@iternio/react-native-auto-play 0.5.3 → 0.5.4

package/android/src/main/java/com/margelo/nitro/swe/iternio/reactnativeautoplay/HybridVoice.kt

@@ -64,6 +64,8 @@ class HybridVoice : HybridVoiceSpec() {
         silenceThresholdMs: Double?,
         maxDurationMs: Double?,
         listeningText: String?,
+        listeningImage: Variant_GlyphImage_AssetImage_RemoteImage?,
+        listeningImageRepeats: Boolean?,
         preferSpeechToText: Boolean?,
         onChunk: ((chunk: VoiceInputChunk) -> Unit)?,
         language: String?

package/ios/hybrid/HybridAutoPlay.swift

@@ -22,7 +22,6 @@ class HybridAutoPlay: HybridAutoPlaySpec {
     private static var renderStateListeners = [String: [RenderStateListener]]()
     private static var safeAreaInsetsListeners = [String: [SafeAreaListener]]()
 
-
     override init() {
         HybridAutoPlay.listeners.removeAll()
         HybridAutoPlay.renderStateListeners.removeAll()

package/ios/hybrid/HybridVoice.swift

@@ -32,6 +32,8 @@ class HybridVoice: HybridVoiceSpec {
         silenceThresholdMs: Double?,
         maxDurationMs: Double?,
         listeningText: String?,
+        listeningImage: Variant_GlyphImage_AssetImage_RemoteImage?,
+        listeningImageRepeats: Bool?,
         preferSpeechToText: Bool?,
         onChunk: ((_ chunk: VoiceInputChunk) -> Void)?,
         language: String?
@@ -49,6 +51,8 @@ class HybridVoice: HybridVoiceSpec {
                 silenceThresholdMs: silenceThresholdMs ?? 1_500,
                 maxDurationMs: maxDurationMs ?? 10_000,
                 listeningText: listeningText ?? "Listening...",
+                listeningImage: listeningImage,
+                listeningImageRepeats: listeningImageRepeats,
                 preferSpeechToText: preferSpeechToText ?? false,
                 onChunk: onChunk,
                 language: language

package/ios/templates/Parser.swift

@@ -6,6 +6,7 @@
 //
 
 import CarPlay
+import ImageIO
 import UIKit
 
 struct HeaderActions {
@@ -939,6 +940,100 @@ class Parser {
         )
     }
 
+    // MARK: - Animated image decoding
+
+    /// Decodes raw image data via ImageIO, walking every frame so animated GIF/APNG/WebP all
+    /// animate. UIImage(data:) only ever decodes the first frame for any of these formats.
+    /// `maxDuration` caps the assembled cycle length; pass `.greatestFiniteMagnitude` to skip capping.
+    static func decodeImage(data: Data, scale: CGFloat, maxDuration: TimeInterval = .greatestFiniteMagnitude)
+        -> UIImage?
+    {
+        guard let source = CGImageSourceCreateWithData(data as CFData, nil) else { return nil }
+        let frameCount = CGImageSourceGetCount(source)
+        guard frameCount > 0 else { return nil }
+
+        guard frameCount > 1 else {
+            guard let cgImage = CGImageSourceCreateImageAtIndex(source, 0, nil) else { return nil }
+            return UIImage(cgImage: cgImage, scale: scale, orientation: .up)
+        }
+
+        var frames: [UIImage] = []
+        var totalDuration: TimeInterval = 0
+        for index in 0..<frameCount {
+            guard let cgImage = CGImageSourceCreateImageAtIndex(source, index, nil) else { continue }
+            totalDuration += frameDuration(source: source, index: index)
+            frames.append(UIImage(cgImage: cgImage, scale: scale, orientation: .up))
+        }
+        guard !frames.isEmpty else { return nil }
+        return UIImage.animatedImage(with: frames, duration: min(totalDuration, maxDuration))
+    }
+
+    /// Reads the per-frame delay from whichever format dictionary ImageIO populated
+    /// (GIF, APNG, or WebP), falling back to a sane default if none is present.
+    private static func frameDuration(source: CGImageSource, index: Int) -> TimeInterval {
+        let defaultDuration: TimeInterval = 0.1
+        guard
+            let properties = CGImageSourceCopyPropertiesAtIndex(source, index, nil) as? [CFString: Any]
+        else { return defaultDuration }
+
+        if let gif = properties[kCGImagePropertyGIFDictionary] as? [CFString: Any] {
+            if let unclamped = gif[kCGImagePropertyGIFUnclampedDelayTime] as? Double, unclamped > 0 {
+                return unclamped
+            }
+            if let delay = gif[kCGImagePropertyGIFDelayTime] as? Double, delay > 0 {
+                return delay
+            }
+        }
+
+        if let png = properties[kCGImagePropertyPNGDictionary] as? [CFString: Any] {
+            if let unclamped = png[kCGImagePropertyAPNGUnclampedDelayTime] as? Double, unclamped > 0 {
+                return unclamped
+            }
+            if let delay = png[kCGImagePropertyAPNGDelayTime] as? Double, delay > 0 {
+                return delay
+            }
+        }
+
+        if let webp = properties[kCGImagePropertyWebPDictionary] as? [CFString: Any],
+            let delay = webp[kCGImagePropertyWebPDelayTime] as? Double, delay > 0
+        {
+            return delay
+        }
+
+        return defaultDuration
+    }
+
+    private static func targetSize(for size: CGSize, max maxSize: CGSize) -> CGSize {
+        guard size.width > maxSize.width || size.height > maxSize.height else { return size }
+        let scale = min(maxSize.width / size.width, maxSize.height / size.height)
+        return CGSize(width: (size.width * scale).rounded(), height: (size.height * scale).rounded())
+    }
+
+    static func resize(_ image: UIImage, max maxSize: CGSize) -> UIImage {
+        let target = targetSize(for: image.size, max: maxSize)
+        guard target != image.size else { return image }
+        return UIGraphicsImageRenderer(size: target).image { _ in
+            image.draw(in: CGRect(origin: .zero, size: target))
+        }
+    }
+
+    /// Resizes every frame of an animated UIImage while preserving the per-frame timing.
+    /// UIImage.draw(in:) only renders the current frame, so resize() alone would collapse
+    /// the animation to a still image.
+    static func resizeAnimated(_ image: UIImage, max maxSize: CGSize) -> UIImage {
+        guard let frames = image.images, !frames.isEmpty else {
+            return resize(image, max: maxSize)
+        }
+        let target = targetSize(for: image.size, max: maxSize)
+        guard target != image.size else { return image }
+        let resizedFrames = frames.map { frame in
+            UIGraphicsImageRenderer(size: target).image { _ in
+                frame.draw(in: CGRect(origin: .zero, size: target))
+            }
+        }
+        return UIImage.animatedImage(with: resizedFrames, duration: image.duration) ?? image
+    }
+
     static func imageFromLanes(
         laneImages: Array<NitroImage>.SubSequence,
         traitCollection: UITraitCollection

