添加音色克隆

test/src/main/AndroidManifest.xml

@@ -26,6 +26,12 @@
             </intent-filter>
         </activity>
 
+        <activity
+            android:name=".ui.activity.VoiceCloneSetupActivity"
+            android:exported="false"
+            android:theme="@style/Theme.DUIX.Test">
+        </activity>
+
         <activity
             android:name=".ui.activity.CallActivity"
             android:exported="false"

test/src/main/java/ai/guiji/duix/test/audio/VoiceCloneRecorder.kt

@@ -0,0 +1,63 @@
+package ai.guiji.duix.test.audio
+
+import ai.guiji.duix.test.service.DashScopeConfig
+import android.media.MediaRecorder
+import android.os.SystemClock
+import java.io.File
+
+class VoiceCloneRecorder {
+
+    private var mediaRecorder: MediaRecorder? = null
+    private var recordingFile: File? = null
+    private var startAtMs: Long = 0L
+
+    fun start(outputFile: File) {
+        release()
+        outputFile.parentFile?.mkdirs()
+        mediaRecorder = MediaRecorder().apply {
+            setAudioSource(MediaRecorder.AudioSource.MIC)
+            setOutputFormat(MediaRecorder.OutputFormat.MPEG_4)
+            setAudioEncoder(MediaRecorder.AudioEncoder.AAC)
+            setAudioChannels(1)
+            setAudioSamplingRate(DashScopeConfig.AUDIO_SAMPLE_RATE)
+            setAudioEncodingBitRate(96000)
+            setOutputFile(outputFile.absolutePath)
+            prepare()
+            start()
+        }
+        recordingFile = outputFile
+        startAtMs = SystemClock.elapsedRealtime()
+    }
+
+    fun stop(): VoiceRecording {
+        val file = recordingFile ?: throw IllegalStateException("还没有开始录音")
+        val recorder = mediaRecorder ?: throw IllegalStateException("录音器不可用")
+        val duration = (SystemClock.elapsedRealtime() - startAtMs).coerceAtLeast(0L)
+        try {
+            recorder.stop()
+        } catch (e: RuntimeException) {
+            file.delete()
+            throw IllegalStateException("录音太短，请再试一次", e)
+        } finally {
+            release()
+        }
+        return VoiceRecording(file, duration)
+    }
+
+    fun release() {
+        mediaRecorder?.runCatching {
+            reset()
+        }
+        mediaRecorder?.runCatching {
+            release()
+        }
+        mediaRecorder = null
+        recordingFile = null
+        startAtMs = 0L
+    }
+
+    data class VoiceRecording(
+        val file: File,
+        val durationMs: Long
+    )
+}

test/src/main/java/ai/guiji/duix/test/model/AvatarCatalog.kt

@@ -0,0 +1,20 @@
+package ai.guiji.duix.test.model
+
+import ai.guiji.duix.test.R
+
+object AvatarCatalog {
+
+    val models = listOf(
+        AvatarModel("晨煦", "Ethan", "https://dl.xn--876a.net/models/bendi3_20240518.zip", R.mipmap.ethan),
+        AvatarModel("詹妮弗", "Maia", "https://dl.xn--876a.net/models/Emma.zip", R.mipmap.jennifer),
+        AvatarModel("月白", "Moon", "https://dl.xn--876a.net/models/Kai.zip", R.mipmap.moon),
+        AvatarModel("艾登", "Aiden", "https://dl.xn--876a.net/models/Leo.zip", R.mipmap.leo),
+        AvatarModel("四月", "Serena", "https://dl.xn--876a.net/models/Lily.zip", R.mipmap.lily),
+        AvatarModel("奥利弗", "Andre", "https://dl.xn--876a.net/models/Oliver.zip", R.mipmap.oliver),
+        AvatarModel("索菲亚", "Sonrisa", "https://dl.xn--876a.net/models/Sofia.zip", R.mipmap.sofia),
+    )
+
+    fun findByModelUrl(modelUrl: String?): AvatarModel? {
+        return models.firstOrNull { it.modelUrl == modelUrl }
+    }
+}

