Contentful Personalization & Analytics
    Preparing search index...

    Integrating the Optimization iOS SDK in a UIKit app

    Use this guide to add Contentful personalization to a UIKit app with the ContentfulOptimization Swift Package. By the end of the quick start, the SDK is running in your scene and one screen event has passed the SDK's consent gate, with a visible label confirming it.

    New to personalization? Here is the whole idea in five points:

    • In Contentful you author variants of an entry and attach them to an experience — a rule that decides which visitors see which variant.
    • As the app runs, Contentful's Experience API looks at who the visitor is and picks the variant for each experience. Swapping a fetched entry for its picked variant is called resolving the entry.
    • The Experience API also returns a profile: the anonymous, per-visitor identity and state used to keep personalization consistent across requests or app launches.
    • Your app hands a Contentful entry to the SDK at the point where that entry becomes output. The SDK gives back the selected variant, or the original entry when no variant applies—the baseline fallback.
    • You render the returned entry with the same application components you already use.

    The iOS SDK persists the profile in UserDefaults across app launches when persistence consent allows it.

    That is enough to start. The guide introduces policy and optional capabilities at the point you need them.

    You will get there in two milestones:

    • Milestone 1 — the SDK initialized in your scene and one accepted screen event (the quick start below). Once your app also hands the SDK a fetched Contentful entry, that entry resolves to a variant or the baseline through resolveOptimizedEntry (the Contentful fetching and entry resolution section). This is complete and shippable on its own.
    • Milestone 2 — the opt-in layers (later). Consent handoff, interaction tracking, identity, Custom Flags, live updates, the preview panel, runtime locale changes, and offline delivery, each introduced by the section that needs it. Start with Consent and privacy-policy handoff.

    This guide uses ContentfulOptimization. UIKit apps drive the SDK through the imperative OptimizationClient: you create and initialize one client, hold it for the scene or app lifetime, and inject it into the view controllers that track events or resolve entries. The SDK does not replace your app's Contentful client — your UIKit app still owns Contentful fetching, link resolution, consent UX, identity policy, navigation, caching, and rendering. If your app renders through SwiftUI views instead, use the Integrating the Optimization iOS SDK in a SwiftUI app guide. A UIKit app that hosts some screens in SwiftUI through UIHostingController can use that guide's SwiftUI view surface on those screens with this same client: those views read the client from the SwiftUI environment, so inject it with .environmentObject(client) on the hosted view rather than wrapping the screen in OptimizationRoot, which creates and initializes a client of its own.

    Most UIKit + Contentful apps share one shape: a SceneDelegate builds the window and a root view controller, and a UIViewController presents content. This quick start assumes that shape and proves the smallest result: the SDK initializes in your scene and one screen event is accepted, and a visible label flips to confirm it. It owns one OptimizationClient in the scene, initializes it, injects it into the first view controller, and tracks the current screen from viewDidAppear(_:).

    This quick start assumes your application policy permits Optimization to start with accepted consent and renders no end-user consent UI, so it configures StorageDefaults(consent: true) — the shorthand that accepts both consent axes at once. Read that as a startup default, not a one-time seed: a configured value takes precedence over whatever is stored in UserDefaults on every launch, so a shipped consent: true re-grants consent on each launch even after a user revoked it. If personalization must wait for a consent decision, keep this structure and add the Consent and privacy-policy handoff step before you ship, which explains the two axes, the split form that sets them separately, and why an app that collects a choice leaves defaults unset.

    1. Add the ContentfulOptimization Swift Package to your app target from https://github.com/contentful/optimization.swift (in Xcode: File > Add Package Dependencies), then build and run the app target once on a simulator or a device so Swift Package Manager resolves and compiles the package. The package supports iOS 15+.

    2. Own one client in your existing SceneDelegate, initialize it, and inject it into your first view controller. initialize(config:) is synchronous and throws — it loads the SDK's bridge and runs bridge setup inline on the main actor — so call it with try and no await. The bridge is the SDK's embedded JavaScript runtime: the iOS SDK runs the same Optimization core as the other SDKs in the suite inside a JavaScriptCore context, one per client. clientId is your Optimization client ID; Before you start says where to find it in the Contentful web app.

      Adapt this to your use case:

       import UIKit
      +import ContentfulOptimization
      
       final class SceneDelegate: UIResponder, UIWindowSceneDelegate {
           var window: UIWindow?
      
      +    // Own one client for the whole scene, then inject this same instance
      +    // into the view controllers that track events or resolve entries.
      +    let client = OptimizationClient()
      
           func scene(
               _ scene: UIScene,
               willConnectTo _: UISceneSession,
               options _: UIScene.ConnectionOptions
           ) {
               guard let windowScene = scene as? UIWindowScene else { return }
      
      +        // Synchronous throws, not async: call with try and no await. Catching
      +        // and printing keeps a failed startup visible instead of silent.
      +        do {
      +            try client.initialize(config: OptimizationConfig(
      +                clientId: "your-optimization-client-id",
      +                // Startup default, not a one-time seed: this wins over a stored choice.
      +                defaults: StorageDefaults(consent: true),
      +                logLevel: .debug,
      +                onEventBlocked: { blocked in
      +                    // If the label reads "blocked", this prints why.
      +                    print("Optimization blocked \(blocked.method): \(blocked.reason)")
      +                }
      +            ))
      +        } catch {
      +            print("Optimization initialize failed: \(error)")
      +        }
      
      -        let home = HomeViewController()
      +        let home = HomeViewController(client: client)
               window = UIWindow(windowScene: windowScene)
               window?.rootViewController = UINavigationController(rootViewController: home)
               window?.makeKeyAndVisible()
           }
       }
      

      The unchanged lines above are illustrative context to match against your own SceneDelegate, not a block to paste over it. StorageDefaults is the SDK config type that holds the SDK's startup state, consent included; StorageDefaults(consent: true) sets both consent axes — event consent and persistence consent — as the startup default the SDK resolves before it reads a stored choice.

    3. Track the current screen from a view controller and reflect the outcome in a label. HomeViewController below is illustrative app shape — adapt it to a screen you already render, keeping the client-injection initializer and the trackCurrentScreen call in viewDidAppear.

      Adapt this to your use case:

      import ContentfulOptimization
      import UIKit

      final class HomeViewController: UIViewController {
      private let client: OptimizationClient
      private let statusLabel = UILabel()

      init(client: OptimizationClient) {
      self.client = client
      super.init(nibName: nil, bundle: nil)
      }

      @available(*, unavailable)
      required init?(coder: NSCoder) { fatalError("init(coder:) is not supported") }

      override func viewDidLoad() {
      super.viewDidLoad()
      statusLabel.text = "Waiting for Optimization"
      statusLabel.textAlignment = .center
      statusLabel.translatesAutoresizingMaskIntoConstraints = false
      view.addSubview(statusLabel)
      NSLayoutConstraint.activate([
      statusLabel.centerXAnchor.constraint(equalTo: view.centerXAnchor),
      statusLabel.centerYAnchor.constraint(equalTo: view.centerYAnchor),
      ])
      }

      override func viewDidAppear(_ animated: Bool) {
      super.viewDidAppear(animated)
      Task { @MainActor in
      // Track the current screen once UIKit has made it visible.
      let result = try? await client.trackCurrentScreen(name: "Home")
      statusLabel.text = result?.accepted == true
      ? "Optimization screen event accepted"
      : "Optimization screen event blocked"
      }
      }
      }

      If your first screen comes from a storyboard, do not paste the init(coder:) above: UIKit instantiates storyboard controllers through exactly that initializer, so the fatalError crashes the app on launch. Keep UIKit's initializer and take the client through a settable property that the scene delegate fills in after the storyboard has built the controller — or read it from an app-level dependency container the controller can reach on its own.

