App Intents logo

App Intents

Community
rshankras
app-intents

App Intents for Siri, Shortcuts, Spotlight, and Apple Intelligence integration including intent modes, interactive snippets, visual intelligence, and entity indexing. Use when implementing Siri integration, App Shortcuts, or Spotlight indexing.

Overview

Publisherrshankras
Repositoryclaude-code-apple-skills
Skill nameapp-intents
Stars
744
Forks
70
Bundled files
3
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by rshankras on GitHub. Read the source before you install it.

Installation

Install the App Intents AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/rshankras/claude-code-apple-skills.git /tmp/claude-code-apple-skills
mkdir -p .claude/skills
cp -r /tmp/claude-code-apple-skills/skills/apple-intelligence/app-intents .claude/skills/app-intents
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable App Intents in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads App Intents on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so App Intents is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

App Intents

Build intents that expose your app's functionality to Siri, Shortcuts, Spotlight, and Apple Intelligence. Covers the full App Intents framework from basic actions through advanced features like interactive snippets, intent modes, visual intelligence integration, and Spotlight entity indexing.

When This Skill Activates

  • User wants to add Siri or Shortcuts integration
  • User asks about App Intents, AppIntent, or AppEntity
  • User needs Spotlight indexing for app content
  • User wants to create App Shortcuts with voice phrases
  • User is implementing interactive snippets for Siri results
  • User asks about intent modes (foreground, background)
  • User needs visual intelligence integration via App Intents
  • User wants onscreen entity support for Siri/ChatGPT
  • User asks about Swift package support for App Intents

Decision Tree

What do you need?
|
+-- Expose an action to Siri/Shortcuts
|   +-- Simple action, no UI needed
|   |   --> Basic AppIntent (intents-basics.md)
|   +-- Needs to show UI or ask user questions
|   |   --> Intent Modes + Interactive Snippets (advanced-features.md)
|   +-- Needs a predictable voice phrase
|       --> App Shortcuts (intents-basics.md)
|
+-- Make content searchable
|   +-- In Spotlight
|   |   --> IndexedEntity + @Property (entities-spotlight.md)
|   +-- Runnable from Spotlight on Mac
|   |   --> parameterSummary visibility gates (entities-spotlight.md)
|   +-- In Visual Intelligence
|   |   --> IntentValueQuery + SemanticContentDescriptor (advanced-features.md)
|   +-- As onscreen entities for Siri/ChatGPT
|       --> annotation APIs + EntityIdentifier (advanced-features.md)
|
+-- Let Siri execute intents from natural language
|   --> App Schemas: @AppIntent(schema:) / @AssistantIntent (advanced-features.md)
|
+-- Feed entities to Apple Intelligence (Use Model action)
|   --> AttributedString params + entity JSON + Find actions (entities-spotlight.md)
|
+-- Hand entities to other apps as content/files
|   --> Transferable / FileEntity (entities-spotlight.md)
|
+-- Show rich results in Siri
|   +-- Static display only
|   |   --> .result(view:) snippet (advanced-features.md)
|   +-- Interactive buttons/controls
|   |   --> SnippetIntent protocol (advanced-features.md)
|   +-- Custom spoken dialog
|       --> IntentDialog(full:supporting:) (advanced-features.md)
|
+-- Present choices to the user
|   --> requestChoice(between:) (advanced-features.md)
|
+-- Teach Siri from in-app UI actions
|   --> IntentDonationManager (advanced-features.md)
|
+-- Share intents via Swift Package
    --> AppIntentsPackage protocol (advanced-features.md)

