When running a mobile app through Xcode, why can't the associated watch app be installed within the watch app? How can this problem be solved?

quencoder 2.5 locally is making great plans. Beautiful plans amazing plans that are so correct I want to hug it. But when I want to implement it there is no button asking me if I want to go ahead. I tried /exit-plan and the command could not be completed (I only see /plan, /rename, and /exit-planP. I do not see the markdown in a separate pane that's copyable as shown in this video https://developer.apple.com/videos/play/wwdc2026/259/?time=597 (I also can't get the conversation in a separate window by pressing option shift click, I have to manually open the conversation in a new tab). I get a plan interleaved with code snippets and I can't copy the whole plan to a new conversation. I would very much like this to work. (I also really like being able to run locally because I do not agree with google, openai, or anthropic ethics. thanks for giving me a choice) (I also like how not like a sycophant the tone of the response is. refreshing) (Note I started planning first because without a plan it started mucking about in the plist editing status bar keys when the problem was a navigation bar which I thought was clear from the picture I attached) (But I'd really like to say go and I have no buttons to exit plan mode.) (I am going to stay on Tahoe till October) I asked it how to copy and it said: In the opencode session where the plan was discussed, type /share — it'll copy a shareable link to your clipboard (like opncd.ai/s/...). Paste that link here and I can fetch the plan directly. That was a hallucination when I typed /share the agent just repeated its plan for a third time. It's a good plan, the best plan, I don't see any algae in this plan. It also said "I can see from the repository exploration that there's no written plan document, and the other agent is working on "[the plan]." However, I don't have access to the previous conversation's last response. Maybe there should be a button to copy a whole response or open a new conversation (not in plan mode) with a response? I see that I can configure agents https://opencode.ai/docs/agents/. Should I? Shouldn't Xcode do it for me? { "$schema": "https://opencode.ai/config.json", "agent": { "build": { "mode": "primary", "model": "anthropic/claude-sonnet-4-20250514", "prompt": "{file:./prompts/build.txt}", "permission": { "edit": "allow", "bash": "allow" } }, "plan": { "mode": "primary", "model": "anthropic/claude-haiku-4-20250514", "permission": { "edit": "deny", "bash": "deny" } }, "code-reviewer": { "description": "Reviews code for best practices and potential issues", "mode": "subagent", "model": "anthropic/claude-sonnet-4-20250514", "prompt": "You are a code reviewer. Focus on security, performance, and maintainability.", "permission": { "edit": "deny" } } } }

How do I select between GPT and Claude, and the model, when opening a new task for the coding agent with the ‘New Conversation’ button in the toolbar?

I've created a basic Multiplatform Project in Xcode, by going to File > New > Project. I've also included the default widget extension target. When I run the widget on Mac, and I control-click and choose "Edit Widget" the widget just kinda turns gray and floats over other windows, but does not let me configure anything. We leave the option to use configure with app intents turned on when adding the widget extension target. When I run the same project on an iPhone simulator, I can configure the widget without any issue. If someone else has a few minutes, can they see if this repeats for them? We are running this test because we are seeing the same experience in a real widget we are trying to develop but we can't get it to go into configuration mode on Mac.

I can not seem to get ll the pieces working here. Has anyone set up grok? URL - https://api.x.ai/v1/chat/completions api-key ---- API Key Header x-api-key Doesn't work. Any help would be greatly appreciated. Thanks, Rob

