Siri AI Opens to All Developers: App Intents Bridges the App Ecosystem
Apple opens Siri AI to all developers in iOS 27.3 betas, letting apps be invoked by voice via App Intents.
Apple has launched a Developer Preview in the iOS 27.3, iPadOS 27.3, macOS 27.3, and visionOS 27.3 betas, opening Siri AI access to all developers. Any app can integrate its content and actions into the system assistant through App Intents, gaining personal context understanding, screen awareness, and natural language processing. The consumer version is expected in a few months, but the feature is not yet available in the EU and has daily usage limits for server-side models.
Apple has officially opened Siri AI access to all developers in the beta versions of iOS 27.3, iPadOS 27.3, macOS 27.3, and visionOS 27.3. This means any app can now integrate its content and actions directly into Siri AI through App Intents, marking a structural shift in how developers reach users.
From Closed Integration to Full Openness
Previously, third-party Siri integrations were concentrated in a handful of scenarios—navigation, to-dos, messaging, and audiobooks—involving thousands of developers. The Developer Preview released this time removes the barrier: any app can invoke Siri AI's new capabilities, including personal context understanding, on-screen content awareness, and natural language processing. When users make requests in everyday language, Siri AI can execute custom actions built by developers—checking package deliveries, booking restaurants, placing orders.
For complex needs, Apple has also introduced a skills mechanism that allows developers to teach Siri AI to combine multiple actions into a single complete experience. Another easily overlooked change is that Siri AI can now present any entity from the Spotlight semantic index, no longer limited to types conforming to existing schemas. This effectively exposes an app's internal data structures directly to the system-level assistant.
App Intents is the app action declaration framework Apple introduced starting with iOS 16, replacing the earlier SiriKit intent system. Developers declaratively define "actions" in Swift (such as querying orders or starting playback), along with their parameters, return values, and entity types; the system then automatically generates callable capabilities across Siri, Shortcuts, Spotlight, and other entry points. Compared with SiriKit, App Intents is no longer confined to fixed domains like navigation and messaging—it allows any custom domain and enables bidirectional integration with an app's data model. The key to this opening is that Siri AI can read these declarations and execute them with semantic understanding, so apps can be driven by natural language without building a separate conversational interface. The trade-off is that action descriptions must be precise, or intent recognition will easily fail.
Spotlight Semantic Index refers to apps registering their internal entities into the system index in a structured way, for Spotlight search and system-level assistant invocation. Previously, the system recognized only a few types conforming to established schemas (such as contacts and media items), and custom app data structures were hard for the system to understand. With this opening, Siri AI can present any entity in the index—effectively letting the system assistant directly "see" business objects inside apps, such as a specific order, project, or ticket. This provides the data foundation for cross-app natural language queries, but it also means developers must proactively design which fields are exposed in the index and at what granularity, and assess privacy boundaries.
Apple's Own Apps Set the Example
Apple has already rolled out the new capabilities across a batch of its own apps, including Calendar, Files, Finder, Journal, Mail, Maps, Messages, Music, Notes, Photos, Preview, Reminders, Safari, Shortcuts, and Weather. Users can ask Siri AI to find the coffee shop they visited last week, or archive files they haven't opened in over a year. These actions also provide reference examples for third-party developers.
From a product perspective, this is a two-way ecosystem expansion: users take fewer steps, and apps gain another entry point. For utility, lifestyle service, and content apps, the value of a voice entry point lies in shortening the path from intent to execution—not in building yet another chat interface.
Phased Rollout and Regional Restrictions
This release is a Developer Preview, with early developer access preceding the consumer launch, which is expected in a few months; the App Store submission channel will open in the spring. Apple also said it will later publish terms and usage policy details for experiences delivered through Siri AI.
On the language front, Siri AI currently supports English, French, Japanese, Korean, Portuguese, and Spanish, and will expand to Traditional Chinese, Norwegian, Turkish, and Vietnamese. Device requirements are Apple Intelligence–capable models with the system language in the supported list.
