Apple's strict contracts for App schema domains are great if you have something which fits into that domain. There are endless options with AppSchema domains outside that of what apple have created. Does anyone know if apple will open the door to custom AppSchema domains? This would be a "game-changer". Is there any insight on what the future holds?

i cant use mlx on swift for some reason, i would like for them to add the support to add it as a package

Hi guys, I'm testing the Foundation Models Framework with the on-device model in iOS 27 (beta 4) and macOS 27 (beta 4) and is completely failing to respond. There are many errors. For starters, the model doesn't respond to prompts directly, you need to specify instructions, otherwise it refuses to provide an answer. It is always looking for tools, even when no tool has been provided, and returns an error saying that it couldn't find the tool. Then, when it produces a response, it shows all the thinking process first, which completely ruins the response. Most of the time, the response begins with all the JSON code. And when I try to have a long conversation, it just says "I cannot write content or generate text." I wonder if someone is experiencing the same issues or maybe the way to implement this model changed and I'm missing something? Here is a screenshot of one of my interactions when I asked the model to describe a unicorn. It tried to access a tool that doesn't exist. (the app just prints the value of the content property) Here is the code. It is performing a simple request. struct ContentView: View { @State private var response = "" var body: some View { VStack { Button("Send") { let prompt = "Write a paragraph describing a unicorn" let session = LanguageModelSession { "Respond to the user's request. Never acknowledge the request, add preamble, or comment on what you are about to write." } if !session.isResponding { Task { do { let answer = try await session.respond(to: prompt) response = answer.content } catch { response = "Error accessing the model: \(error)" } } } } .buttonStyle(.borderedProminent) Text(response) .font(Font.system(size: 18)) .padding() Spacer() } .padding() } }

On macOS Golden Gate Developer Beta 4 the following code: import CoreSpotlight import FoundationModels let tool = SpotlightSearchTool() let session = LanguageModelSession(tools: [tool]) let response = try await session.respond(to: "What hikes have I gone on?") , returns the following error: Model Catalog error: Error Domain=com.apple.UnifiedAssetFramework Code=5000 "There are no underlying assets (neither atomic instance nor asset roots) for consistency token for asset set com.apple.modelcatalog" UserInfo={NSLocalizedFailureReason=There are no underlying assets (neither atomic instance nor asset roots) for consistency token for asset set com.apple.modelcatalog} , although the model is available in general and can return responses without using the tool. The code: print(SystemLanguageModel.default.availability) returns 'available'. What am I doing wrong?

On watchOS, once the "Turn on [App] shortcuts with Siri?" prompt has been declined for an app, is there a user-facing way to turn it back on? On iOS I found the per-app Siri toggle, and re-enabling it restored voice invocation immediately. I cannot find any equivalent on watchOS, in the watch Shortcuts app, in watch Settings, or in the paired iPhone's Watch app. If one exists I would appreciate a pointer. If it does not, I would like to know whether that is expected. What I am seeing I have an iOS app with a companion watchOS app. Each declares its own AppShortcutsProvider and App Intent. Modern App Intents only, no SiriKit, no Intents extension. On one Apple Watch, saying the trigger phrase transcribes correctly on screen, then falls through to a generic Maps business search. The intent is never invoked. On that same watch the App Shortcut appears in the watch Shortcuts app and runs correctly when tapped, and Siri is otherwise healthy. The same phrase works on the paired iPhone. Why I believe the app is fine On a second iPhone and Apple Watch on a different Apple Account, the phrase spoken to the watch returned my intent's own dialog from perform(). Two control apps also behave correctly, including one with an iOS host plus embedded companion watch app matching my production structure, which prompted and worked on both devices. Reading Metadata.appintents/extract.actionsdata from the built watch app shows the provider registered and all phrase templates well formed. The state survives deleting the app. I deleted it from both devices, power cycled both, and reinstalled. The iPhone worked immediately without re-prompting. The watch continued to fail without prompting. Both retained their prior opt-in state. The app's bundle identifiers changed during a rename several months ago. I believe the prompt appeared on the watch then and was dismissed. Environment Apple Watch Ultra, watchOS 26.6. iPhone on iOS 18.6. Xcode 26.1.1. Deployment targets iOS 18.6 / watchOS 11.6. Companion watch app, WKRunsIndependentlyOfCompanionApp = NO. Already ruled out on device, each with a clean install and watch restart: CFBundleSpokenName, one word and spaced Watch PRODUCT_NAME / CFBundleName matched to the spoken app name Removing updateAppShortcutParameters() from the watch App.init() Adding the com.apple.developer.siri entitlement to the watch target Phrase collision between phone and watch providers, a watch-only phrase also failed Siri domain collision on the trigger words, a neutral phrase also failed Intent identifier collisions between the phone and watch bundles A brand new, uniquely named intent with canonical syntax and no gating was also ignored Questions Is there a user-facing setting on watchOS, or in the paired iPhone's Watch app, to re-enable App Shortcuts for a specific app after the prompt has been declined? If not, is there any supported recovery short of erasing the watch and setting it up as new? Restoring from backup appears to bring the state back. Is there any API for an app to detect that its App Shortcuts opt-in is off, so it can show guidance instead of a silently dead feature? INPreferences.siriAuthorizationStatus() is not a valid proxy, it reports authorized on the affected device. Happy to provide a sysdiagnose with the App Intents logging profile, or either control project.