package/ios/utils/VoiceInputManager.swift

@@ -65,6 +65,8 @@ class VoiceInputManager {
         silenceThresholdMs: Double,
         maxDurationMs: Double,
         listeningText: String,
+        listeningImage: Variant_GlyphImage_AssetImage_RemoteImage?,
+        listeningImageRepeats: Bool?,
         preferSpeechToText: Bool,
         onChunk: ((_ chunk: VoiceInputChunk) -> Void)?,
         language: String?
@@ -82,6 +84,8 @@ class VoiceInputManager {
                     silenceThresholdMs: silenceThresholdMs,
                     maxDurationMs: maxDurationMs,
                     listeningText: listeningText,
+                    listeningImage: listeningImage,
+                    listeningImageRepeats: listeningImageRepeats,
                     preferSpeechToText: preferSpeechToText,
                     onChunk: onChunk,
                     box: box,
@@ -128,6 +132,8 @@ class VoiceInputManager {
         silenceThresholdMs: Double,
         maxDurationMs: Double,
         listeningText: String,
+        listeningImage: Variant_GlyphImage_AssetImage_RemoteImage?,
+        listeningImageRepeats: Bool?,
         preferSpeechToText: Bool,
         onChunk: ((_ chunk: VoiceInputChunk) -> Void)?,
         box: ResultBox,
@@ -142,13 +148,20 @@ class VoiceInputManager {
         try session.setActive(true)
 
         if let interfaceController {
-            presentVoiceTemplate(interfaceController: interfaceController, listeningText: listeningText)
+            presentVoiceTemplate(
+                interfaceController: interfaceController,
+                listeningText: listeningText,
+                listeningImage: listeningImage,
+                listeningImageRepeats: listeningImageRepeats
+            )
         }
 
         var activeRecognitionRequest: SFSpeechAudioBufferRecognitionRequest? = nil
 
         if preferSpeechToText, SFSpeechRecognizer.authorizationStatus() == .authorized,
-            let recognizer = language != nil ? SFSpeechRecognizer(locale: Locale(identifier: language!)) : SFSpeechRecognizer(locale: Locale.current),
+            let recognizer = language != nil
+                ? SFSpeechRecognizer(locale: Locale(identifier: language!))
+                : SFSpeechRecognizer(locale: Locale.current),
             recognizer.isAvailable
         {
             let request = SFSpeechAudioBufferRecognitionRequest()
@@ -311,12 +324,81 @@ class VoiceInputManager {
         return VoiceInputResult(transcription: nil, audio: buffer)
     }
 