      Adapt this to your use case:

      final class HomeViewController: UIViewController {
      // Reader-owned injection point instead of an initializer parameter, so
      // UIKit's init(coder:) stays intact. viewDidAppear is the first read.
      var client: OptimizationClient!

      // ...viewDidLoad and viewDidAppear exactly as above...
      }

      // In SceneDelegate.scene(_:willConnectTo:), after the storyboard built the window:
      if let home = (window?.rootViewController as? UINavigationController)?
      .viewControllers.first as? HomeViewController
      {
      // Inject the one scene-owned client, not a new one.
      home.client = client
      }
    4. Verify the first run. Launch the app on a simulator or a device; the label reads Optimization screen event accepted. trackCurrentScreen returns an EventEmissionResult — an SDK result type whose accepted flag is true when the event passed the SDK's local consent and allow-list gate and was emitted or queued for delivery. accepted does not confirm that Contentful received the event, only that the local gate let it through. Because StorageDefaults(consent: true) sets consent and screen is on the SDK's default pre-consent allow-list, the event is accepted.

    If the label reads Optimization screen event blocked, the two print calls from step 2 name the cause in the Xcode console. A line prefixed Optimization blocked means the consent gate rejected the event and names the reason and method. A line prefixed Optimization initialize failed means the client never initialized, so nothing reached the gate at all. Troubleshooting covers what each branch usually is.

    Table of Contents

    The sections below walk the integration in order. First, gather the few things you can only get from outside this guide:

    • A UIKit app and Xcode, with your own Contentful fetching already working and the ability to add a Swift package and run an Xcode build. The SDK is added through Swift Package Manager and supports iOS 15+.

    • Contentful delivery credentials — space ID, delivery token, environment, and one concrete locale — read from your app's configuration layer.

    • A configured contentful.swift client — Contentful's Swift delivery SDK, added to your app target and able to fetch one entry with a concrete locale. Every entry example below starts from a Contentful.Entry this client returned.

    • At least one entry with a variant attached to an experience, authored in Contentful. Without an authored variant, the integration can still run correctly while returning the baseline, so you cannot yet distinguish working personalization from a content-authoring gap. For the first personalized-content test, target all visitors so the test request or visitor matches automatically.

    • Your Optimization project values — client ID and environment, from your Optimization project settings. In the Contentful web app the path depends on which navigation your organization uses: in classic navigation, go to Apps → Installed apps → Contentful Personalization → SDK keys; in new navigation (the Contentful app with ExO navigation enabled), go to Platform/Apps → Installed apps → Contentful Personalization → SDK keys. The client ID and environment are listed there.

      OptimizationConfig.environment defaults to main, so pass it only when your setup differs. That default belongs to OptimizationConfig alone: the preview panel's own Contentful client (ContentfulHTTPPreviewClient, in Preview panel) defaults its environment to master, so pass yours explicitly there. The Experience API (which picks variants) and the Insights API (which receives event and interaction delivery) each have a base URL that defaults correctly; you only set them for mocks or non-default hosts (see Package installation and SDK configuration).

    You do not need a setup inventory up front. Everything else — consent, entry resolution, screen tracking, interaction tracking, identity, live updates, preview, runtime locale changes, offline delivery — is introduced by the section that needs it.

    Note

    Read the SDK client ID, Contentful credentials, and any base-URL overrides from your app's own configuration layer — an xcconfig value, a build setting, or a generated config type. This guide's examples use inline placeholder strings for clarity; the iOS reference app centralizes these in a shared AppConfig because it runs against shared mock defaults. Use whatever configuration convention your app already uses and keep it consistent.

    Integration category: Required for first integration

    The quick start added the package through Xcode. What is new here is the SwiftPM-manifest form of the same dependency, and the full OptimizationConfig surface behind the handful of keys the quick start passed.

    Adapt this to your use case:

    dependencies: [
    .package(url: "https://github.com/contentful/optimization.swift", from: "<version>"),
    ],
    targets: [
    .target(
    name: "MyApp",
    dependencies: [
    .product(name: "ContentfulOptimization", package: "optimization.swift"),
    ]
    ),
    ],

    Configure the SDK with your Optimization client ID and the environment that matches your Contentful setup. Only clientId is required by the initializer. An app builds one OptimizationConfig and calls initialize(config:) once, in the scene or app startup the quick start edited; the config snippets in later sections add keys to that same config rather than introducing a second one.

    1. Pass clientId from your configuration layer.
    2. Pass environment only when it is not the default main.
    3. Pass locale when Experience API requests and event context must use the same language as the Contentful Delivery API (CDA) entries you render.
    4. Set api base URLs (experienceBaseUrl/insightsBaseUrl) only for mock, staging, or other non-default endpoints — both default correctly otherwise.
    5. Keep logLevel at its default .error in production unless your operational policy allows more verbose logging.

    Adapt this to your use case:

    let appLocale = "en-US"

    let config = OptimizationConfig(
    clientId: "your-optimization-client-id",
    // environment defaults to "main"; pass it only when your setup differs.
    // Keep SDK event and Experience locale aligned with rendered CDA entries.
    locale: appLocale
    )

    For package-level installation notes, see the Optimization iOS SDK README.

    Integration category: Required for first integration

    UIKit integrations use OptimizationClient directly. Keep one initialized client alive for the scene or app lifetime, then inject that instance into every controller or view that resolves entries or tracks events.

    1. Create the client in SceneDelegate, AppDelegate, or an app-level dependency container, and call initialize(config:) before presenting content that uses Optimization.
    2. Pass the initialized client through initializers instead of creating separate clients in child controllers. For controllers UIKit instantiates itself — storyboard or nib scenes — use a settable property or an app-level dependency container the controller reads from, and keep UIKit's init(coder:) intact.
    3. Return to the main actor before calling the client from asynchronous callbacks; OptimizationClient is @MainActor.
    4. Gate UI on readiness when needed: the client publishes isInitialized, so observe client.$isInitialized when a screen must wait for setup before it reads SDK state.