So I'm having an issue with the FoundationModels framework but idk if this is just my feeling or not, the issue comes up after I updated my Mac into 26.6 the code was very simple actually: #Playground { let model = SystemLanguageModel.default let session = LanguageModelSession(model: model) print(model.availability) var query = "How to hide button" Task { do { let response = try await session.respond(to: query) print(response.content) } catch { print("\(error)") } } } the code works before I updated the version, but then after I updated the version it says: Error Domain=FoundationModels.LanguageModelSession.GenerationError Code=-1 "The operation couldn’t be completed. (FoundationModels.LanguageModelError error -1.)" UserInfo={NSMultipleUnderlyingErrorsKey=( "Error Domain=FoundationModels.LanguageModelError Code=-1 "(null)" UserInfo={NSMultipleUnderlyingErrorsKey=(\n "Error Domain=ModelManagerServices.ModelManagerError Code=1026 \"(null)\" UserInfo={NSMultipleUnderlyingErrorsKey=(\n)}"\n)}" ), NSLocalizedDescription=The operation couldn’t be completed. (FoundationModels.LanguageModelError error -1.)} this is runned in Xcode 26.6, additional information I have also coder 27 beta 4 installed in my Mac, is this problem occurring because the Xcode 26.6 and Xcode 27 beta 4?? can u guys help me