Title: Xcode 27 beta 27A5194q: aimodelc reports “Core AI requires the Metal Toolchain” although Metal Toolchain is installed Description: I am trying to compile a Core AI asset using Xcode 27 beta on macOS 27. CoreAI.framework is available and the Metal Toolchain component is installed, but aimodelc refuses to compile with: error: Core AI requires the Metal Toolchain. note: Download it from Xcode > Settings > Components > Other Components. Environment: macOS 27 beta Xcode 27.0 beta Xcode build: 27A5194q Metal Toolchain installed build: 27A5194o Apple Silicon Mac Diagnostics: xcodebuild -version reports: Xcode 27.0 Build version 27A5194q xcodebuild -showComponent MetalToolchain -json reports: { "buildVersion": "27A5194o", "status": "installed", "toolchainIdentifier": "com.apple.dt.toolchain.Metal.32023.917", "toolchainSearchPath": "/private/var/run/com.apple.security.cryptexd/mnt/com.apple.MobileAsset.MetalToolchain-v27.1.5194.15.4aqDVJ" } The actual tools resolve: xcrun --find aimodelc /Applications/Xcode-beta.app/Contents/Developer/usr/bin/aimodelc xcrun --find coreai-build /var/run/com.apple.security.cryptexd/mnt/com.apple.MobileAsset.MetalToolchain-v27.1.5194.15.4aqDVJ/Metal.xctoolchain/usr/bin/coreai-build xcrun --find metallib /var/run/com.apple.security.cryptexd/mnt/com.apple.MobileAsset.MetalToolchain-v27.1.5194.15.4aqDVJ/Metal.xctoolchain/usr/bin/metallib Reproduction: Install Xcode 27 beta build 27A5194q. Install Metal Toolchain from Xcode > Settings > Components > Other Components. Confirm xcodebuild -showComponent MetalToolchain -json reports Metal Toolchain installed. Attempt to compile a Core ML package to Core AI: xcrun aimodelc compile --output /private/tmp/vesta-coreai-smoke/ManualPatchCheck.coreai /private/tmp/vesta-coreai-smoke/AFM27CoreAISmoke.mlpackage Actual result: aimodelc fails immediately: error: Core AI requires the Metal Toolchain. note: Download it from Xcode > Settings > Components > Other Components. Expected result: aimodelc should recognize the installed Metal Toolchain and compile the Core AI asset, or Xcode should provide a Metal Toolchain component compatible with Xcode build 27A5194q. Additional details: Running: xcodebuild -downloadComponent MetalToolchain downloads and installs Metal Toolchain 27A5194o. Running: xcodebuild -downloadComponent MetalToolchain -buildVersion 27A5194q fails with: xcodebuild: error: Failed fetching catalog for assetType (com.apple.MobileAsset.MetalToolchain), serverParameters ({ RequestedBuild = 27A5194q; }) This suggests Apple’s component catalog currently provides Metal Toolchain 27A5194o while Xcode is build 27A5194q, and aimodelc does not accept this installed component as compatible. Please confirm whether Metal Toolchain 27A5194o is intended to work with Xcode 27A5194q. If not, please publish the matching Metal Toolchain component or adjust aimodelc/component compatibility detection.

I am receiving this error when uploading to AppStore Connect: Validation failed This bundle Payload/Document IT.app/Frameworks/iOSAppSDK.framework is invalid. The Info.plist file is missing the required key: CFBundleVersion. Please find more information about CFBundleVersion at https://developer.apple.com/documentation/bundleresources/information_property_list/cfbundleversion (ID: b4774d8b-1f34-4c18-90d9-d47a76d3c3dd) My PLIST in Xcode appears to be correct, and my Build Settings appear to be correct. Any ideas why I am seeing this issue? I uploaded a version for the iPhone that uploaded correctly.

I can't use cloud-based AI providers, so I'm very excited about the potential of local models with Xcode's Coding Assistant. I've been playing around with a local MLX model, and have been generally pretty satisfied with it. But I'm curious - what features am I missing out on by using a local model instead of a larger, cloud-based model? Is Xcode's ability to run tests, build the project, etc. limited by the use of a local model? Does it limit the ability to do more complex agentic workflows? If so, how? Basically, local models in Xcode seem to be working pretty well for me, and I'm curious to know what cloud-based models can offer that I don't already have.