API Availability

FeatureMinimum OSFramework
AppIntent protocoliOS 16 / macOS 13AppIntents
AppEntity protocoliOS 16 / macOS 13AppIntents
AppShortcutsProvideriOS 16 / macOS 13AppIntents
@Parameter macroiOS 16 / macOS 13AppIntents
IndexedEntity protocoliOS 18 / macOS 15AppIntents
@Property with indexingKeyiOS 18 / macOS 15AppIntents
Intent Modes (supportedModes)iOS 26 / macOS 26AppIntents
requestChoice(between:)iOS 26 / macOS 26AppIntents
@ComputedPropertyiOS 26 / macOS 26AppIntents
@DeferredPropertyiOS 26 / macOS 26AppIntents
SnippetIntent protocoliOS 26 / macOS 26AppIntents
AppIntentsPackage protocoliOS 26 / macOS 26AppIntents
Onscreen entities (.userActivity())iOS 26 / macOS 26AppIntents
@UnionValueiOS 18 / macOS 15AppIntents
Assistant Schemas (@AssistantIntent)iOS 18AppIntents
Transferable entities, FileEntityiOS 18 / macOS 15AppIntents
UndoableIntentiOS 26 / macOS 26AppIntents
App Schemas on @AppIntent(schema:)iOS 27 / macOS 27AppIntents
IntentDonationManager, OwnershipProvidingEntityiOS 27 / macOS 27AppIntents
AppIntentsTesting frameworkXcode 26 cycle (WWDC26)AppIntentsTesting

Quick Reference

TaskType/APIReference File
Define an actionAppIntent protocolintents-basics.md
Accept parameters@Parameter macrointents-basics.md
Create voice phrasesAppShortcutsProviderintents-basics.md
Define a data entityAppEntity protocolentities-spotlight.md
Index in SpotlightIndexedEntity protocolentities-spotlight.md
Mark indexable fields@Property(indexingKey:)entities-spotlight.md
Run in background/foregroundsupportedModesadvanced-features.md
Continue in foregroundcontinueInForeground()advanced-features.md
Show result UI.result(view:)advanced-features.md
Interactive result UISnippetIntent protocoladvanced-features.md
Present choicesrequestChoice(between:)advanced-features.md
Visual intelligence searchIntentValueQueryadvanced-features.md
Onscreen entity association.userActivity() modifieradvanced-features.md
Computed/deferred properties@ComputedProperty, @DeferredPropertyadvanced-features.md
Share via packagesAppIntentsPackageadvanced-features.md
Make intents Siri-executable@AppIntent(schema:), @AssistantIntentadvanced-features.md
Update-intent "clear vs leave unchanged"valueState (.set/.set(nil)/.unset)advanced-features.md
Custom Siri dialogIntentDialog(full:supporting:)advanced-features.md
Donate in-app UI actionsIntentDonationManageradvanced-features.md
Shared-content confirmationsOwnershipProvidingEntityadvanced-features.md
Undo intent actionsUndoableIntentadvanced-features.md
Export entities as content/filesTransferable, FileEntityentities-spotlight.md
Run from Spotlight on MacparameterSummary gatesentities-spotlight.md
Accept model-generated rich textAttributedString parametersentities-spotlight.md

Process

1. Identify Integration Needs

Read the user's code or requirements to determine:

  • What actions should be exposed to Siri/Shortcuts
  • What content should be searchable in Spotlight
  • Whether interactive snippets are needed for Siri results
  • Whether the intent needs foreground UI or can run in background
  • Target platform and minimum OS version

2. Load Relevant Reference Files

Based on the need, read from this directory:

  • intents-basics.md -- AppIntent protocol, @Parameter, perform(), App Shortcuts
  • entities-spotlight.md -- AppEntity, IndexedEntity, Spotlight indexing, @Property
  • advanced-features.md -- Intent modes, interactive snippets, visual intelligence, onscreen entities, choices, packages

3. Review or Implement

Apply patterns from the reference files. Check for common mistakes (see Top Mistakes below).

4. Cross-Reference

  • For Visual Intelligence camera search, see apple-intelligence/visual-intelligence/
  • For Foundation Models on-device LLM, see apple-intelligence/foundation-models/
  • For deep linking from intents, see generators/deep-linking/ skill

Top Mistakes

These are the most frequent errors when implementing App Intents.