I’ve been developing mlx-recurrence, a plug-in framework of fused Metal GPU kernels for linear recurrences on Apple silicon—roughly analogous to flash linear attention for MLX. Sequential recurrences are difficult for MLX to fuse automatically. Architectures such as state-space models, gated linear attention, and diagonal RNNs ordinarily require a loop across the sequence length. When that loop is implemented in Python, a sequence of length L can require L separate Python-to-Metal dispatches. These kernels instead execute the entire recurrence in a single Metal dispatch. The training path uses segment checkpointing with recomputation during the backward pass. In validated M3 Max tests, the checkpoint-and-recompute kernels reduced peak recurrent-state memory by approximately 12–18× at the kernel level and lowered total training peak memory from 23.88 GB to 10.34 GB. At the same batch size, end-to-end training throughput improved by roughly 1.4×, while individual fused forward-and-backward kernels ran approximately 1.5–1.9× faster than the original full-state implementations. Results will vary with recurrence type, sequence length, state dimensions, batch size, datatype, model architecture, and hardware. Current kernels: ssd_scan Mamba-2-style, head-wise SSD selective scan. Intended for Mamba-2 and other SSM hybrid architectures. State shape: [B, H, Dh, N] gla_scan Gated Linear Attention with a scalar forget gate and outer-product write. Intended for GLA and linear-attention hybrid architectures. State shape: [B, H, Dh, Dh] rglru_scan RG-LRU diagonal recurrence. Intended for Griffin and RecurrentGemma-style architectures. State shape: [B, D] rotlru_scan Rotational LRU using a complex-diagonal recurrence, a magnitude gate, and a per-step rotation of two-dimensional channel pairs. Intended for complex-LRU and S4-style oscillatory memory architectures. State shape: [B, D], represented as interleaved channel pairs. Each kernel is implemented as a self-contained plug-in on a shared chassis located at: mlx_recurrence._chassis The chassis provides: Segment checkpoint-and-recompute infrastructure Shape and argument validation VJP integration Forward and gradient parity-test helpers Common recurrence plug-in handling Adding another recurrence therefore requires implementing its Metal forward and backward source pair and connecting its VJP. The checkpointing, validation, and testing infrastructure does not need to be rebuilt for each operator. The original version 0.1 kernels remain available under: mlx_recurrence.legacy They are also re-exported at the package’s top level for backward compatibility. I’m interested in feedback from developers working with MLX or custom Metal compute kernels, particularly around: Preferred APIs for packaging reusable MLX Metal extensions Threadgroup and memory-layout strategies across Apple GPU generations Numerical stability expectations for long recurrent sequences Benchmarking fused scans against MLX-native implementations Additional recurrent operators that would be valuable to support I would also be interested to know whether others are developing similar fused recurrence primitives for MLX and whether a shared interface for these operations would be useful. My setup is a M3 MAX Macbook Pro with 36GB Ram and I am running on macOS 26.4.1 (25E253). https://github.com/D-CSIL/mlx-recurrence

I have a Core AI model export — a bundle folder (.llmasset, containing multiple .aimodel subfolders plus metadata/tokenizer resources) — added to my app target as a folder reference. Rather than treating the bundle as one opaque resource and copying it into the app bundle as-is (the way .xcassets, .bundle, or any other folder reference behaves), Xcode reaches into it, finds the individual .aimodel subfolders, and adds each one to Compile Sources. When it compiles them there, it's for my build machine's specific chip only — I can't find any setting (Build Settings, scheme, target picker) to compile for multiple architectures/platforms, the way a universal binary would work. Question: Is there a way to make Xcode treat a .llmasset bundle as an atomic resource — copied wholesale, not decomposed into individual .aimodel compile targets? Or is reaching into the bundle and AOT-compiling its components for the active build architecture the intended behavior here, and if so, what's the recommended way to make sure the result works across the actual range of devices the app ships to?

Anyone able to get some neat SiriAI experiences working? Anything that makes you think “man I hope other developers do this in their apps too!”? (I’m willing!)

so i just had 2 convo closed because somehow the message i made was unrelated

Let's say I'm in Notes, or in Pages, or in Files or Live Text if I'm in Photos... will Siri AI be able to COPY the text for me if I ask it to do so? If you have new Siri AI installed, or are DTS Engineer at Apple I'd appreciate a yes/no? Critical for text editing, working with AI output, and general modern work requirements. Presently I have to press the "share" button then select Copy from the Share Sheet, OR in Pages I have to select EXPORT from the More Menu and choose Plain Text to get the contents. Thank you. Be well.