-    private func presentVoiceTemplate(interfaceController: AutoPlayInterfaceController, listeningText: String) {
+    // CPVoiceControlState enforces a maximum image size of 150x150 points.
+    private static let voiceImageMaxSize = CGSize(width: 150, height: 150)
+
+    // CPVoiceControlState also enforces a 0.3s–5s animation cycle; the 0.3s floor is applied
+    // by the system regardless of what we pass, so we only need to clamp our own ceiling.
+    private static let maxVoiceImageCycleDuration: TimeInterval = 5.0
+
+    // Bypasses RCTConvert for asset images: it collapses animated UIImages to a single frame
+    // via CGImage during scale adjustment. Parser.decodeImage preserves animation frames by
+    // walking every frame in the source via ImageIO — UIImage(data:) never builds a multi-frame
+    // .images array itself, for GIF, APNG, or WebP. Tinting is skipped for animated images since
+    // frames cannot be tinted individually.
+    private func loadVoiceImage(image: Variant_GlyphImage_AssetImage_RemoteImage?, traitCollection: UITraitCollection)
+        -> UIImage?
+    {
+        guard let image else { return nil }
+
+        if let assetImage = image.assetImage {
+            guard let url = URL(string: assetImage.uri),
+                let data = try? Data(contentsOf: url),
+                let uiImage = Parser.decodeImage(
+                    data: data,
+                    scale: CGFloat(assetImage.scale),
+                    maxDuration: VoiceInputManager.maxVoiceImageCycleDuration
+                )
+            else { return nil }
+
+            if uiImage.images != nil {
+                return Parser.resizeAnimated(uiImage, max: VoiceInputManager.voiceImageMaxSize)
+            }
+
+            if assetImage.color == nil {
+                return Parser.resize(uiImage, max: VoiceInputManager.voiceImageMaxSize)
+            }
+
+            guard let tinted = Parser.parseAssetImage(assetImage: assetImage, traitCollection: traitCollection) else {
+                return nil
+            }
+            return Parser.resize(tinted, max: VoiceInputManager.voiceImageMaxSize)
+        }
+
+        if let glyphImage = image.glyphImage {
+            return
+                SymbolFont
+                .imageFromGlyph(
+                    glyphImage: glyphImage,
+                    size: 150,  // according to docs on CPVoiceControlState.image
+                    foregroundColor: glyphImage.color,
+                    backgroundColor: glyphImage.backgroundColor,
+                    fontScale: glyphImage.fontScale ?? 1.0,
+                    traitCollection: traitCollection
+                )
+        }
+
+        return nil
+    }
+
+    private func presentVoiceTemplate(
+        interfaceController: AutoPlayInterfaceController,
+        listeningText: String,
+        listeningImage: Variant_GlyphImage_AssetImage_RemoteImage?,
+        listeningImageRepeats: Bool?
+    ) {
+        let traitCollection = SceneStore.getRootTraitCollection() ?? UITraitCollection.current
+        let image = loadVoiceImage(
+            image: listeningImage,
+            traitCollection: traitCollection
+        )
+
+        let repeats = listeningImageRepeats ?? (image?.images != nil)
         let listeningState = CPVoiceControlState(
             identifier: "listening",
             titleVariants: [listeningText],
-            image: nil,
-            repeats: true
+            image: image,
+            repeats: repeats
         )
         let template = CPVoiceControlTemplate(voiceControlStates: [listeningState])
         initTemplate(template: template, id: "voice-input")

package/lib/hybrid/HybridVoice.js

@@ -1,8 +1,10 @@
 import { NitroModules } from 'react-native-nitro-modules';
+import { NitroImageUtil } from '../utils/NitroImage';
 const _native = NitroModules.createHybridObject('Voice');
 const startVoiceInput = async (options) => {
-    const { onChunk, silenceThresholdMs, maxDurationMs, listeningText, preferSpeechToText, language, } = options ?? {};
-    return await _native.startVoiceInput(silenceThresholdMs, maxDurationMs, listeningText, preferSpeechToText, onChunk, language);
+    const { onChunk, silenceThresholdMs, maxDurationMs, listeningText, listeningImage, preferSpeechToText, language, } = options ?? {};
+    const listeningImageRepeats = listeningImage?.type === 'asset' ? listeningImage.repeats : undefined;
+    return await _native.startVoiceInput(silenceThresholdMs, maxDurationMs, listeningText, NitroImageUtil.convert(listeningImage), listeningImageRepeats, preferSpeechToText, onChunk, language);
 };
 export const HybridVoice = {
     /**

package/lib/specs/Voice.nitro.d.ts

@@ -1,11 +1,12 @@
 import type { HybridObject } from 'react-native-nitro-modules';
 import type { VoiceInputChunk, VoiceInputResult } from '../types/Voice';
+import type { NitroImage } from '../utils/NitroImage';
 export interface Voice extends HybridObject<{
     android: 'kotlin';
     ios: 'swift';
 }> {
     hasVoiceInputPermission(): boolean;
     requestVoiceInputPermission(): Promise<boolean>;
-    startVoiceInput(silenceThresholdMs?: number, maxDurationMs?: number, listeningText?: string, preferSpeechToText?: boolean, onChunk?: (chunk: VoiceInputChunk) => void, language?: string): Promise<VoiceInputResult>;
+    startVoiceInput(silenceThresholdMs?: number, maxDurationMs?: number, listeningText?: string, listeningImage?: NitroImage, listeningImageRepeats?: boolean, preferSpeechToText?: boolean, onChunk?: (chunk: VoiceInputChunk) => void, language?: string): Promise<VoiceInputResult>;
     stopVoiceInput(): void;
 }