    OptimizationClient is an ObservableObject, and its reactive values are Combine publishers reached through the $ prefix: client.isInitialized is the current Bool, while client.$isInitialized is the publisher that emits when it changes. Subscribing needs import Combine and somewhere to keep the subscription alive — a Set<AnyCancellable> property, conventionally called cancellables, that lives as long as the object doing the subscribing. Every Combine snippet later in this guide assumes a bag like the one below on whichever object subscribes.

    Adapt this to your use case:

    import Combine
    import ContentfulOptimization
    import UIKit

    final class ProductViewController: UIViewController {
    private let client: OptimizationClient
    // Keeps every Combine subscription in this controller alive.
    private var cancellables = Set<AnyCancellable>()

    // Inject the app-owned client instead of creating a new one here.
    init(client: OptimizationClient) {
    self.client = client
    super.init(nibName: nil, bundle: nil)
    }

    @available(*, unavailable)
    required init?(coder: NSCoder) { fatalError("init(coder:) is not supported") }

    override func viewDidLoad() {
    super.viewDidLoad()
    // Readiness gate: $isInitialized is the publisher behind isInitialized.
    client.$isInitialized
    .sink { [weak self] isReady in self?.setContentHidden(!isReady) }
    .store(in: &cancellables)
    }

    override func viewDidAppear(_ animated: Bool) {
    super.viewDidAppear(animated)
    Task { @MainActor in
    // Hop back to the main actor before any client call.
    _ = try? await client.trackCurrentScreen(name: "ProductList")
    }
    }
    }

    setContentHidden(_:) is reader-owned: your own method for showing a placeholder until the SDK is ready.

    Use destroy() only for test teardown or a deliberate SDK teardown flow, not for normal navigation between UIKit screens. For lifecycle and main-actor mechanics, see iOS SDK runtime and interaction mechanics.

    Integration category: Common but policy-dependent

    Consent policy belongs to your application. The SDK provides the runtime gate; your app or CMP owns notice, user choices, consent records, jurisdiction logic, and withdrawal behavior. Consent has two independent axes: event consent (may the SDK personalize and emit events) and persistence consent (may the SDK store profile continuity in UserDefaults).

    1. Use StorageDefaults(consent: true) at startup only when application policy permits SDK activity at launch, and only when no stored user choice can contradict it.
    2. Leave defaults unset when the app must collect a choice before gated events can emit, and call consent(...) from the app-owned banner, CMP callback, or settings flow.
    3. Use consent(_:) for the boolean shorthand that sets both axes, or consent(events:persistence:) to set them independently.
    4. Pass allowedEventTypes: [] for strict opt-in, so no SDK event emits before event consent.
    5. Observe client.$state when the UI must reflect event consent or persistence consent.

    StorageDefaults values are startup defaults, not one-time seeds. At every launch, initialize resolves the configured values over what is persisted in UserDefaults, so a configured consent or persistenceConsent replaces a stored user choice on that launch and every launch after it. An app that ships StorageDefaults(consent: true) therefore re-grants consent a user revoked, silently. That is the reason step 2 above matters: when your app collects a choice, leave defaults unset so the stored decision is what the SDK starts from, and let consent(...) carry your resolved policy.

    Adapt this to your use case:

    @objc private func acceptTapped() {
    // Boolean consent sets both event emission and durable profile continuity.
    client.consent(true)
    }

    @objc private func rejectTapped() {
    client.consent(false)
    }

    @objc private func allowEventsOnlyTapped() {
    // Split consent: emit events but keep profile continuity session-only.
    client.consent(events: true, persistence: false)
    }

    When allowedEventTypes is unset, the SDK's default pre-consent allow-list lets identify and screen emit before event consent, so a mobile journey can establish profile context and anonymous screen analytics. Before consent, that list is the whole admission rule: entry views, entry taps, custom track events, and page events (the page-view event the SDK shares with the web SDKs; UIKit apps track screens instead) are blocked because they are absent from it, not because consent is undecided. Accepting event consent admits every type at once. A custom allowedEventTypes replaces the default list, so a type you add there emits with no consent decision at all, and allowedEventTypes: [] blocks every SDK event until consent is accepted. consent(false) clears both axes, purges queues, and clears durable continuity while in-memory state stays usable until reset or teardown.

    Adapt this to your use case:

    // The same config the scene initializes with, plus one key.
    let config = OptimizationConfig(
    clientId: "your-optimization-client-id",
    // Replaces the default pre-consent allow-list of identify and screen with
    // strict opt-in: nothing emits until consent is accepted.
    allowedEventTypes: []
    )

    client.state is a snapshot value of type OptimizationState, and client.$state is the publisher behind it. Consent lives on that snapshot rather than on the client itself: read state.consent and state.persistenceConsent (each an optional Bool, where nil means the visitor has not decided yet), plus state.profile and state.changes — the flag and inline field values the Experience API returned for this visitor — when the UI reflects profile-driven values. The client's own published properties are a different set — isInitialized, selectedOptimizations, locale, isPreviewPanelOpen, previewState — so subscribe to client.$state for consent and to client.$selectedOptimizations for selections, not to one for both.

    Adapt this to your use case:

    client.$state
    // Consent is a field on the state snapshot, so this is the subscription a
    // consent banner or privacy screen observes.
    .sink { [weak self] state in
    self?.updateConsentUI(
    eventConsent: state.consent,
    persistenceConsent: state.persistenceConsent
    )
    }
    .store(in: &cancellables)

    updateConsentUI(eventConsent:persistenceConsent:) is reader-owned, and cancellables is the subscription bag from Client lifetime and UIKit injection.

    For the full consent responsibility model, see Consent management in the Optimization SDK Suite.

    Contentful fetching and entry resolution

    Integration category: Required for first integration

    This is where personalization happens. The iOS SDK has no native managed fetch path: your app fetches the entry as it already does and hands the fetched entry to resolveOptimizedEntry, which swaps in the selected variant locally, synchronously, for the current visitor. Start from the call itself, then read what it returns.

    The client is @MainActor, so resolution belongs on a main-actor type — a UIView, a UIViewController, or a cell. Resolve where you configure content, so a re-configured view resolves again. One returned field the call site branches on right away: isEmptyVariant, which is true when the visitor's selection is an empty variant — an authored variant with no content, meaning "show nothing here". The paragraphs after the snippet cover the rest of the result.

    Adapt this to your use case:

    import Contentful
    import ContentfulOptimization
    import UIKit

    @MainActor
    final class ArticleCardView: UIView {
    // Injected app-owned client; see Client lifetime and UIKit injection.
    private let client: OptimizationClient
    private let contentView = CardContentView() // Reader-owned UI.