Description I've encountered a consistent hang/freeze issue in SwiftUI applications when using nested LazyVStack containers with Accessibility Inspector (simulator) or VoiceOver (physical device) enabled. The application becomes completely unresponsive and must be force-quit. Importantly, this hang occurs in a minimal SwiftUI project with no third-party dependencies, suggesting this is a framework-level issue with the interaction between SwiftUI's lazy view lifecycle and the accessibility system. Reproduction Steps I've created a minimal reproduction project available here: https://github.com/pendo-io/SwiftUI_Hang_Reproduction To Reproduce: Create a SwiftUI view with the following nested LazyVStack structure: struct NestedLazyVStackView: View { @State private var outerSections: [Int] = [] @State private var innerRows: [Int: [Int]] = [:] var body: some View { ScrollView { LazyVStack(alignment: .leading, spacing: 24) { ForEach(outerSections, id: \.self) { section in VStack(alignment: .leading, spacing: 8) { Text("Section #\(section)") // Nested LazyVStack LazyVStack(alignment: .leading, spacing: 2) { ForEach(innerRows[section] ?? [], id: \.self) { row in Text("Section #\(section) - Row #\(row)") .onAppear { // Load more data when row appears loadMoreInner(section: section) } } } } .onAppear { // Load more sections when section appears loadMoreOuter() } } } } } } Enable Accessibility Inspector in iOS Simulator: Xcode → Open Developer Tool → Accessibility Inspector Select your running simulator Enable Inspection mode (eye icon) Navigate to the view and start scrolling Result: The application hangs and becomes unresponsive within a few seconds of scrolling Expected Behavior The application should remain responsive when Accessibility Inspector or VoiceOver is enabled, allowing users to scroll through nested lazy containers without freezing. Actual Behavior The application freezes/hangs completely CPU usage may spike The app must be force-quit to recover The hang occurs consistently and is reproducible Workaround 1: Replace inner LazyVStack with VStack LazyVStack { ForEach(...) { section in VStack { // ← Changed from LazyVStack ForEach(...) { row in ... } } } } Workaround 2: Embed in TabView TabView { NavigationStack { NestedLazyVStackView() // ← Same nested structure, but no hang } .tabItem { ... } } Interestingly, wrapping the entire navigation stack in a TabView prevents the hang entirely, even with the nested LazyVStack structure intact. Questions for Apple Is there a known issue with nested LazyVStack containers and accessibility traversal? Why does wrapping the view in a TabView prevent the hang? Are there recommended patterns for using nested lazy containers with accessibility support? Is this a timing issue, a deadlock, or an infinite loop in the accessibility system? Why that happens? Reproduction Project A complete, minimal reproduction project is available at: https://github.com/pendo-io/SwiftUI_Hang_Reproduction

How can I enable an external display in the Device Hub via Xcode 27, similar to the external display option previously available in the Simulator?

Hello, I have been unable to install any development-signed app on any physical device for five months. Builds succeed, code signing passes locally, but every device rejects the app at install time with: Failed to verify code signature of .../extracted/MyApp.app : 0xe8008018 (The identity used to sign the executable is no longer valid.) ApplicationVerificationFailed The app installs briefly, then iOS immediately removes it. This started right after my account (Team ID MB4DXDTDMT) was reinstated following a duplicate-account flag. Background: I had a personal account that was converted to a business account (Wakeout LLC), then created a new personal account, which Apple flagged as a duplicate and later reinstated. The signing failure began immediately after that reinstatement. Isolation already done (this is not a local-setup problem) I have run the full isolation sequence — including every step DTS typically asks for — and the result points squarely at the account/team, not my machine: New blank Xcode project, automatic signing, new bundle ID → same 0xe8008018. Brand-new macOS user account → same failure. Multiple Macs, fresh Xcode installs → same failure. Multiple iOS devices (iPhone 17 Pro, iPhone 15 Pro, others) → same failure. Different Apple ID / different developer team on the same Mac + same device → installs fine. This is the decisive one: the local environment is healthy; only Team MB4DXDTDMT is rejected. Xcode Cloud builds for this same team install fine. Apple's cloud signing trusts MB4DXDTDMT; the device-verification backend does not. That gap can only exist server-side. I have also: revoked/regenerated all certificates multiple times, deleted/recreated all provisioning profiles, cleared ~/Library/MobileDevice/Provisioning Profiles, cleared DerivedData and CoreDevice, removed device pairing records, re-paired devices, confirmed Developer Mode and correct system time. Simulators work. codesign --verify --deep --strict passes. Profile certificate SHA-1 matches the signing cert exactly. Entitlements match. Why I'm posting here This is the same failure documented in thread 755762, where Quinn concluded: "this seems to be tied to your primary developer account and only DevPrograms has access to those details." That matches my evidence exactly: the problem isolates cleanly to one team, and only DevPrograms can see the account-side state. I've already gone through Developer Support on this — an open case has been with them for about five months without a resolution, which is what convinced me the fix isn't something I can reach from the support side. I'm posting here in case a DTS engineer can confirm the diagnosis and point me to the right path. Question for any DTS engineer: given that the failure isolates to a single team — different teams sign and install fine on the same Mac and same device, and Xcode Cloud builds for this same team install fine — can you confirm this is an account-side signing-trust state that has to be reset by Apple, and what's the most direct way to get that reset actioned? Happy to attach a sysdiagnose, full console output, or codesign -dvvv dumps on request. Thank you.