package/lib/templates/Template.d.ts

@@ -4,7 +4,7 @@ import type { NitroMapButton } from '../utils/NitroMapButton';
 export type ActionButton<T> = ActionButtonAndroid<T> | ActionButtonIos<T>;
 export type HeaderActionsIos<T> = {
     /**
-     * @platform iOS - the back button can not be hidden or disabled, if you don't define it iOS will provide a default back button popping the template
+     * @namespace iOS - the back button can not be hidden or disabled, if you don't define it iOS will provide a default back button popping the template
      */
     backButton?: BackButton<T>;
     leadingNavigationBarButtons?: [ActionButtonIos<T>, ActionButtonIos<T>] | [ActionButtonIos<T>];

package/lib/types/Image.d.ts

@@ -66,4 +66,20 @@ export type AutoImage = AutoGlyph | {
     timeoutMs?: number;
     type: 'remote';
 };
+/**
+ * Image for the CarPlay voice control overlay. Animated GIF, APNG, and WebP all animate.
+ * `color` is ignored for animated assets — frames cannot be tinted individually. Static
+ * assets and glyphs are resized/capped to fit within 150×150pt (`fontScale` controls a
+ * glyph's relative size). CarPlay enforces a 0.3s–5s animation cycle duration: shorter
+ * source animations are stretched to 0.3s by the system, longer ones are capped to 5s.
+ * @namespace ios
+ */
+export type VoiceInputImage = AutoGlyph | {
+    type: 'asset';
+    image: ImageSourcePropType;
+    /** Tints the image. Ignored for animated assets — frames cannot be tinted individually. */
+    color?: ThemedColor | string;
+    /** Whether the animation loops. Defaults to `true` for animated images, `false` for static. */
+    repeats?: boolean;
+};
 export {};

package/lib/types/Voice.d.ts

@@ -1,3 +1,4 @@
+import type { VoiceInputImage } from './Image';
 export interface VoiceInputChunk {
     partial?: string;
     audio?: ArrayBuffer;
@@ -10,6 +11,10 @@ export interface VoiceInputOptions {
     silenceThresholdMs?: number;
     maxDurationMs?: number;
     listeningText?: string;
+    /** Image displayed in the CarPlay voice control overlay. See {@link VoiceInputImage}.
+     * @namespace ios
+     */
+    listeningImage?: VoiceInputImage;
     preferSpeechToText?: boolean;
     onChunk?: (chunk: VoiceInputChunk) => void;
     language?: string;

package/nitrogen/generated/android/c++/JHybridVoiceSpec.cpp

@@ -9,6 +9,14 @@
 
 // Forward declaration of `VoiceInputResult` to properly resolve imports.
 namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputResult; }
+// Forward declaration of `GlyphImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct GlyphImage; }
+// Forward declaration of `AssetImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct AssetImage; }
+// Forward declaration of `RemoteImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct RemoteImage; }
+// Forward declaration of `NitroColor` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct NitroColor; }
 // Forward declaration of `VoiceInputChunk` to properly resolve imports.
 namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputChunk; }
 