    // entry is the Contentful.Entry your app fetched for this card.
    func configure(with entry: Contentful.Entry) {
    let result = client.resolveOptimizedEntry(
    baseline: entry,
    selectedOptimizations: client.selectedOptimizations
    )

    // An empty variant means "show nothing here for this visitor".
    contentView.isHidden = result.isEmptyVariant
    guard !result.isEmptyVariant else { return }

    // Content type IDs and field names below are your content model's.
    switch result.entry.contentTypeId {
    case "hero" where result.entry.hasField("headline"):
    let headline: String? = result.entry.getField("headline")
    contentView.showHero(headline: headline)
    case "cta" where result.entry.hasField("label"):
    let label: String? = result.entry.getField("label")
    contentView.showCTA(label: label)
    default:
    contentView.showUnsupportedContent()
    }
    }
    }
    1. Fetch one entry with one concrete locale, either by its entry ID or by a route slug. Pass that fetched entry to resolution — never the ID or the slug. Do not pass all-locale payloads (locale=* or all-locale helpers): the resolver cannot read locale-keyed field maps, so those fall back to baseline.
    2. Include linked entries deeply enough to resolve the optimization links. nt_experiences (plural) is the SDK-fixed link field the SDK reads on an optimized entry; it links that entry's nt_experience (singular) experiences, and each experience links its nt_variants and nt_audience entries. These are SDK-owned Optimization content-model names, not names you choose; your fetch must include deeply enough to pull them back in one payload. include: 10 is the reference implementation's pattern.
    3. Keep the app's Contentful locale aligned with SDK locale when rendered content and events must use the same language.
    4. Resolve entries during view, cell, or wrapper configuration.
    5. Branch on result.entry.contentTypeId, check hasField(...), and then read the matching field with getField(...).
    6. Use result.selectedOptimization and result.optimizationContextId only when building tracking payloads.

    Both lookups are ordinary contentful.swift queries through the Contentful client your app already owns. The slug form filters on content type and the slug field as exact equalities and asks for two items, so a duplicate slug is something you can detect rather than silently resolve: return the entry only for exactly one item, send zero items through your not-found path, and treat more than one as an authoring or configuration error. Replace page and slug with your content type and slug-field IDs. The SDK never performs this request or reads these lookup values.

    fetchArray reports through a completion handler, so one continuation makes it awaitable and lets the call site handle a fetch failure instead of dropping it.

    Together the two snippets are the whole path, and it is worth reading in one direction: the view controller fetches a baseline entry, hands it to cardView.configure(with:), and that method — the ArticleCardView above — calls resolveOptimizedEntry, which swaps in the visitor's selected variant locally and synchronously before anything renders. Nothing in the fetch is personalized; the entry that comes back from Contentful is the same for every visitor. Resolution is the only step that differs per visitor, which is why it belongs in view configuration rather than in the fetch: a re-configured or reused cell resolves again against the current selectedOptimizations, while a cached fetch stays valid. The view controller owns the OptimizationClient it injected into the card, so both objects resolve against the same client state.

    Adapt this to your use case:

    import Contentful
    import ContentfulOptimization
    import UIKit

    // One entry by its Contentful entry ID.
    func entryIdQuery(_ entryId: String, locale: String) -> Query {
    Query.where(sys: .id, .equals(entryId))
    .include(10)
    // One concrete locale; an all-locale response falls back to baseline.
    .localizeResults(withLocaleCode: locale)
    }

    // Or one entry by route slug, for routes that carry a public slug.
    func slugQuery(_ routeSlug: String, locale: String) -> Query {
    Query.where(contentTypeId: "page")
    .where(field: "slug", .equals(routeSlug))
    .include(10)
    .localizeResults(withLocaleCode: locale)
    // Two, so a duplicate slug is detectable instead of silently resolved.
    .limit(to: 2)
    }

    @MainActor
    final class ArticleViewController: UIViewController {
    // Both clients are app-owned and injected; see Client lifetime and UIKit injection.
    private let client: OptimizationClient
    private let contentfulClient: Contentful.Client
    private let appLocale: String
    // The view from the snippet above. It holds the same OptimizationClient, and its
    // configure(with:) is the only place resolution happens.
    private let cardView: ArticleCardView

    func loadCard(slug: String) async {
    do {
    let items = try await fetchEntries(matching: slugQuery(slug, locale: appLocale))
    switch items.count {
    case 1:
    // Hand the baseline entry to the card. configure(with:) calls
    // resolveOptimizedEntry, so the variant swap happens inside this line.
    cardView.configure(with: items[0])
    case 0:
    showNotFound() // Reader-owned not-found path.
    default:
    // More than one match is an authoring or configuration error.
    reportAmbiguousSlug(slug)
    }
    } catch {
    reportFetchFailure(error) // Reader-owned error path.
    }
    }

    // fetchArray reports through a completion handler; one continuation makes it awaitable.
    private func fetchEntries(matching query: Query) async throws -> [Contentful.Entry] {
    let response: HomogeneousArrayResponse<Contentful.Entry> =
    try await withCheckedThrowingContinuation { continuation in
    contentfulClient.fetchArray(of: Contentful.Entry.self, matching: query) { result in
    continuation.resume(with: result)
    }
    }
    return response.items
    }
    }

    Now the values the call returns. client.selectedOptimizations (plural) is the SDK's current set of selected optimizations — one selection per experience the visitor's profile matched, published on the client and updated from Experience API responses. resolveOptimizedEntry(baseline:selectedOptimizations:) returns a ResolvedOptimizedEntry — an SDK result type that wraps the resolved entry, the single selectedOptimization (singular) that was applied to it, and an optimizationContextId identifying the optimization context, the profile-and-selection state that produced the variant. Note the one-letter difference: selectedOptimizations is the set you pass in (or the SDK resolves against), while selectedOptimization is the one selection returned on the result. Passing nil for selectedOptimizations uses current client state; an explicit snapshot locks resolution to that selection (see Live updates and locked variants).

    On an empty variant the result still hands back the baseline entry. That is there so interaction tracking keeps its context, not for you to render — UIKit does no hiding for you, which is why the snippet checks result.isEmptyVariant before touching its content. The SDK sets the flag only for a literal boolean true; an absent, false, or otherwise invalid value renders normally.

    ResolvedOptimizedEntry.entry is the SDK-owned CTEntry wrapper around the resolved entry. A selected linked variant can use any Contentful content type, and contentTypeId identifies that type without validating its fields — which is why the pattern is branch on contentTypeId, confirm the field with hasField(...), then read it with getField(...).

    resolveOptimizedEntry is synchronous and fail-soft: it never throws or breaks the UI. Two different situations hand back the baseline, and they differ in what metadata comes with it:

    • A client-side failure — the client is not initialized, the baseline cannot be serialized, or the bridge result cannot be parsed — returns the baseline entry unchanged with selectedOptimization and optimizationContextId nil, and logs a warning.
    • A resolver-side fallback — no matching selection, unresolved nt_experiences or nt_variants links, an all-locale payload, or a selection that points at the baseline instead of a variant — also returns the baseline, but it can arrive with selected-optimization metadata attached. Treat non-nil metadata as "the resolver had a selection for this entry", not as proof that a variant replaced it.

    That distinction is what Entry interaction tracking builds on when it derives a TrackingMetadata from result.selectedOptimization. For the shared resolution and fallback rules, see Entry optimization and variant resolution.