1. Missing static metadata

swift
// ❌ Wrong -- no title or description
struct MyIntent: AppIntent {
    func perform() async throws -> some IntentResult {
        return .result()
    }
}

// ✅ Correct -- static title is required
struct MyIntent: AppIntent {
    static var title: LocalizedStringResource = "Do Something"
    static var description: IntentDescription = "Performs the action"

    func perform() async throws -> some IntentResult {
        return .result()
    }
}

2. Forgetting to index entities after changes

swift
// ❌ Wrong -- entities updated but Spotlight not notified
func saveRecipe(_ recipe: Recipe) {
    database.save(recipe)
}

// ✅ Correct -- reindex after mutations
func saveRecipe(_ recipe: Recipe) async throws {
    database.save(recipe)
    try await CSSearchableIndex.default().indexAppEntities()
}

3. Using foreground intent for background-safe work

swift
// ❌ Wrong -- forces app to foreground for a simple toggle
struct ToggleFavoriteIntent: AppIntent {
    static var title: LocalizedStringResource = "Toggle Favorite"
    static var openAppWhenRun = true  // Unnecessary

    func perform() async throws -> some IntentResult {
        toggleFavorite()
        return .result()
    }
}

// ✅ Correct -- runs silently in background
struct ToggleFavoriteIntent: AppIntent {
    static var title: LocalizedStringResource = "Toggle Favorite"

    static let supportedModes: IntentModes = .background

    func perform() async throws -> some IntentResult {
        toggleFavorite()
        return .result()
    }
}

4. Not providing EntityStringQuery for entities

swift
// ❌ Wrong -- entity has no way to be queried
struct NoteEntity: AppEntity {
    var id: String
    var title: String
    // Missing: static var defaultQuery
}

// ✅ Correct -- provides a query so Siri can resolve entities
struct NoteEntity: AppEntity {
    var id: String
    var title: String

    static var defaultQuery = NoteEntityQuery()
    // ... typeDisplayRepresentation, displayRepresentation
}

5. Returning too many Spotlight results

swift
// ❌ Wrong -- indexing thousands of items at once blocks the main thread
func indexAll() async throws {
    let allItems = database.fetchAll()  // 50,000 items
    try await CSSearchableIndex.default().indexAppEntities()
}

// ✅ Correct -- batch index and run off main thread
func indexAll() async throws {
    try await CSSearchableIndex.default().indexAppEntities(
        of: RecipeEntity.self
    )
}

Design Guidelines

How to decide what to expose and how it should behave — from Apple's design sessions.

App Shortcuts (WWDC22)

  • Pick self-contained, straightforward features completable without the app in focus.
  • Hard cap is 10 App Shortcuts; aim for 2–5 high-quality ones.
  • Invocation phrase: brief, memorable, and must include the app name — provide natural synonym variants per language ("Start a run" / "Begin a run").
  • At most one dynamic parameter per phrase; values must come from a finite, front-of-mind list, ordered by recency/frequency — the first value becomes the top Spotlight suggestion.
  • Three dialog flows: Parameter Confirmation (assume the likely value, confirm it), Disambiguation (short list that teaches the available values), Intent Confirmation only for consequential actions (financial, destructive, high-risk).
  • Snippet visuals: semitranslucent material + vibrant label colors — never opaque backgrounds.
  • Suppress spoken dialog when the snippet fully communicates the result, but keep the dialog complete for voice-only contexts (AirPods, CarPlay).
  • Surface in-app education right before or after the user performs the action they'd repeat — that's when the phrase sticks.

What deserves to be an intent (WWDC24)

  • "Anything your app does should be an app intent." Scope by task, not by a feature checklist.
  • Start from fundamental verbs — Create, Open, Search — then specialize.
  • Consolidate near-duplicates into one flexible intent with parameters ("Start Workout" with a workout-type parameter, not five separate intents).
  • Intents represent tasks, never UI gestures — "save the draft", not "tap the save button".
  • Live Activity and audio apps should expose background read-intents (current state, now playing), not just actions.
  • Parameter summaries must read as complete sentences for every possible value.
  • Make parameters optional by default so the intent runs immediately with no follow-up questions; binary states default to a toggle rather than explicit on/off.
  • Intents that end in the UI get an "Open When Run" toggle, default on — users composing shortcuts may want to suppress it.