@@ -20,6 +28,16 @@ namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputC
 #include <optional>
 #include <NitroModules/ArrayBuffer.hpp>
 #include <NitroModules/JArrayBuffer.hpp>
+#include "GlyphImage.hpp"
+#include "AssetImage.hpp"
+#include "RemoteImage.hpp"
+#include <variant>
+#include "JVariant_GlyphImage_AssetImage_RemoteImage.hpp"
+#include "JGlyphImage.hpp"
+#include "NitroColor.hpp"
+#include "JNitroColor.hpp"
+#include "JAssetImage.hpp"
+#include "JRemoteImage.hpp"
 #include "VoiceInputChunk.hpp"
 #include <functional>
 #include "JFunc_void_VoiceInputChunk.hpp"
@@ -80,9 +98,9 @@ namespace margelo::nitro::swe::iternio::reactnativeautoplay {
       return __promise;
     }();
   }
-  std::shared_ptr<Promise<VoiceInputResult>> JHybridVoiceSpec::startVoiceInput(std::optional<double> silenceThresholdMs, std::optional<double> maxDurationMs, const std::optional<std::string>& listeningText, std::optional<bool> preferSpeechToText, const std::optional<std::function<void(const VoiceInputChunk& /* chunk */)>>& onChunk, const std::optional<std::string>& language) {
-    static const auto method = _javaPart->javaClassStatic()->getMethod<jni::local_ref<JPromise::javaobject>(jni::alias_ref<jni::JDouble> /* silenceThresholdMs */, jni::alias_ref<jni::JDouble> /* maxDurationMs */, jni::alias_ref<jni::JString> /* listeningText */, jni::alias_ref<jni::JBoolean> /* preferSpeechToText */, jni::alias_ref<JFunc_void_VoiceInputChunk::javaobject> /* onChunk */, jni::alias_ref<jni::JString> /* language */)>("startVoiceInput_cxx");
-    auto __result = method(_javaPart, silenceThresholdMs.has_value() ? jni::JDouble::valueOf(silenceThresholdMs.value()) : nullptr, maxDurationMs.has_value() ? jni::JDouble::valueOf(maxDurationMs.value()) : nullptr, listeningText.has_value() ? jni::make_jstring(listeningText.value()) : nullptr, preferSpeechToText.has_value() ? jni::JBoolean::valueOf(preferSpeechToText.value()) : nullptr, onChunk.has_value() ? JFunc_void_VoiceInputChunk_cxx::fromCpp(onChunk.value()) : nullptr, language.has_value() ? jni::make_jstring(language.value()) : nullptr);
+  std::shared_ptr<Promise<VoiceInputResult>> JHybridVoiceSpec::startVoiceInput(std::optional<double> silenceThresholdMs, std::optional<double> maxDurationMs, const std::optional<std::string>& listeningText, const std::optional<std::variant<GlyphImage, AssetImage, RemoteImage>>& listeningImage, std::optional<bool> listeningImageRepeats, std::optional<bool> preferSpeechToText, const std::optional<std::function<void(const VoiceInputChunk& /* chunk */)>>& onChunk, const std::optional<std::string>& language) {
+    static const auto method = _javaPart->javaClassStatic()->getMethod<jni::local_ref<JPromise::javaobject>(jni::alias_ref<jni::JDouble> /* silenceThresholdMs */, jni::alias_ref<jni::JDouble> /* maxDurationMs */, jni::alias_ref<jni::JString> /* listeningText */, jni::alias_ref<JVariant_GlyphImage_AssetImage_RemoteImage> /* listeningImage */, jni::alias_ref<jni::JBoolean> /* listeningImageRepeats */, jni::alias_ref<jni::JBoolean> /* preferSpeechToText */, jni::alias_ref<JFunc_void_VoiceInputChunk::javaobject> /* onChunk */, jni::alias_ref<jni::JString> /* language */)>("startVoiceInput_cxx");
+    auto __result = method(_javaPart, silenceThresholdMs.has_value() ? jni::JDouble::valueOf(silenceThresholdMs.value()) : nullptr, maxDurationMs.has_value() ? jni::JDouble::valueOf(maxDurationMs.value()) : nullptr, listeningText.has_value() ? jni::make_jstring(listeningText.value()) : nullptr, listeningImage.has_value() ? JVariant_GlyphImage_AssetImage_RemoteImage::fromCpp(listeningImage.value()) : nullptr, listeningImageRepeats.has_value() ? jni::JBoolean::valueOf(listeningImageRepeats.value()) : nullptr, preferSpeechToText.has_value() ? jni::JBoolean::valueOf(preferSpeechToText.value()) : nullptr, onChunk.has_value() ? JFunc_void_VoiceInputChunk_cxx::fromCpp(onChunk.value()) : nullptr, language.has_value() ? jni::make_jstring(language.value()) : nullptr);
     return [&]() {
       auto __promise = Promise<VoiceInputResult>::create();
       __result->cthis()->addOnResolvedListener([=](const jni::alias_ref<jni::JObject>& __boxedResult) {

package/nitrogen/generated/android/c++/JHybridVoiceSpec.hpp

@@ -56,7 +56,7 @@ namespace margelo::nitro::swe::iternio::reactnativeautoplay {
     // Methods
     bool hasVoiceInputPermission() override;
     std::shared_ptr<Promise<bool>> requestVoiceInputPermission() override;
-    std::shared_ptr<Promise<VoiceInputResult>> startVoiceInput(std::optional<double> silenceThresholdMs, std::optional<double> maxDurationMs, const std::optional<std::string>& listeningText, std::optional<bool> preferSpeechToText, const std::optional<std::function<void(const VoiceInputChunk& /* chunk */)>>& onChunk, const std::optional<std::string>& language) override;
+    std::shared_ptr<Promise<VoiceInputResult>> startVoiceInput(std::optional<double> silenceThresholdMs, std::optional<double> maxDurationMs, const std::optional<std::string>& listeningText, const std::optional<std::variant<GlyphImage, AssetImage, RemoteImage>>& listeningImage, std::optional<bool> listeningImageRepeats, std::optional<bool> preferSpeechToText, const std::optional<std::function<void(const VoiceInputChunk& /* chunk */)>>& onChunk, const std::optional<std::string>& language) override;
     void stopVoiceInput() override;
 
   private:

package/nitrogen/generated/android/kotlin/com/margelo/nitro/swe/iternio/reactnativeautoplay/HybridVoiceSpec.kt

@@ -37,12 +37,12 @@ abstract class HybridVoiceSpec: HybridObject() {
   @Keep
   abstract fun requestVoiceInputPermission(): Promise<Boolean>
   
