCode: nil, merchantName: nil, transactionType: .deposit, status: .booked, transactionDate: 2025-08-20 13:44:26 +0000, postedDate: 2025-08-20 13:44:26 +0000 ) As shown: transactionDescription is just "Deposit" merchantName is nil No indication of the source account In contrast, the Wallet app clearly shows context for Apple Savings account deposits, such as: "Apple Card" (daily cash) "Bank of America" (external transfer) "Interest Paid" (we do see "Interest" come through correctly) Without this metadata, third-party apps cannot replicate Wallet’s clarity about where a deposit came from. Every deposit simply appears as "Deposit", which is ambiguous. Request: Please expose additional metadata for Apple Savings account transactions, for example: sourceAccountName (e.g. “Apple Card” or “Bank of America”) transactionOriginType (cashback, external bank transfer, interest) institutionIdentifier or similar for external banks This would allow developers to show clear, Wallet-quality transaction details and avoid confusing users. Impact: The lack of source info makes Savings deposits nearly indistinguishable from one another, even though Wallet provides this context. For apps leveraging FinanceKit, this results in a poorer experience compared to Apple’s own Wallet. Thanks!

I am developing an app that allows the user to ask it to process the clipboard contents and do something with it. In developing a XC UI Test, I find the app stops while it waits for the user to give permission. That breaks the automation. I tried: let springboard = XCUIApplication(bundleIdentifier: "com.apple.springboard") let allowButton = springboard.buttons["Allow Paste"] But that does not work. Is there a way to tell the framework to automatically give the test permission to access the Paste clipboard (or to allow me to write tests to grant this)?

Hi Team, We’re encountering a device-specific issue with our SMS Message Filter extension. The extension works as expected on an iPhone 11 running iOS 16.6, but it does not trigger on an iPhone 12 Pro running iOS 16.7. Key Observations: The extension is implemented using ILMessageFilterExtension and calls messageFilterOffline(appGroupIdentifier:for:) from our shared library. The App Group is properly configured and accessible across the app and extension. The extension is enabled under Settings > Messages > Unknown & Spam. There are no crashes or error logs reported on the affected device. The issue is consistently reproducible — it works on one device but not the other. We’re wondering if this could be a regression or a device-specific behavior change introduced in iOS 16.7. Has anyone encountered similar inconsistencies in Message Filter extensions across different iOS versions or device models? Any guidance or suggestions would be greatly appreciated. Thanks in advance!

Can someone please give me a clue who to go about creating an automation that will open an app click log in, close 2 pop up windows the select an item on screen spin a when collect my prize then log out of the app. I want it to do this once a hour.

Hi everyone, I observed a behavior with AirPrint from an iPhone and wanted to confirm if this is expected behavior from iOS. Scenario tested: File type: Photo Print-quality selected by the user: Normal Observation (from packet capture): When checking the PCAP for the request sent from the iPhone, the print-quality attribute is always sent as high, even though the user selected Normal in the UI. Question: Is this an expected behavior in iOS/AirPrint where photos are always sent with print-quality=high regardless of the user-selected print quality? Or could this be a bug?

"Bu beta şu anda yeni testçileri kabul etmiyor" Lütfen bu hatayı düzeltin

My back camera freezes after first frame is displayed on the screen. Issue is present in all the apps that use the back camera. Same issue with front camera, however front camera works with Facetime and WhatsApp but not the native camera app.