I'm trying to get some clarity on how the new Siri deals with IndexedEntities and whether it's worth adopting, considering our app does not fit into any of the predefined domain schemas. In running some tests with the TravelTracking sample app, it seems the only way I can get Siri to show any of the referenced entities is by using the exact phrasing (or extremely close to it) in one of the donated shortcuts. If I ask Siri to "Find closest landmark in TravelTracking" produces a result from the App in the form of an app snippet. But, if I then ask it "Text the description to Jane", it seeds the text with something like, "Niagara Falls is located in North America", instead of what's in the description field of the entity. General questions about the indexed data fail to show any results at all in Siri. For example: "Show me some landmarks from TravelTracking" or "Find Mount Fuji in TravelTracking" produce no results, even though the landmarks are indexed. My original assumption was that indexing data from your app would make it available to Siri, but it only seems to show up in on-device search and not in conversation with Siri itself. So is it the case that such data is only available through a Siri conversation if either you can adopt a domain schema or create a shortcut and use very close to the exact phraseology? And in the case of the latter, you can't really act on the returned entities because basically all you get is what is shown in a snippet? Maybe the on-screen intelligence picks up something here (seems to), but nothing deeper, even if it is defined in the entity. I've put in a feedback request (FB23796681) for a general database domain with schema for common database operations. Perhaps something like this and way to describe record types to aid in understanding from the LLM would go a long way toward making Siri more flexible for agentic use? I can get Siri to do a lot of the things that were shown at WWDC, but that tends to make you think you can do similar things with other types of apps and when you can't because of the domain limitations, it's very frustrating and feels limiting. It seems the domain types fit the apps Apple ships with the OS (Mail, Photos, Notes, etc), but not other types of apps that don't fit that criteria. If I'm missing something here, any guidance would be appreciated.

still no sign of the new siri no app no nothing im on the ios 27 beta 2 and iphone 15 pro max what is this apple

We recently had to change our MLModel's architecture to include custom layers, which means the model can't run on the Neural Engine anymore. After the change, we observed a lot of crashes being reported on A13 devices. It turns out that the memory consumption when running the prediction with the new model on the GPU is much higher than before, when it was running on the Neural Engine. Before, the peak memory load was ~350 MB, now it spikes over 2 GB, leading to a crash most of the time. This only seems to happen on the A13. When forcing the model to only run on the CPU, the memory consumption is still high, but the same as running the old model on the CPU (~750 MB peak). All tested on iOS 16.1.2. We profiled the process in Instruments and found that there are a lot of memory buffers allocated by Core ML that are not freed after the prediction. The allocation stack trace for those buffers is the following: We ran the same model on a different device and found the same buffers in Instruments, but there they are only 4 KB in size. It seems, Core ML is somehow massively over-allocating memory when run on the A13 GPU. So far we limit the model to only run on CPU for those devices, but this is far from ideal. Is there any other model setting or workaround that we can use to avoid this issue?

We're implementing the App Intents Phone schema domain in our app to enable Siri to initiate calls to our contact entities via our voip. We've implemented a .phone.startCall intent and registered our entities as .phone.phonePerson. The intent provides both the required destination and audioVisualMode parameters, and the perform() method is implemented to handle the call. However, the perform() method is never invoked. Instead, Siri either: Says that the phone number is not linked, or Announces that it is calling, but our app intent is never executed. Anybody implemented this Phone schema domain and it s working successfully ? Sample Code: struct StartCallIntent: AudioRecordingIntent, AudioPlaybackIntent { var destination: CallDestination var audioVisualMode: CallAVMode init(contact: ContactEntity, mode: CallAVMode = .audio) { self.destination = .phonePerson(contact) self.audioVisualMode = mode } func perform() async throws -> some IntentResult { print("Call Initiating to contact") return .result() } @AppEnum(schema: .phone.audioVisualMode) enum CallAVMode: String, CaseIterable { case audio case video } @UnionValue enum CallDestination: Sendable { case phonePerson(ContactEntity) case group([ContactEntity]) } @AppEntity(schema: .phone.phonePerson) struct ContactEntity: IndexedEntity { static var defaultQuery = ContactEntityQuery() let id: UUID var person: IntentPerson }

Can anyone please tell if a M5 Pro Macbook Pro can connect to a M3 ultra Mac studio over thunderbolt 5 using RDMA for LLM inference? Thanks

Hi everyone, I’m testing the latest iOS 27 beta and I’ve noticed an issue with the new Siri. When I ask very simple questions that should be handled locally or through basic reasoning, Siri consistently responds with: “Uh oh, something went wrong.” For example, asking: “When is the next Friday the 13th?” results in the error message instead of an answer. I’ve reproduced this multiple times and it seems to happen with other straightforward informational queries as well. I’ve already tried restarting the device and checking my network connection, but the issue persists. Has anyone else experienced this behavior with the new Siri in the iOS 27 beta? If so, were you able to find a workaround or identify what’s causing it? Any help or confirmation would be greatly appreciated. Thanks!