-  abstract fun startVoiceInput(silenceThresholdMs: Double?, maxDurationMs: Double?, listeningText: String?, preferSpeechToText: Boolean?, onChunk: ((chunk: VoiceInputChunk) -> Unit)?, language: String?): Promise<VoiceInputResult>
+  abstract fun startVoiceInput(silenceThresholdMs: Double?, maxDurationMs: Double?, listeningText: String?, listeningImage: Variant_GlyphImage_AssetImage_RemoteImage?, listeningImageRepeats: Boolean?, preferSpeechToText: Boolean?, onChunk: ((chunk: VoiceInputChunk) -> Unit)?, language: String?): Promise<VoiceInputResult>
   
   @DoNotStrip
   @Keep
-  private fun startVoiceInput_cxx(silenceThresholdMs: Double?, maxDurationMs: Double?, listeningText: String?, preferSpeechToText: Boolean?, onChunk: Func_void_VoiceInputChunk?, language: String?): Promise<VoiceInputResult> {
-    val __result = startVoiceInput(silenceThresholdMs, maxDurationMs, listeningText, preferSpeechToText, onChunk?.let { it }, language)
+  private fun startVoiceInput_cxx(silenceThresholdMs: Double?, maxDurationMs: Double?, listeningText: String?, listeningImage: Variant_GlyphImage_AssetImage_RemoteImage?, listeningImageRepeats: Boolean?, preferSpeechToText: Boolean?, onChunk: Func_void_VoiceInputChunk?, language: String?): Promise<VoiceInputResult> {
+    val __result = startVoiceInput(silenceThresholdMs, maxDurationMs, listeningText, listeningImage, listeningImageRepeats, preferSpeechToText, onChunk?.let { it }, language)
     return __result
   }
   

package/nitrogen/generated/ios/c++/HybridVoiceSpecSwift.hpp

@@ -16,6 +16,14 @@ namespace ReactNativeAutoPlay { class HybridVoiceSpec_cxx; }
 namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputResult; }
 // Forward declaration of `ArrayBufferHolder` to properly resolve imports.
 namespace NitroModules { class ArrayBufferHolder; }
+// Forward declaration of `GlyphImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct GlyphImage; }
+// Forward declaration of `AssetImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct AssetImage; }
+// Forward declaration of `RemoteImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct RemoteImage; }
+// Forward declaration of `NitroColor` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct NitroColor; }
 // Forward declaration of `VoiceInputChunk` to properly resolve imports.
 namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputChunk; }
 
@@ -25,6 +33,11 @@ namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputC
 #include <optional>
 #include <NitroModules/ArrayBuffer.hpp>
 #include <NitroModules/ArrayBufferHolder.hpp>
+#include "GlyphImage.hpp"
+#include "AssetImage.hpp"
+#include "RemoteImage.hpp"
+#include <variant>
+#include "NitroColor.hpp"
 #include "VoiceInputChunk.hpp"
 #include <functional>
 
@@ -94,8 +107,8 @@ namespace margelo::nitro::swe::iternio::reactnativeautoplay {
       auto __value = std::move(__result.value());
       return __value;
     }
-    inline std::shared_ptr<Promise<VoiceInputResult>> startVoiceInput(std::optional<double> silenceThresholdMs, std::optional<double> maxDurationMs, const std::optional<std::string>& listeningText, std::optional<bool> preferSpeechToText, const std::optional<std::function<void(const VoiceInputChunk& /* chunk */)>>& onChunk, const std::optional<std::string>& language) override {
-      auto __result = _swiftPart.startVoiceInput(silenceThresholdMs, maxDurationMs, listeningText, preferSpeechToText, onChunk, language);
+    inline std::shared_ptr<Promise<VoiceInputResult>> startVoiceInput(std::optional<double> silenceThresholdMs, std::optional<double> maxDurationMs, const std::optional<std::string>& listeningText, const std::optional<std::variant<GlyphImage, AssetImage, RemoteImage>>& listeningImage, std::optional<bool> listeningImageRepeats, std::optional<bool> preferSpeechToText, const std::optional<std::function<void(const VoiceInputChunk& /* chunk */)>>& onChunk, const std::optional<std::string>& language) override {
+      auto __result = _swiftPart.startVoiceInput(silenceThresholdMs, maxDurationMs, listeningText, listeningImage, listeningImageRepeats, preferSpeechToText, onChunk, language);
       if (__result.hasError()) [[unlikely]] {
         std::rethrow_exception(__result.error());
       }

package/nitrogen/generated/ios/swift/HybridVoiceSpec.swift

@@ -15,7 +15,7 @@ public protocol HybridVoiceSpec_protocol: HybridObject {
   // Methods
   func hasVoiceInputPermission() throws -> Bool
   func requestVoiceInputPermission() throws -> Promise<Bool>
-  func startVoiceInput(silenceThresholdMs: Double?, maxDurationMs: Double?, listeningText: String?, preferSpeechToText: Bool?, onChunk: ((_ chunk: VoiceInputChunk) -> Void)?, language: String?) throws -> Promise<VoiceInputResult>
+  func startVoiceInput(silenceThresholdMs: Double?, maxDurationMs: Double?, listeningText: String?, listeningImage: Variant_GlyphImage_AssetImage_RemoteImage?, listeningImageRepeats: Bool?, preferSpeechToText: Bool?, onChunk: ((_ chunk: VoiceInputChunk) -> Void)?, language: String?) throws -> Promise<VoiceInputResult>
   func stopVoiceInput() throws -> Void
 }
 