    Before you start assumes you already have a contentful.swift client. If you do not, construct one from your space ID, environment, and Delivery API token — the queries above run through it:

    Adapt this to your use case:

    import Contentful

    let contentfulClient = Client(
    spaceId: "<your-space-id>",
    environmentId: "<your-environment-id>",
    accessToken: "<your-delivery-api-token>"
    )

    Integration category: Required for first integration

    The quick start tracked one screen. Real UIKit navigation repeats lifecycle callbacks across modal, tab, and navigation-controller transitions, so choose the method that matches the event you want.

    trackCurrentScreen(name:properties:routeKey:) and screen(name:properties:) both emit the same underlying screen event and return the same result shape; the difference is dedupe. trackCurrentScreen deduplicates the current route in the SDK by routeKey (which defaults to name), so a repeat of the same current screen from a repeated viewDidAppear(_:) is skipped, and a blocked attempt is retried once consent allows. Plain screen(name:properties:) re-emits on every call with no dedupe — call it directly only when your app wants that: an intentional one-off event, or navigation tracking where every appearance should count as a fresh screen event rather than deduping repeats of the same route.

    1. Emit from viewDidAppear(_:) so UIKit has completed the visible transition.
    2. Use a stable screen name that maps to your analytics model.
    3. Pass a stable routeKey when several instances of one destination should still count as the same current screen, or when the default name-based key would collide.
    4. Add properties only when the downstream analysis needs them.

    Adapt this to your use case:

    override func viewDidAppear(_ animated: Bool) {
    super.viewDidAppear(animated)
    Task { @MainActor in
    // entryId comes from your route or Contentful lookup; when it is a Contentful entry ID, use the opaque ID.
    _ = try? await client.trackCurrentScreen(
    name: "ProductDetail",
    properties: ["entryId": entryId],
    // Stable route key prevents duplicate current-screen events when the
    // lifecycle callback repeats for the same destination.
    routeKey: "product-detail-\(entryId)"
    )
    }
    }

    For shared tracking mechanics and event delivery, see iOS SDK runtime and interaction mechanics.

    Entry interaction tracking

    Integration category: Common but policy-dependent

    UIKit does not automatically infer when a user tapped a Contentful entry or when an entry met a visibility threshold, so your app owns the geometry and the app decides whether these events are allowed by its Analytics and privacy policy. Entry views deliver on the wire as component events; entry taps as component_click.

    Entry taps. Build a TrackingMetadata (an SDK helper type that derives componentId/experienceId/variantIndex from an entry and its selected optimization) from the resolution you already rendered, then pass its fields to a TrackClickPayload (an SDK payload type). Building the metadata from the stored resolution — not by re-resolving at tap time — makes the tap carry the same optimization context that produced the rendered variant.

    1. Resolve and render the entry, and store both it and the ResolvedOptimizedEntry you rendered from.
    2. On tap, build TrackingMetadata from the stored entry and the resolution's selectedOptimization.
    3. Call client.trackClick(TrackClickPayload(...)) from a UIControl action or gesture recognizer. For gesture recognizers, gate the dispatch to the completed gesture state instead of suppressing later taps for the view's lifetime.

    Keep the fetched Contentful.Entry and the resolution it produced side by side, both typed. TrackingMetadata takes the baseline as a dictionary rather than a typed entry, and reads exactly one value out of it: sys.id, which becomes componentId. The experience, variant index, and sticky flag all come from selectedOptimization. Encode with CTEntry at that call and nowhere else — the dictionary is what one SDK initializer accepts, not a shape to store in your own view, which stays typed. Both stored properties and both methods below belong to the same view, cell, or view controller that resolved the entry.

    Adapt this to your use case:

    // Reader-owned: your view or cell stores the entry and resolution it rendered,
    // both typed.
    private var latestBaselineEntry: Contentful.Entry?
    private var latestResolution: ResolvedOptimizedEntry?

    func configure(with entry: Contentful.Entry) {
    let result = client.resolveOptimizedEntry(
    baseline: entry,
    selectedOptimizations: client.selectedOptimizations
    )
    latestBaselineEntry = entry
    latestResolution = result
    contentView.isHidden = result.isEmptyVariant
    guard !result.isEmptyVariant else { return }
    contentView.configure(with: result.entry) // contentView is reader-owned UI.
    }

    @objc private func primaryButtonTapped() {
    guard let entry = latestBaselineEntry, let result = latestResolution else { return }

    // TrackingMetadata carries the optimization context that produced the
    // rendered variant, so the tap matches what the visitor actually saw. The
    // CTEntry encode stays at this boundary, so the view keeps a typed entry.
    let metadata = TrackingMetadata(
    entry: CTEntry(entry).toDictionary(),
    optimizationContextId: result.optimizationContextId,
    selectedOptimization: result.selectedOptimization
    )

    Task { @MainActor in
    try? await client.trackClick(TrackClickPayload(
    componentId: metadata.componentId,
    experienceId: metadata.experienceId,
    optimizationContextId: metadata.optimizationContextId,
    variantIndex: metadata.variantIndex
    ))
    }
    }

    Entry views. Feed app-owned scroll or layout geometry to a ViewTrackingController — the SDK's imperative view-timing engine for UIKit — and it applies the same timing model and emits a TrackViewPayload (an SDK payload type) through the client for you. The controller uses a fixed model: a view session begins when the entry reaches 10% visibility and qualifies after a continuous 1000 ms dwell. A qualified view session produces two normal event records whose type field is component: the first when it qualifies and the second, final record when visibility falls below 10% or another ending occurs. Both records carry the same SDK-owned viewId. The viewDurationMs value is milliseconds measured from the moment that view session began at 10% visibility, so it includes the qualifying dwell. A view session that ends before qualification emits no record, and active view sessions emit no periodic duration records.

    The controller observes app background and foreground transitions automatically. Backgrounding ends and resets the current view session. On foreground, the controller checks the last measured entry and viewport positions; if the entry is still visible, it starts a fresh view session that must satisfy the continuous 1000 ms dwell again. Your UIKit integration must still feed geometry updates and call onDisappear() when the tracked view leaves the screen or its tracking metadata is replaced. The controller also dedupes its own sticky views.

    Those call sites are the part a UIKit app is responsible for, and there are three: geometry updates while the entry is on screen, a rebuild when a new resolution changes the tracking metadata, and onDisappear() when the entry leaves the screen. Skip the last one and a view session that remains above the threshold can stay open, so the final record is not emitted — and a reused cell keeps the previous entry's view session open. contentHost is a reader-owned container: the class's omitted initializer and layout code must add it with addSubview(contentHost) and size or constrain it before configure() runs; the SDK does not create or mount that container. ViewTrackingController and TrackingMetadata read the baseline entry in dictionary form — they have no typed-entry initializer — so a typed Contentful.Entry is encoded once with CTEntry and that one dictionary is reused for every rebuild.