test/src/main/java/ai/guiji/duix/test/model/VoiceCloneProfile.kt

@@ -0,0 +1,22 @@
+package ai.guiji.duix.test.model
+
+data class VoiceCloneProfile(
+    val voice: String,
+    val targetModel: String,
+    val preferredName: String,
+    val samplePath: String,
+    val avatarName: String,
+    val avatarVoice: String,
+    val avatarModelUrl: String,
+    val avatarRes: Int,
+    val createdAt: Long
+) {
+    fun toAvatarModel(): AvatarModel {
+        return AvatarModel(
+            name = avatarName,
+            voice = avatarVoice,
+            modelUrl = avatarModelUrl,
+            avatarRes = avatarRes
+        )
+    }
+}

test/src/main/java/ai/guiji/duix/test/realtime/OmniRealtimeClient.java

@@ -1,5 +1,7 @@
 package ai.guiji.duix.test.realtime;
 
+import ai.guiji.duix.test.service.DashScopeConfig;
+
 import android.util.Base64;
 import android.util.Log;
 
@@ -24,9 +26,9 @@ import okhttp3.WebSocketListener;
 public class OmniRealtimeClient {
 
     private static final String TAG = "OmniRealtimeClient";
-    private static final String DEFAULT_API_KEY = "sk-7c32083ace99472e97209e611b9ecbb1";
+    private static final String DEFAULT_API_KEY = DashScopeConfig.DEFAULT_API_KEY;
     public static final String DEFAULT_MODEL = "qwen3.5-omni-plus-realtime";
-    private static final String DEFAULT_BASE_URL = "wss://dashscope.aliyuncs.com/api-ws/v1/realtime";
+    private static final String DEFAULT_BASE_URL = DashScopeConfig.REALTIME_BASE_URL;
     private static final String INPUT_AUDIO_TRANSCRIPTION_MODEL = "gummy-realtime-v1";
     private static final int MAX_IMAGE_BYTES = 500 * 1024;
     private static final long MIN_IMAGE_APPEND_INTERVAL_MS = 800L;
@@ -66,6 +68,8 @@ public class OmniRealtimeClient {
     private final AtomicBoolean imageWindowOpen = new AtomicBoolean(false);
     private final AtomicBoolean hasSentAnyAudio = new AtomicBoolean(false);
     private final AtomicLong lastImageAppendAtMs = new AtomicLong(0L);
+    private final AtomicBoolean assistantTextFinalized = new AtomicBoolean(false);
+    private final StringBuilder assistantTextBuffer = new StringBuilder();
 
     public OmniRealtimeClient(SessionOptions sessionOptions, Callback callback) {
         this.sessionOptions = sessionOptions;
@@ -155,10 +159,16 @@ public class OmniRealtimeClient {
         }
         try {
             JSONObject session = new JSONObject();
-            session.put("modalities", new JSONArray().put("text").put("audio"));
+            JSONArray modalities = new JSONArray().put("text");
-            session.put("voice", options.getVoice());
+            if (options.isAudioOutputEnabled()) {
+                modalities.put("audio");
+            }
+            session.put("modalities", modalities);
+            if (options.isAudioOutputEnabled()) {
+                session.put("voice", options.getVoice());
+                session.put("output_audio_format", "pcm");
+            }
             session.put("input_audio_format", "pcm");
-            session.put("output_audio_format", "pcm");
             session.put("instructions", options.getInstructions());
             session.put("enable_search", options.isEnableSearch());
 
@@ -226,6 +236,7 @@ public class OmniRealtimeClient {
                 case "response.created":
                     assistantResponding.set(true);
                     imageWindowOpen.set(false);
+                    resetAssistantTextState();
                     break;
 
                 case "conversation.item.input_audio_transcription.completed":
@@ -237,18 +248,23 @@ public class OmniRealtimeClient {
                     break;
 
                 case "response.audio_transcript.delta":
-                    String textDelta = event.optString("delta", "");
+                    handleAssistantTextDelta(event.optString("delta", ""));
-                    if (callback != null && !textDelta.isEmpty()) {
-                        callback.onAssistantTextDelta(textDelta);
-                    }
                     break;
 