We have an app that supports iOS 18+ We have a couple of AppEntity(s) that we are keen to make work with the new schemas along with several AppIntent(s). We cannot increase our floor to iOS 27 for obvious reasons. All the documentation suggests using the macros, e.g. @AppEntity(schema: .audio.song) struct SongEntity { ... } This refuses to compile below iOS 26. It's possible to add availability checks, e.g. @available(anyAppleOS 27, *) @AppEntity(schema: .audio.song) struct SongEntity { ... } But then the whole entity becomes unavailable on pre-27 OSes. So I tried moving the macro onto an extension, e.g. struct SongEntity { ... } @available(anyAppleOS 27, *) @AppEntity(schema: .audio.song) extension SongEntity { ... } But this results in a compiler error: 'extension' macro cannot be attached to extension (extension of 'SongEntity') One other option is to create a new entity with a totally different name and mark it as isAssistantOnly but this has a lot of quite negative downstream effects that make it unworkable. For example: a lot of code duplication duplication in search indexes if we index both sets of entities awkwardness trying to use NSUserActivity when we have 2 different entity types pain in downstream AppIntent arguments which would require duplicating every AppIntent which has more cascading effects The same issues are present in AppIntent schemas too where even trying to add the most basic @AppIntent(schema: .system.open) to our existing OpenIntent doesn't seem possible for all the same reasons. I am really struggling with how to structure code so we can support schemas, currently I don't really see a path forward here until our floor raises to iOS 27. Is there a way to make this work nicely with the current APIs? What are others doing here? How can apps can ship in September and support both this and pre iOS 27 cleanly? Thinking about solutions here, my ideal would be that the macros are improved to either: be able to be applied to an extension rather than the structure itself. expand in such a way that they still build the core AppEntity / AppIntent on pre 27 OSes but then add the iOS 27 schema additions behind @available internally so they can be used with older targets as essentially no-ops on the current definitions.

We have a list app that implements INAddTasksIntent so users can add items to our app with Siri. We're now working on implementing an App Intent for the .reminders.createReminder schema for iOS 27. Our app still supports iOS 18, so it implements both INAddTasksIntent and the .reminders.createReminder schema. Observed behavior (iOS 27 beta 4): When we say "Siri, add eggs to my grocery list in AppName", Siri routes the request to the legacy INAddTasksIntent handler in our SiriKit extension. Our new CreateReminderIntent is never invoked. I confirmed this with breakpoints and logging in both handlers. The CreateReminderIntent does seem to be set up correctly, because it appears in the Shortcuts app and I can invoke it via AppIntentsTesting. Also, after using the above phrase, I was able to say "Siri, add cookies to my grocery list" and the item got added to my app via the INAddTasksIntent, even though I didn't specify the app name in the request. This also worked with a version of our app that does not contain CreateReminderIntent running on iOS 26.5. Isn't the app name normally required for INAddTasksIntent to be invoked? Questions: Is Siri activating the INAddTasksIntent instead of the new CreateReminderIntent expected behavior? Are users on iOS 27 going to have a worse experience adding items to our app with Siri if we support both INAddTasksIntent and the new CreateReminderIntent? If so, how do you recommend we proceed? Thank you for any guidance you can provide.

I wanted to activate the new Siri AI on my iPhone 12 running iOS 27 Developer Beta 4. I knew it didn't support Apple Intelligence features, but I switched the language to English anyway—and then I ran into this bug: I went into the Siri tab and saw this.