    Follow this pattern:

    final class OptimizedEntryView: UIView {
    private let client: OptimizationClient
    private let contentHost = UIView()
    private let entry: Contentful.Entry
    // ViewTrackingController takes the baseline as a dictionary, so the typed
    // entry is encoded once here and reused for every rebuild.
    private lazy var entryDictionary: [String: Any] = CTEntry(entry).toDictionary()
    private weak var scrollView: UIScrollView?
    private var trackingController: ViewTrackingController?
    private var offsetObservation: NSKeyValueObservation?

    // Call site 1: resolve the entry when the view is configured, then (re)build
    // the controller for that resolution — the same place you render the entry.
    func configure() {
    let result = client.resolveOptimizedEntry(baseline: entry)
    rebuildTracking(result: result)
    contentHost.isHidden = result.isEmptyVariant
    if !result.isEmptyVariant {
    // ...render result.entry inside contentHost with your own view code...
    }
    }

    // Rebuild the controller whenever a newly resolved variant changes the
    // tracking metadata, ending the previous view session first.
    private func rebuildTracking(result: ResolvedOptimizedEntry) {
    trackingController?.onDisappear()
    trackingController = ViewTrackingController(
    client: client,
    entry: entryDictionary,
    optimizationContextId: result.optimizationContextId,
    selectedOptimization: result.selectedOptimization
    )
    observeScrollOffset()
    emitVisibility()
    }

    // Call site 2: feed geometry on every scroll change and every layout pass so
    // the controller can run its timing model.
    private func observeScrollOffset() {
    offsetObservation?.invalidate()
    offsetObservation = scrollView?.observe(\.contentOffset, options: [.new]) { [weak self] _, _ in
    Task { @MainActor in self?.emitVisibility() }
    }
    }

    override func layoutSubviews() {
    super.layoutSubviews()
    emitVisibility()
    }

    override func didMoveToWindow() {
    super.didMoveToWindow()
    guard window != nil else { return }
    observeScrollOffset()
    emitVisibility()
    }

    // Call site 3: end the view session when the entry leaves the screen, and stop
    // observing with it. A reusable cell runs the same teardown from
    // prepareForReuse().
    override func willMove(toWindow newWindow: UIWindow?) {
    super.willMove(toWindow: newWindow)
    guard newWindow == nil else { return }
    trackingController?.onDisappear()
    offsetObservation?.invalidate()
    offsetObservation = nil
    }

    // Reader-owned geometry: your app computes the element's position and feeds
    // it to the controller, which owns timing, consent checks, and the view session's
    // single final record.
    private func emitVisibility() {
    guard let controller = trackingController, let scrollView else { return }
    let frameInScroll = convert(bounds, to: scrollView)
    controller.updateVisibility(
    elementY: frameInScroll.minY,
    elementHeight: bounds.height,
    scrollY: scrollView.contentOffset.y,
    viewportHeight: scrollView.bounds.height
    )
    }
    }

    ViewTrackingController is the recommended path because it applies the SDK's visibility timing for you. If your app already computes its own visibility and duration — or needs a single one-off view event — call client.trackView(TrackViewPayload(...)) directly instead of using the controller; it is the lower-level async throws primitive the controller wraps, and you then own the timing the controller would otherwise apply.

    To opt an entry out of view or tap tracking, do not install its controller or gesture recognizer. For shared tracking mechanics, see iOS SDK runtime and interaction mechanics.

    Integration category: Common but policy-dependent

    Identity policy belongs to the application. The SDK can identify a visitor, update selected optimizations and changes (the inline field and flag values the Experience API returned for the visitor) from Experience API responses, persist profile-continuity state when allowed, and reset SDK-managed profile state, but it does not decide when a user becomes known or how account data is governed.

    1. Call identify(userId:traits:) after sign-in or when the app has a stable application user ID.
    2. Wait for SDK state or rendered content before assuming the profile has affected visible entries.
    3. Call reset() on logout, account switch, or a privacy flow that must clear SDK-managed profile, selected-optimization, change, and anonymous-ID state.
    4. Preserve or clear app-owned user identifiers according to your account and privacy policy; the SDK does not clear your application storage.

    Adapt this to your use case:

    Task { @MainActor in
    // identify links the app-owned user ID to the current mobile profile.
    _ = try? await client.identify(
    userId: user.id,
    traits: ["plan": user.plan]
    )
    }

    Copy this:

    // reset() clears profile continuity but preserves consent state.
    client.reset()

    reset() clears profile continuity (profile, changes, selected optimizations, the anonymous ID, the current-screen dedupe tracker, and sticky-view keys) and any preview-panel overrides currently applied, but preserves consent state, and it no-ops before initialization. When persistence consent is allowed, the SDK writes continuity to UserDefaults before publishing SDK state from an Experience response. In tests and relaunch flows, wait for SDK-derived UI or state instead of adding arbitrary storage delays. The SDK persists to UserDefaults under the com.contentful.optimization. prefix, not to cookies, and provides no built-in cross-platform identity handoff — implement any web, server, or account continuity in application code. For the identifier model, see Consent management in the Optimization SDK Suite.

    Integration category: Optional

    Use custom events for business actions that are not tied to a Contentful entry swap, and the event streams for local diagnostics or app-owned analytics forwarding.

    1. Call track(event:properties:) for a business event.
    2. Subscribe to eventStream for accepted events; subscribe to blockedEventStream (or configure onEventBlocked at startup) for events stopped by consent or the allow-list.
    3. Subscribe before the events you want to observe fire — eventStream is a passthrough publisher that does not replay earlier events to late subscribers.

    Copy this:

    Task { @MainActor in
    // A custom business event, not tied to a Contentful entry swap.
    _ = try? await client.track(event: "Purchase Completed", properties: ["sku": "ABC-123"])
    }

    Both streams are Combine publishers, so they need import Combine and the cancellables bag from Client lifetime and UIKit injection on whichever object subscribes.

    Adapt this to your use case:

    // eventStream is a passthrough publisher with no replay: subscribe before the
    // events you want to observe fire, or you miss the earlier ones.
    client.eventStream
    // analyticsDebugStore is reader-owned: your own diagnostic sink.
    .sink { event in analyticsDebugStore.append(event) }
    .store(in: &cancellables)

    // blockedEventStream surfaces events stopped by consent or the allow-list —
    // the diagnostic for a missing event during integration.
    client.blockedEventStream
    .sink { blocked in print("blocked \(blocked.method): \(blocked.reason)") }
    .store(in: &cancellables)

    When forwarding SDK events to third-party destinations, apply the same app-owned consent policy, deduplication, and data-minimization rules that govern the destination. For destination mapping, consent, identity, dedupe, and governance guidance, see Forwarding Optimization SDK context to analytics and tag-management tools.

    Integration category: Optional

    A Custom Flag is a named value an experience personalizes instead of swapping a whole entry — a badge label, a discount percentage, a boolean that turns a section on. The Experience API returns the visitor's values as the changes on SDK state, and getFlag(_:) reads one of them by name. That name is not one you invent: it must match the flag authored in your Optimization data, the way "show-promo" does below. Merge tags are the Rich Text counterpart: profile-driven text substitutions inside a Rich Text field. Both read from SDK state separately from entry-variant resolution.