                 case "response.audio_transcript.done":
-                    String fullTranscript = event.optString("transcript", "");
+                    dispatchAssistantFinalText(event.optString("transcript", event.optString("text", "")));
-                    Log.d(TAG, "Assistant said: " + fullTranscript);
+                    break;
-                    if (callback != null) {
+
-                        callback.onAssistantTranscriptDone(fullTranscript);
+                case "response.text.delta":
-                    }
+                    handleAssistantTextDelta(event.optString("delta", ""));
+                    break;
+
+                case "response.text.done":
+                    dispatchAssistantFinalText(event.optString("text", event.optString("transcript", "")));
+                    break;
+
+                case "response.output_item.done":
+                    dispatchAssistantFinalTextIfNeeded(extractTextFromItem(event.optJSONObject("item")));
                     break;
 
                 case "response.audio.delta":
@@ -270,6 +286,7 @@ public class OmniRealtimeClient {
                     Log.d(TAG, "Response done");
                     assistantResponding.set(false);
                     imageWindowOpen.set(false);
+                    dispatchAssistantFinalTextIfNeeded(extractTextFromResponse(event.optJSONObject("response")));
                     if (callback != null) {
                         callback.onResponseDone();
                     }
@@ -401,6 +418,95 @@ public class OmniRealtimeClient {
         imageWindowOpen.set(false);
         hasSentAnyAudio.set(false);
         lastImageAppendAtMs.set(0L);
+        resetAssistantTextState();
+    }
+
+    private void handleAssistantTextDelta(String delta) {
+        if (delta == null || delta.isEmpty()) {
+            return;
+        }
+        synchronized (assistantTextBuffer) {
+            assistantTextBuffer.append(delta);
+        }
+        if (callback != null) {
+            callback.onAssistantTextDelta(delta);
+        }
+    }
+
+    private void dispatchAssistantFinalTextIfNeeded(String text) {
+        if (assistantTextFinalized.get()) {
+            return;
+        }
+        dispatchAssistantFinalText(text);
+    }
+
+    private void dispatchAssistantFinalText(String text) {
+        String finalText = text;
+        if (finalText == null || finalText.isEmpty()) {
+            synchronized (assistantTextBuffer) {
+                finalText = assistantTextBuffer.toString();
+            }
+        }
+        if (finalText == null || finalText.isEmpty()) {
+            return;
+        }
+        synchronized (assistantTextBuffer) {
+            assistantTextBuffer.setLength(0);
+            assistantTextBuffer.append(finalText);
+        }
+        assistantTextFinalized.set(true);
+        Log.d(TAG, "Assistant said: " + finalText);
+        if (callback != null) {
+            callback.onAssistantTranscriptDone(finalText);
+        }
+    }
+
+    private void resetAssistantTextState() {
+        synchronized (assistantTextBuffer) {
+            assistantTextBuffer.setLength(0);
+        }
+        assistantTextFinalized.set(false);
+    }
+
+    private String extractTextFromResponse(JSONObject response) {
+        if (response == null) {
+            return "";
+        }
+        StringBuilder builder = new StringBuilder();
+        JSONArray output = response.optJSONArray("output");
+        if (output == null) {
+            return "";
+        }
+        for (int i = 0; i < output.length(); i++) {
+            JSONObject item = output.optJSONObject(i);
+            String itemText = extractTextFromItem(item);
+            if (!itemText.isEmpty()) {
+                builder.append(itemText);
+            }
+        }
+        return builder.toString();
+    }
+
+    private String extractTextFromItem(JSONObject item) {
+        if (item == null) {
+            return "";
+        }
+        StringBuilder builder = new StringBuilder();
+        JSONArray content = item.optJSONArray("content");
+        if (content == null) {
+            return "";
+        }
+        for (int i = 0; i < content.length(); i++) {
+            JSONObject part = content.optJSONObject(i);
+            if (part == null) {
+                continue;
+            }
+            String text = part.optString("text", "");
+            if (!text.isEmpty()) {
+                builder.append(text);
+            }
+        }
+        return builder.toString();
     }
 
     private String nextEventId() {
@@ -408,6 +514,11 @@ public class OmniRealtimeClient {
     }
 