Interactive snippet design (WWDC25)

  • Max content height 340pt — beyond that is clipped, not scrolled.
  • Use larger-than-default type; snippets are glanced at from a distance.
  • Show only essential info — link into the app for anything more.
  • Use ContainerRelativeShape for margins that adapt to the snippet's corner radius.
  • Vibrant brand backgrounds are fine, but check contrast for distance viewing.
  • Two types: Result (outcome + Done button) vs Confirmation (action-verb button like "Order"); Confirmation → Result is the canonical flow.
  • The snippet must be self-sufficient even if Siri dialog is never shown or heard.

Review Checklist

Before shipping App Intents integration:

  • Every AppIntent has a static var title and static var description
  • Every AppEntity has typeDisplayRepresentation, displayRepresentation, and defaultQuery
  • @Parameter properties have descriptive titles
  • Entities used in Shortcuts have EntityStringQuery or EntityPropertyQuery
  • IndexedEntity types call CSSearchableIndex.default().indexAppEntities() after data changes
  • @Property fields used in indexing have indexingKey set
  • App Shortcuts have clear, natural-language phrases with \(.applicationName)
  • Intent modes match the actual work: background for data ops, foreground for UI
  • Interactive snippets use SnippetIntent (not plain AppIntent)
  • SnippetIntent perform() never mutates state — mutations live in the button intents (WWDC25 275)
  • parameterSummary includes every required parameter without a default — the Spotlight-on-Mac visibility gate (WWDC25 260)
  • Text-accepting parameters use AttributedString, not String (Use Model rich text, WWDC25 260)
  • perform() handles errors gracefully and returns meaningful dialog
  • Intents are tested in the order: AppIntentsTesting → Shortcuts → Spotlight → Siri (WWDC26 240)
  • Deep links from entity results navigate to the correct screen

WWDC Session References

Rules throughout the reference files carry inline attributions to these sessions:

SessionCovers
WWDC24 10133 — Bring your app to SiriAssistant Schemas, 12 iOS 18 domains, semantic search
WWDC24 10210 — Bring your app's core features to usersCore doctrine: intents, entities, queries, reuse across surfaces
WWDC24 10134 — App Intents framework additionsIndexedEntity, Transferable, FileEntity, @UnionValue
WWDC24 10176 — Design App Intents for system experiencesScope + parameter design rules, Open When Run
WWDC25 244 — Get to know App IntentsProtocol shapes, metadata extraction, ID contract, packaging
WWDC25 275 — Advances in App IntentsSnippetIntent, intent modes, undo, onscreen entities
WWDC25 260 — Develop for Shortcuts and SpotlightUse Model action, Find actions, Mac Spotlight gates
WWDC26 240 — Build intelligent Siri experiences with App SchemasUnified schema macros, testing ladder
WWDC26 343 — Advanced App Intents features for SiriDialogs, donations, ownership, annotation APIs
WWDC26 344 — Code-along: Make your app available to SiriCanonical integration sequence, valueState

References

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the App Intents AI skill do?

App Intents for Siri, Shortcuts, Spotlight, and Apple Intelligence integration including intent modes, interactive snippets, visual intelligence, and entity indexing. Use when implementing Siri integration, App Shortcuts, or Spotlight indexing.

Why use App Intents on TypingMind?

Because you install it once and use it with any model. App Intents is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install App Intents in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/rshankras/claude-code-apple-skills/tree/main/skills/apple-intelligence/app-intents. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use App Intents?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with App Intents?

As many as you like. As long as a model supports skills, you can use App Intents with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the App Intents AI skill free?

Yes. It is published on GitHub by rshankras under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