    1. Read a flag once with getFlag(_:) when a synchronous value is enough.
    2. Subscribe with flagPublisher(_:) when the UI must update as flag values change.
    3. Resolve merge tags with getMergeTagValue(mergeTagEntry:) from your app-owned Rich Text renderer.

    Copy this:

    // Non-reactive one-shot read; returns nil before init or when unresolved.
    let flagValue = client.getFlag("show-promo")

    Adapt this to your use case:

    // Subscribing registers an observeFlag subscription. A flag subscription emits
    // a component flag-view event (an analytics exposure) when consent and profile
    // allow, so treat it as tracked exposure, not a free read, and govern it like
    // any other event.
    client.flagPublisher("show-promo")
    .receive(on: RunLoop.main)
    // applyPromoFlag(_:) is reader-owned: your own UI update for the value.
    .sink { [weak self] value in self?.applyPromoFlag(value) }
    .store(in: &cancellables)

    nt_mergetag is the SDK-fixed Optimization content type for a merge tag — a profile-driven text substitution embedded inline in Rich Text; it is not a name you choose. Your app owns extracting the embedded nt_mergetag entry from the Rich Text node before calling the SDK, which resolves the selector against the current profile and returns the resolved string or nil.

    Follow this pattern:

    // mergeTagEntry is reader-owned: the expanded embedded-entry-inline node's
    // data.target you extracted from Rich Text.
    let resolved = client.getMergeTagValue(mergeTagEntry: mergeTagEntry)
    // resolved is String?; nil means neither the profile value nor the merge tag's
    // own configured fallback field resolved.

    The resolver already falls back to the merge tag's configured fallback field, so nil means neither the profile value nor that fallback resolved. Substituting your own placeholder text on nil is defensive rendering, not a required step.

    For the deeper data model, see Entry optimization and variant resolution.

    Integration category: Optional

    Live updates in UIKit exist, and they are entirely app-driven. There is no view wrapper and no automatic mechanism: when the SDK publishes new selections, nothing on screen changes until your code resolves again and redraws. Both options in this section are therefore policies you implement — lock a screen to the selections its first render used, or subscribe and redraw — and you choose between them by what you pass for selectedOptimizations and whether you subscribe at all.

    1. To lock a screen, capture client.selectedOptimizations ?? [] at the screen's first resolution and pass that explicit snapshot to every resolveOptimizedEntry call on the screen. Do not pass nil for locked screens, because nil tells the resolver to use current SDK state.
    2. To update live, pass nil (or the current client.selectedOptimizations) and subscribe to client.$selectedOptimizations to redraw affected views when selections change.
    3. Decide what an open preview panel means for your screens. Nothing forces live updates in UIKit, so overrides applied in the panel appear only if you redraw: subscribe to client.$isPreviewPanelOpen and client.$previewState alongside $selectedOptimizations, and treat an open panel as a reason to re-resolve even on an otherwise locked screen.

    Adapt this to your use case:

    // nil until this screen has resolved once; an explicit snapshot afterwards.
    private var lockedOptimizations: [[String: Any]]?

    // Call site: every render on this screen goes through here — the first one from
    // viewDidLoad or cell configuration, and every later redraw.
    func render(entry: Contentful.Entry) {
    // The first render is what locks. Empty array locks to no selections; nil
    // would keep asking for current SDK state on every later render.
    if lockedOptimizations == nil {
    lockedOptimizations = client.selectedOptimizations ?? []
    }

    let result = client.resolveOptimizedEntry(
    baseline: entry,
    selectedOptimizations: lockedOptimizations
    )
    contentView.isHidden = result.isEmptyVariant
    guard !result.isEmptyVariant else { return }
    contentView.configure(with: result.entry)
    }

    Locking on the first render keeps the screen from waiting on a separate locking step. Selections come from an Experience API response, though, so the first render can happen before any selection exists — which locks the screen to none. The SwiftUI view surface handles this by locking on the first non-nil value instead; in UIKit that decision is yours: to match it, set lockedOptimizations back to nil and render again when a $selectedOptimizations subscription delivers the first non-nil value.

    Adapt this to your use case:

    client.$selectedOptimizations
    // @Published fires in willSet, so hop to the next run-loop turn to read the
    // committed selections before re-resolving.
    .receive(on: RunLoop.main)
    .sink { [weak self] _ in
    // liveUpdates and reloadVisibleContent() are reader-owned: your screen's
    // own policy flag and its redraw.
    guard self?.client.isPreviewPanelOpen == true || self?.liveUpdates == true else { return }
    self?.reloadVisibleContent()
    }
    .store(in: &cancellables)

    cancellables is again the subscription bag from Client lifetime and UIKit injection, so this subscription needs import Combine on the file that holds it.

    For the precedence between live updates, locked variants, and preview-panel state, see iOS SDK runtime and interaction mechanics.

    Integration category: Optional

    PreviewPanelViewController hosts the SDK preview panel from a UIKit view controller. Gate it behind a debug or internal-build condition so production users cannot open local audience and variant overrides.

    1. Supply a Contentful client for the space and environment holding your Optimization entries. Pass an existing contentful.swift Contentful.Client directly and the SDK wraps it for you, or use the built-in ContentfulHTTPPreviewClient when the app has no Contentful client to share. Either one fetches the nt_audience and nt_experience definitions.
    2. Add the floating button to a host controller with addFloatingButton(to:client:contentfulClient:), passing the same initialized OptimizationClient the rest of the app uses so overrides affect the same resolver and event state.
    3. Keep the preview panel out of public production builds unless your release policy explicitly allows it for an internal audience.

    The button attaches to a view controller that already exists, so the natural place for this is where the scene builds its root controller — the same SceneDelegate method the quick start edited, where both the client and the host controller are in scope.

    Adapt this to your use case:

    func scene(
    _ scene: UIScene,
    willConnectTo _: UISceneSession,
    options _: UIScene.ConnectionOptions
    ) {
    // ...window and root controller setup from the quick start...

    #if DEBUG
    let previewContentfulClient = ContentfulHTTPPreviewClient(
    spaceId: "your-space-id",
    accessToken: "your-cda-token",
    // Defaults to "master" on this client, so pass yours explicitly.
    environment: "main"
    )

    PreviewPanelViewController.addFloatingButton(
    to: home,
    // Pass the app-owned client so overrides affect the same resolver and state.
    client: client,
    contentfulClient: previewContentfulClient
    )
    #endif
    }

    Passing contentfulClient is what loads audience and experience definitions by name. Without it the panel can still open, but no definitions are loaded: the audience section is empty, audience and variant override controls are unavailable, and existing override summaries can fall back to raw identifiers.

    Integration category: Optional

    Use this section when the app can change language or locale after SDK startup. The SDK locale and the Contentful CDA locale are separate inputs, even when they usually carry the same value.