Regional restrictions deserve separate mention: Siri AI is currently unavailable on iOS, iPadOS, and watchOS in the EU, and Apple says it is in talks with EU authorities to find a solution. Additionally, the feature is not available to users under 13.
Apple Intelligence is Apple's generative AI system introduced in 2024 that combines on-device and cloud processing; it requires specific chips (A17 Pro / M series and newer) and sufficient memory, so only newer devices are supported. The system language must be in the supported list because language models and speech recognition resources are trained and distributed per language. AFM 3 Cloud refers to the cloud version of Apple Foundation Models: lightweight requests are handled on-device, while complex requests go to Apple's Private Cloud Compute, and are therefore subject to quotas and billing policies. This on-device/cloud split means developers must distinguish which Siri AI capabilities can run reliably offline and which are affected by network, quotas, and policy fluctuations.
Cost Items Developers Need to Assess in Advance
One easily overlooked detail is the usage limit on server-side models. Apple Intelligence features that rely on server-side models are subject to daily usage caps—Siri AI, intelligent photo editing, Image Playground, and the AFM 3 Cloud model in Shortcuts are all included. Limits will fluctuate with features, request complexity, system load, and policy changes, and expanded access will be charged in the future.
For developers, this directly affects feature design: high-frequency, heavy Siri AI call scenarios need fallback paths or should move some logic on-device. The early Preview phase is also a window to observe real call behavior and assess cost structures. Apple will host Apple Intelligence Group Labs on November 4–5, where developers can ask Apple engineers questions and get advice.
What It Means for the App Ecosystem
Siri AI's opening turns the system-level assistant into a programmable distribution layer. Apps are no longer found only through icons and search—they can also be invoked through a user's natural language request. For developers, early integration means polishing the experience before the official launch and building an understanding of intent recognition; for Apple, this is an important step in using the developer ecosystem to fill in Apple Intelligence's app breadth.
Risks exist as well. The degree of semantic index openness, the compositional complexity of skills, the experience gaps caused by daily limits, and the absence from the EU market will all affect the actual reach of this mechanism. The signal from the Preview phase is clear: the door is open, but the rules are still taking shape.
One-line Summary
Apple opens Siri AI access to all developers in iOS 27.3 and other betas, letting apps be invoked by voice via App Intents.
Paragraph Summary
Apple has launched a Developer Preview in the iOS 27.3, iPadOS 27.3, macOS 27.3, and visionOS 27.3 betas, opening Siri AI access to all developers. Any app can integrate its content and actions into the system assistant through App Intents, gaining personal context understanding, screen awareness, and natural language processing. A new skills mechanism combines multiple actions into a single complete experience, and Siri AI can present any entity from the Spotlight semantic index. Fifteen Apple apps—including Calendar, Mail, and Maps—have already integrated as examples. The feature is currently a developer preview, with a consumer version expected in a few months and App Store submissions opening in the spring; it is not yet available on iOS, iPadOS, and watchOS in the EU, nor to users under 13. Features relying on server-side models are subject to daily limits, and expanded access will be charged in the future.
Related articles
How Accurate Are Sci-Fi's Predictions About AI?
How accurate are sci-fi's predictions about AI? Drawing on a Hacker News discussion, this article analyzes sci-fi's foresight and blind spots in AI development, arguing its true value lies in inspiring thought rather than precise prediction.
SwarmEvolve: Latent-Space Agent Swarms Accelerate Evolutionary Search
SwarmEvolve lets hundreds of low-cost agents swarm-reason in latent space, sharing info via KV caches to achieve fast evolutionary search for math and science problems—reaching in a median of 3–8 minutes the best scores other frameworks need 60 minutes for.
trainlint: Static Checks Catch ML Training Bugs That Don't Crash
trainlint is a static analysis tool that finds ML training bugs that don't throw errors, including data leakage, leftover eval mode, and softmax/CrossEntropyLoss conflicts across 14 categories. This article covers its principles, validation results, and limitations.