Hello? Does anyone have a program to "Create Shortcuts" in Swift? With: Program Name: Program Path: Program Image:

We developed a "Default Translation App" following the guide: https://aninterestingwebsite.com/documentation/translationuiprovider/preparing-your-app-to-be-the-default-translation-app. I have already configured everything that needs to be configured according to the document, but there is still this problem

I wonder how one would use IOBluetoothHandsFree APIs to interact from macOS app with a bluetooth device that implements bluetooth hands free profile. My current observation is as follows: IOBluetoothDevice object representing the device correctly identifies it as a hands free device, i.e.: there is a proper record in services array, that matches the kBluetoothSDPUUID16ServiceClassHandsFree uuid, the IOBluetoothDevice handsFreeDevice property returns 1 Attempt to create IOBluetoothHandsFreeDevice using IOBluetoothDevice as described above (i.e. [[IOBluetoothHandsFreeDevice alloc] initWithDevice: myIOBluetoothDeviceThatHasHandsFreeDevicePropertySetTo1 delegate: self]) results in the following output in debugger console: SRS-XB20 is not a hands free device but trying anyways. Subsequent call to connect on an object constructed as above results in the following stream of messages: API MISUSE: <CBClassicPeer: 0x1442447b0 6D801974-5457-9ECE-0A9B-8343EC4F60AA, SRS-XB20, connected, Paired, b8:d5:0b:03:62:70, devType: 19, PID: 0x1582, VID: 0x0039> Invalid RFCOMM CID -[IOBluetoothRFCOMMChannel setupRFCOMMChannelForDevice] No channel <IOBluetoothRFCOMMChannel: 0x600003e5de00 SRS-XB20, b8-d5-0b-03-62-70, CID: 0, UUID: 110F > AddInstanceForFactory: No factory registered for id <CFUUID 0x600000b5e3e0> F8BB1C28-BAE8-11D6-9C31-00039315CD46 -[IOBluetoothRFCOMMChannel setupRFCOMMChannelForDevice] No channel <IOBluetoothRFCOMMChannel: 0x600003e5de00 SRS-XB20, b8-d5-0b-03-62-70, CID: 0, UUID: 110F > API MISUSE: <CBClassicPeer: 0x1442447b0 6D801974-5457-9ECE-0A9B-8343EC4F60AA, SRS-XB20, connected, Paired, b8:d5:0b:03:62:70, devType: 19, PID: 0x1582, VID: 0x0039> Invalid RFCOMM CID Note that this device's handsFreeServiceRecord looks as follows: ServiceName: Hands-free unit RFCOMM ChannelID: 1 Attributes: { 0 = "uint32(65539)"; 256 = "string(Hands-free unit)"; 9 = "{ { uuid32(00 00 11 1e), uint32(262) } }"; 785 = "uint32(63)"; 1 = "uuid32(00 00 11 1e)"; 6 = "{ uint32(25966), uint32(106), uint32(256) }"; 4 = "{ { uuid32(00 00 01 00) }, { uuid32(00 00 00 03), uint32(1) } }"; } and explicit attempt to open RFCOMM channel no 1 ends like this: WARNING: Unknown error: 911 Failed to open RFCOMM channel -[IOBluetoothRFCOMMChannel setupRFCOMMChannelForDevice] No channel <IOBluetoothRFCOMMChannel: 0x6000002036c0 SRS-XB20, b8-d5-0b-03-62-70, CID: 1, UUID: 111E > AddInstanceForFactory: No factory registered for id <CFUUID 0x600003719260> F8BB1C28-BAE8-11D6-9C31-00039315CD46 -[IOBluetoothRFCOMMChannel waitforChanneOpen] CID:1 - timed out waiting to open -[IOBluetoothDevice openRFCOMMChannelSync:withChannelID:delegate:] CID:1 error -536870212 call returned: -536870212

Hello. Our application performs encoding and decoding of large json files which sometimes may take up to 1,5GB of RAM memory. We understand that this may be a problem for devices with low RAM memory (3GB). But before iOS 26.4 we didn't have much occurrences (3 for the last 90 days). Starting from iOS 26.4 it started to crash a lot. Can the reason be that iOS 26.4 occupies more RAM memory so there is less memory left for our app? Or maybe starting from iOS 26.4 there is less RAM memory allocated per app? The crash message is Fatal error: failed to allocate 392 bytes of memory with alignment 8

I have noticed that there are no new kdk builds since mid December Is everything ok at apple I am aware of NXUs but the kernel version has changed much since then