    1. Derive the next app locale from your navigation, i18n, account, or settings layer.
    2. Call setLocale(_:) to update the SDK Experience and event locale. It updates the SDK locale only — it does not refetch Contentful entries or refresh profile state — and it throws before init or on an invalid locale.
    3. Refetch Contentful entries with the same locale and re-resolve visible entries once the localized payload and SDK state are both ready.
    4. Invalidate app-owned content caches using locale-aware cache keys.

    Adapt this to your use case:

    // Called from your language picker or settings screen. The refetch is async, so
    // the whole sequence lives in one async method (or a Task) on that controller.
    @MainActor
    func applyLocale(_ nextLocale: String) async {
    do {
    // Updates the SDK Experience/event locale only; throws on an invalid locale.
    try client.setLocale(nextLocale)

    // Reader-owned refetch in the same locale, then re-resolve and redraw.
    entries = try await contentfulEntryService.fetchEntries(
    ids: entryIds,
    include: 10,
    locale: nextLocale
    )
    reloadVisibleContent()
    } catch {
    // Reader-owned: keep the previous locale's content on screen.
    reportLocaleChangeFailure(error)
    }
    }

    For the full locale model, see Locale handling in the Optimization SDK Suite.

    Integration category: Advanced or production-only

    The iOS SDK monitors network reachability, queues events while offline, flushes when connectivity returns, and flushes as the app moves toward the background. No setup is required for the default offline path: NWPathMonitor drives the SDK online state and flushes on reconnect, and the app lifecycle handler flushes on willResignActive.

    1. Add QueuePolicy only when production telemetry needs queue limits or lifecycle callbacks. The offline Experience queue holds up to 100 events by default (tunable via QueuePolicy.offlineMaxEvents); queues are in-memory only and do not survive process death.
    2. Use queue callbacks for operational diagnostics, not for resending blocked or dropped events.
    3. Keep Contentful entry caching in the application layer — the SDK does not cache CDA responses for UIKit rendering.
    4. Call flush() only for deliberate release, test, or lifecycle flows; the SDK already flushes on background and reconnect.

    Adapt this to your use case:

    // Again the one config the scene initializes with, plus one key.
    let config = OptimizationConfig(
    clientId: "your-optimization-client-id",
    queuePolicy: QueuePolicy(
    offlineMaxEvents: 500,
    onOfflineDrop: { event in
    // event is a QueueEvent with a type and a context dictionary;
    // diagnostics is reader-owned: your own operational logging.
    diagnostics.record("optimization-offline-drop", context: event.context)
    },
    onFlushFailure: { event in
    diagnostics.record("optimization-flush-failure", context: event.context)
    },
    onFlushRecovered: { event in
    diagnostics.record("optimization-flush-recovered", context: event.context)
    }
    )
    )

    Before release, verify the UIKit integration against these checks:

    • Credentials and runtime configuration — The app uses the intended Optimization client ID, Contentful environment, SDK locale, and CDA locale. Non-default API base URLs and .debug logging are absent from production builds unless explicitly approved.

    • Consent behavior — Startup consent, CMP wiring, refusal, withdrawal, split event and persistence consent, and reset() behavior match the app's legal and privacy requirements.

    • Event delivery — Screen, custom, tap, view, identify, and flag-view events appear when allowed and are blocked or omitted when policy denies them.

    • Content fallback behavior — Baseline entries render when selected optimizations are missing, unresolved links are returned, variants are out of range, or the visitor is not qualified. Every supported resolved content type maps to a renderer; a variant content type that differs from the baseline is rendered as that variant, not treated as fallback.

    • Duplicate tracking prevention — UIKit lifecycle hooks, reusable cells, gesture recognizers, and visibility observers do not emit duplicate screen, tap, or view events for one intended interaction or visibility cycle.

    • Privacy and governance — Preview-panel access, event forwarding, profile IDs, user traits, app-owned caches, and diagnostics follow the app's data-minimization and retention policy.

    • Confirm in Live Events — in addition to local log and status checks, open the target Contentful space and environment's Live Events view in the Contentful web app, trigger a real flow from the app (a screen view, an entry view or tap, an identify() call, or a custom track() call), and confirm the corresponding event arrives with the expected wire type (identify, screen, component, component_click, or track) and payload fields.

    • Local validation path — Compare your integration against the iOS reference implementation. The repository's maintainers validate UIKit behavior with an XCUITest suite driven from implementations/ios-sdk/; that runner is a maintainer command, not an app command.

      Reference excerpt:

      # From implementations/ios-sdk/ in the optimization monorepo — a maintainer
      # command that builds the JS bridge, starts the mock server, and runs XCUITest.
      APP_SHELL=uikit ./scripts/run-e2e.sh
    • The build fails with No such module 'ContentfulOptimization' — Confirm the package is added to the app target's dependencies (the product is named ContentfulOptimization), build the app target once so Swift Package Manager resolves and compiles it, and confirm the target's minimum deployment version is iOS 15 or later. If import Contentful is what fails to resolve, add contentful.swift to the app target as well.
    • The app crashes on launch with init(coder:) is not supported — A storyboard or nib is instantiating a view controller whose init(coder:) was replaced by the constructor-injection form. Keep UIKit's init(coder:), take the client through a settable property or an app-level dependency container, and set it before the view appears.
    • The quick-start label reads Optimization screen event blocked — Two causes. If the consent gate rejected the event, onEventBlocked prints a line prefixed Optimization blocked naming the reason and method, so search the Xcode console for that prefix, then check consent and allowedEventTypes. If the client never initialized, the caught error prints instead, and the SDK's logLevel: .debug output under the com.contentful.optimization subsystem shows the failed init. initialize throws a configError for an invalid locale and a resourceLoadError when the packaged bridge resource cannot be loaded, so the printed error names which of the two you hit.
    • Optimized entries always render the baseline — Confirm the app fetched a single-locale entry, requested enough include depth for nt_experiences and nt_variants, initialized the client, and has non-empty client.selectedOptimizations for the visitor.
    • Tap or view events do not appear — Check consent, allowedEventTypes, the componentId from TrackingMetadata, UIKit gesture wiring, and whether the view reached the configured visibility threshold long enough to emit.
    • Screen events appear more than once — Review viewDidAppear(_:) calls for modal, tab, and navigation-controller transitions, and prefer trackCurrentScreen with a stable routeKey over raw screen for lifecycle tracking.
    • Preview panel opens but shows identifiers — Pass a PreviewContentfulClient that can fetch nt_audience and nt_experience entries from the correct space and environment.
    • Identified variants disappear after relaunch — Verify persistence consent is true, wait for SDK-published profile or selected-optimization state before terminating tests, and confirm logout or withdrawal flows are not calling reset().
    • iOS reference implementation — Maintained SwiftUI and UIKit shells that exercise the native iOS bridge against the shared mock API: accepted-consent startup, single-locale CDA fetching, entry resolution, screen tracking, interaction tracking, Custom Flags and merge tags, live updates, offline queueing, and preview-panel overrides. Use it as the comparison and validation target for UIKit integration behavior.