Is there a known workaround (other than forking) for watchOS builds in Xcode 27 that depend on packages without a declared watchOS 9.0 floor? I depend on several Apple SPM packages with no minimum deployment target, such as https://github.com/apple/swift-algorithms. My project's minimum is 26.0. Xcode 27 will not build watchOS targets with these packages because watchOS 8.0 is below its range of supported targets. /.../SourcePackages/checkouts/swift-algorithms/Package.swift The watchOS Simulator deployment target 'WATCHOS_DEPLOYMENT_TARGET' is set to 8.0, but the range of supported deployment target versions is 9.0 to 27.0.x.

I’m seeing a reproducible issue in both Xcode 26 and Xcode 27 where Coding Assistant appears to stop autocompleting when a project contains a larger number of Swift Package dependencies. Symptoms: Autocomplete suggestions from Coding Assistant stop appearing. File targeting with @ stops working. The issue typically appears shortly after opening the project If I start a brand-new chat immediately after opening the project, Coding Assistant sometimes works briefly before failing Reducing or removing Swift Package dependencies fixes the issue I’ve created a sample project that reproduces the problem and have also filed Feedback Assistant report: FB23341243 The issue seems related to package count or project complexity rather than any specific package, although I haven’t yet identified the exact threshold that triggers it. Has anyone else run into this or has anyone found a workaround?

iPhoneSimulator.app used to have the ability to show an external display / CarPlay display, but this feature does not appear to be available in DeviceHub.app, making it difficult (/impossible :( ) to test CarPlay features from the simulator.

Hi all, The minimum deployment target for Xcode 27 seems to be macOS 12. While the Changelog states it should be 11. The minimum supported deployment target on macOS has been increased to 11.0. https://developer.apple.com/documentation/xcode-release-notes/xcode-27-release-notes#Deprecations-in-Xcode-27-Beta Which one will it finally be? We need this information to make a product portfolio decision. I already opened a Radar addressing the mismatch between documentation and actual shipped SDK, see FB23106692. I waited to post here if the conflict get's resolved in Xcode Beta 2. But the issue still persists.

When working with an Objective-C project in Xcode 26.5, I'm unable to establish outlet and action connections by dragging between XIB files and the corresponding .h header files. This drag-and-drop connection works perfectly after switching to Xcode 26.3.0. In addition, under Xcode 26.5, drag-and-drop linking works normally for projects using Swift files together with XIB files. The issue only occurs with Objective-C .h files and XIB files.

A empty macos project created using xcode. run build failed with the following error message. nothing in detail. retry no help. seems indexing takes forever!! The Xcode build system has terminated due to an error. Build again to continue.