package/nitrogen/generated/ios/swift/HybridVoiceSpec_cxx.swift

@@ -156,7 +156,7 @@ open class HybridVoiceSpec_cxx {
   }
   
   @inline(__always)
-  public final func startVoiceInput(silenceThresholdMs: bridge.std__optional_double_, maxDurationMs: bridge.std__optional_double_, listeningText: bridge.std__optional_std__string_, preferSpeechToText: bridge.std__optional_bool_, onChunk: bridge.std__optional_std__function_void_const_VoiceInputChunk_____chunk______, language: bridge.std__optional_std__string_) -> bridge.Result_std__shared_ptr_Promise_VoiceInputResult___ {
+  public final func startVoiceInput(silenceThresholdMs: bridge.std__optional_double_, maxDurationMs: bridge.std__optional_double_, listeningText: bridge.std__optional_std__string_, listeningImage: bridge.std__optional_std__variant_GlyphImage__AssetImage__RemoteImage__, listeningImageRepeats: bridge.std__optional_bool_, preferSpeechToText: bridge.std__optional_bool_, onChunk: bridge.std__optional_std__function_void_const_VoiceInputChunk_____chunk______, language: bridge.std__optional_std__string_) -> bridge.Result_std__shared_ptr_Promise_VoiceInputResult___ {
     do {
       let __result = try self.__implementation.startVoiceInput(silenceThresholdMs: { () -> Double? in
         if bridge.has_value_std__optional_double_(silenceThresholdMs) {
@@ -179,6 +179,35 @@ open class HybridVoiceSpec_cxx {
         } else {
           return nil
         }
+      }(), listeningImage: { () -> Variant_GlyphImage_AssetImage_RemoteImage? in
+        if bridge.has_value_std__optional_std__variant_GlyphImage__AssetImage__RemoteImage__(listeningImage) {
+          let __unwrapped = bridge.get_std__optional_std__variant_GlyphImage__AssetImage__RemoteImage__(listeningImage)
+          return { () -> Variant_GlyphImage_AssetImage_RemoteImage in
+            let __variant = bridge.std__variant_GlyphImage__AssetImage__RemoteImage_(__unwrapped)
+            switch __variant.index() {
+              case 0:
+                let __actual = __variant.get_0()
+                return .first(__actual)
+              case 1:
+                let __actual = __variant.get_1()
+                return .second(__actual)
+              case 2:
+                let __actual = __variant.get_2()
+                return .third(__actual)
+              default:
+                fatalError("Variant can never have index \(__variant.index())!")
+            }
+          }()
+        } else {
+          return nil
+        }
+      }(), listeningImageRepeats: { () -> Bool? in
+        if bridge.has_value_std__optional_bool_(listeningImageRepeats) {
+          let __unwrapped = bridge.get_std__optional_bool_(listeningImageRepeats)
+          return __unwrapped
+        } else {
+          return nil
+        }
       }(), preferSpeechToText: { () -> Bool? in
         if bridge.has_value_std__optional_bool_(preferSpeechToText) {
           let __unwrapped = bridge.get_std__optional_bool_(preferSpeechToText)

package/nitrogen/generated/shared/c++/HybridVoiceSpec.hpp

@@ -15,6 +15,12 @@
 
 // Forward declaration of `VoiceInputResult` to properly resolve imports.
 namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputResult; }
+// Forward declaration of `GlyphImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct GlyphImage; }
+// Forward declaration of `AssetImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct AssetImage; }
+// Forward declaration of `RemoteImage` to properly resolve imports.
+namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct RemoteImage; }
 // Forward declaration of `VoiceInputChunk` to properly resolve imports.
 namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputChunk; }
 
@@ -22,6 +28,10 @@ namespace margelo::nitro::swe::iternio::reactnativeautoplay { struct VoiceInputC
 #include "VoiceInputResult.hpp"
 #include <optional>
 #include <string>
+#include "GlyphImage.hpp"
+#include "AssetImage.hpp"
+#include "RemoteImage.hpp"
+#include <variant>
 #include "VoiceInputChunk.hpp"
 #include <functional>
 
@@ -58,7 +68,7 @@ namespace margelo::nitro::swe::iternio::reactnativeautoplay {
       // Methods
       virtual bool hasVoiceInputPermission() = 0;
       virtual std::shared_ptr<Promise<bool>> requestVoiceInputPermission() = 0;
-      virtual std::shared_ptr<Promise<VoiceInputResult>> startVoiceInput(std::optional<double> silenceThresholdMs, std::optional<double> maxDurationMs, const std::optional<std::string>& listeningText, std::optional<bool> preferSpeechToText, const std::optional<std::function<void(const VoiceInputChunk& /* chunk */)>>& onChunk, const std::optional<std::string>& language) = 0;
+      virtual std::shared_ptr<Promise<VoiceInputResult>> startVoiceInput(std::optional<double> silenceThresholdMs, std::optional<double> maxDurationMs, const std::optional<std::string>& listeningText, const std::optional<std::variant<GlyphImage, AssetImage, RemoteImage>>& listeningImage, std::optional<bool> listeningImageRepeats, std::optional<bool> preferSpeechToText, const std::optional<std::function<void(const VoiceInputChunk& /* chunk */)>>& onChunk, const std::optional<std::string>& language) = 0;
       virtual void stopVoiceInput() = 0;
 
     protected:

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@iternio/react-native-auto-play",
-  "version": "0.5.3",
+  "version": "0.5.4",
   "description": "Android Auto and Apple CarPlay for react-native",
   "main": "lib/index",
   "module": "lib/index",

package/src/hybrid/HybridVoice.ts

@@ -1,6 +1,7 @@
 import { NitroModules } from 'react-native-nitro-modules';
 import type { Voice } from '../specs/Voice.nitro';
 import type { VoiceInputOptions, VoiceInputResult } from '../types/Voice';
+import { NitroImageUtil } from '../utils/NitroImage';
 
 const _native = NitroModules.createHybridObject<Voice>('Voice');
 
@@ -17,14 +18,20 @@ const startVoiceInput: StartVoiceInput = async (options?: VoiceInputOptions) =>
     silenceThresholdMs,
     maxDurationMs,
     listeningText,
+    listeningImage,
     preferSpeechToText,
     language,
   } = options ?? {};
 
+  const listeningImageRepeats =
+    listeningImage?.type === 'asset' ? listeningImage.repeats : undefined;
+
   return await _native.startVoiceInput(
     silenceThresholdMs,
     maxDurationMs,
     listeningText,
+    NitroImageUtil.convert(listeningImage),
+    listeningImageRepeats,
     preferSpeechToText,
     onChunk,
     language

package/src/specs/Voice.nitro.ts

@@ -1,5 +1,6 @@
 import type { HybridObject } from 'react-native-nitro-modules';
 import type { VoiceInputChunk, VoiceInputResult } from '../types/Voice';
+import type { NitroImage } from '../utils/NitroImage';
 
 export interface Voice extends HybridObject<{ android: 'kotlin'; ios: 'swift' }> {
   hasVoiceInputPermission(): boolean;
@@ -8,6 +9,8 @@ export interface Voice extends HybridObject<{ android: 'kotlin'; ios: 'swift' }>
     silenceThresholdMs?: number,
     maxDurationMs?: number,
     listeningText?: string,
+    listeningImage?: NitroImage,
+    listeningImageRepeats?: boolean,
     preferSpeechToText?: boolean,
     onChunk?: (chunk: VoiceInputChunk) => void,
     language?: string

package/src/templates/Template.ts

@@ -8,7 +8,7 @@ export type ActionButton<T> = ActionButtonAndroid<T> | ActionButtonIos<T>;
 
 export type HeaderActionsIos<T> = {
   /**
-   * @platform iOS - the back button can not be hidden or disabled, if you don't define it iOS will provide a default back button popping the template
+   * @namespace iOS - the back button can not be hidden or disabled, if you don't define it iOS will provide a default back button popping the template
    */
   backButton?: BackButton<T>;
   leadingNavigationBarButtons?: [ActionButtonIos<T>, ActionButtonIos<T>] | [ActionButtonIos<T>];

package/src/types/Image.ts

@@ -79,3 +79,22 @@ export type AutoImage =
       timeoutMs?: number;
       type: 'remote';
     };
+
+/**
+ * Image for the CarPlay voice control overlay. Animated GIF, APNG, and WebP all animate.
+ * `color` is ignored for animated assets — frames cannot be tinted individually. Static
+ * assets and glyphs are resized/capped to fit within 150×150pt (`fontScale` controls a
+ * glyph's relative size). CarPlay enforces a 0.3s–5s animation cycle duration: shorter
+ * source animations are stretched to 0.3s by the system, longer ones are capped to 5s.
+ * @namespace ios
+ */
+export type VoiceInputImage =
+  | AutoGlyph
+  | {
+      type: 'asset';
+      image: ImageSourcePropType;
+      /** Tints the image. Ignored for animated assets — frames cannot be tinted individually. */
+      color?: ThemedColor | string;
+      /** Whether the animation loops. Defaults to `true` for animated images, `false` for static. */
+      repeats?: boolean;
+    };

package/src/types/Voice.ts

@@ -1,3 +1,5 @@
+import type { VoiceInputImage } from './Image';
+
 export interface VoiceInputChunk {
   partial?: string;
   audio?: ArrayBuffer;
@@ -12,6 +14,10 @@ export interface VoiceInputOptions {
   silenceThresholdMs?: number;
   maxDurationMs?: number;
   listeningText?: string;
+  /** Image displayed in the CarPlay voice control overlay. See {@link VoiceInputImage}.
+   * @namespace ios
+   */
+  listeningImage?: VoiceInputImage;
   preferSpeechToText?: boolean;
   onChunk?: (chunk: VoiceInputChunk) => void;
   language?: string;