Summary: After upgrading to Xcode 26.4, app-hosted XCTest execution on physical devices running iOS 16 fails at test-planning time with "Logic Testing Unavailable." The same project and same device work under Xcode 26.2. The failure reproduces with a standalone minimal Xcode project, which suggests this is an Xcode regression rather than a project-configuration issue. Environment: Xcode: 26.4 (17E192) macOS: Tahoe 26.4 Failing device: iPhone 11 (iPhone12,1), iOS 16.0.3 Working comparisons: Xcode 26.2 + same iPhone 11 (iOS 16.0.3): works Xcode 26.4 + iPad Pro 11-inch (4th generation) on iOS 26.3: works Regression: Yes. Works on Xcode 26.2. Fails on Xcode 26.4. Same project, same signing setup, same physical iOS 16 device. Minimal reproduction: A minimal sample project with: one iOS app target one app-hosted unit-test target one UI-test target (for comparison; not required to reproduce) Steps for the unit-test repro on a physical device running iOS 16: Build for testing: xcodebuild -project TestProject.xcodeproj \ -scheme TestProject-UnitTests \ -sdk iphoneos \ -destination 'platform=iOS,id=<DEVICE_UDID>' \ -derivedDataPath DerivedData-Device-Unit \ -resultBundlePath Results-Device-Unit-Build.xcresult \ DEVELOPMENT_TEAM=<TEAM_ID> \ CODE_SIGN_STYLE=Automatic \ build-for-testing Run tests without building: xcodebuild -project TestProject.xcodeproj \ -scheme TestProject-UnitTests \ -sdk iphoneos \ -destination 'platform=iOS,id=<DEVICE_UDID>' \ -derivedDataPath DerivedData-Device-Unit \ -resultBundlePath Results-Device-Unit-Test.xcresult \ DEVELOPMENT_TEAM=<TEAM_ID> \ CODE_SIGN_STYLE=Automatic \ test-without-building Note: xcodebuild test (combined build and test) also fails with the same error. Expected result: The hosted unit tests run successfully on the connected physical device, as they do under Xcode 26.2. Actual result: Step 1 (build-for-testing) succeeds. Step 2 (test-without-building) fails immediately (within ~1 second) with: 2026-03-26 11:23:28.683 xcodebuild[51930:731004] DVTDeviceOperation: Encountered a build number "" that is incompatible with DVTBuildVersion. 2026-03-26 11:23:28.725 xcodebuild[51930:730966] [MT] DVTDeviceOperation: Encountered a build number "" that is incompatible with DVTBuildVersion. 2026-03-26 11:23:29.239 xcodebuild[51930:730966] Writing error result bundle to /var/folders/jl/knmkq18x4cg_3w087zgpfldm0000gn/T/ResultBundle_2026-26-03_11-23-0029.xcresult xcodebuild: error: Failed to build project TestProject with scheme TestProject-UnitTests.: Cannot test target “redacted” on “redacted”: Logic Testing Unavailable Logic Testing on iOS devices is not supported. You can run logic tests on the Simulator. Why this looks like an Xcode regression: The same project and same physical iOS 16.0.3 device work under Xcode 26.2. Under Xcode 26.4, build-for-testing still succeeds, so signing, provisioning, and bundle construction appear valid. The failure happens only when Xcode plans or launches the on-device test run. Before the failure, Xcode 26.4 logs: DVTDeviceOperation: Encountered a build number "" that is incompatible with DVTBuildVersion. A newer physical device running iOS 26.3 works under Xcode 26.4 without this warning. The issue also reproduces with a minimal standalone Xcode project, which makes a project-specific configuration problem unlikely. The Xcode-generated .xctestrun files for the passing Xcode 26.2 case and failing Xcode 26.4 case are structurally equivalent, making an .xctestrun format difference unlikely to be the cause.

Hello! I have windows pc and an iphone, I want to develop apps and use xcode... Please help me on that and what should I doThank YOU !

Environment: MacBook Air M4 macOS 27 Golden Gate Beta Xcode 27 Beta iPhone running iOS 27 Beta Apple Watch paired to the iPhone Steps to Reproduce: Install macOS 27 Beta and Xcode 27 Beta. Pair an Apple Watch with an iPhone running iOS 27 Beta. Enable Developer Mode on the iPhone. Connect the iPhone to the Mac and launch Xcode. Open Window → Devices and Simulators. Attempt to locate the paired Apple Watch or enable Developer Mode on the watch. Expected Result: The paired Apple Watch appears in Xcode for development purposes. A Developer Mode option is available on the Apple Watch. Xcode is able to recognize and communicate with the watch through the paired iPhone. Actual Result: The Apple Watch does not appear in Xcode. No Developer Mode option is visible on the watch. The iPhone is successfully paired and running iOS 27 Beta, but the watch remains unavailable for development workflows. Additional Notes: The iPhone and Apple Watch are already paired and functioning normally. Developer Mode is enabled on the iPhone. Restarting devices and reconnecting the iPhone did not resolve the issue. Unsure whether this is a limitation of the current beta builds or a regression in Xcode/watchOS development tooling.