     public static class SessionOptions {
+        public enum OutputMode {
+            AUDIO_AND_TEXT,
+            TEXT_ONLY
+        }
+
         private final String apiKey;
         private final String baseUrl;
         private final String model;
@@ -416,15 +527,20 @@ public class OmniRealtimeClient {
         private final boolean enableSearch;
         private final boolean enableTurnDetection;
         private final boolean enableInputAudioTranscription;
+        private final OutputMode outputMode;
 
         public SessionOptions(String voice, boolean enableSearch) {
+            this(voice, enableSearch, OutputMode.AUDIO_AND_TEXT);
+        }
+
+        public SessionOptions(String voice, boolean enableSearch, OutputMode outputMode) {
             this(DEFAULT_API_KEY, DEFAULT_BASE_URL, DEFAULT_MODEL, voice, DEFAULT_INSTRUCTIONS,
-                    enableSearch, true, true);
+                    enableSearch, true, true, outputMode);
         }
 
         public SessionOptions(String apiKey, String baseUrl, String model, String voice,
                 String instructions, boolean enableSearch, boolean enableTurnDetection,
-                boolean enableInputAudioTranscription) {
+                boolean enableInputAudioTranscription, OutputMode outputMode) {
             this.apiKey = apiKey == null ? "" : apiKey;
             this.baseUrl = baseUrl == null || baseUrl.trim().isEmpty() ? DEFAULT_BASE_URL : baseUrl;
             this.model = model == null || model.trim().isEmpty() ? DEFAULT_MODEL : model;
@@ -434,6 +550,7 @@ public class OmniRealtimeClient {
             this.enableSearch = enableSearch;
             this.enableTurnDetection = enableTurnDetection;
             this.enableInputAudioTranscription = enableInputAudioTranscription;
+            this.outputMode = outputMode == null ? OutputMode.AUDIO_AND_TEXT : outputMode;
         }
 
         public String getApiKey() {
@@ -463,6 +580,10 @@ public class OmniRealtimeClient {
         public boolean isEnableInputAudioTranscription() {
             return enableInputAudioTranscription;
         }
+
+        public boolean isAudioOutputEnabled() {
+            return outputMode == OutputMode.AUDIO_AND_TEXT;
+        }
     }
 
     public interface Callback {

test/src/main/java/ai/guiji/duix/test/realtime/QwenTtsRealtimeClient.kt

@@ -0,0 +1,154 @@
+package ai.guiji.duix.test.realtime
+
+import ai.guiji.duix.test.service.DashScopeConfig
+import android.util.Base64
+import okhttp3.OkHttpClient
+import okhttp3.Request
+import okhttp3.Response
+import okhttp3.WebSocket
+import okhttp3.WebSocketListener
+import org.json.JSONObject
+import java.util.UUID
+import java.util.concurrent.TimeUnit
+import java.util.concurrent.atomic.AtomicBoolean
+
+class QwenTtsRealtimeClient(
+    private val sessionOptions: SessionOptions,
+    private val callback: Callback
+) {
+
+    private val client = OkHttpClient.Builder()
+        .connectTimeout(30, TimeUnit.SECONDS)
+        .readTimeout(120, TimeUnit.MINUTES)
+        .writeTimeout(30, TimeUnit.SECONDS)
+        .pingInterval(30, TimeUnit.SECONDS)
+        .build()
+
+    private val connected = AtomicBoolean(false)
+    private var webSocket: WebSocket? = null
+
+    fun connect() {
+        if (connected.get()) {
+            return
+        }
+        val request = Request.Builder()
+            .url(sessionOptions.baseUrl + "?model=" + sessionOptions.model)
+            .addHeader("Authorization", "Bearer " + sessionOptions.apiKey)
+            .build()
+
+        webSocket = client.newWebSocket(request, object : WebSocketListener() {
+            override fun onOpen(webSocket: WebSocket, response: Response) {
+                connected.set(true)
+                sendSessionUpdate()
+                callback.onConnected()
+            }
+
+            override fun onMessage(webSocket: WebSocket, text: String) {
+                handleMessage(text)
+            }
+
+            override fun onFailure(webSocket: WebSocket, t: Throwable, response: Response?) {
+                connected.set(false)
+                callback.onError(t.message ?: "连接失败")
+                callback.onDisconnected()
+            }
+
+            override fun onClosed(webSocket: WebSocket, code: Int, reason: String) {
+                connected.set(false)
+                callback.onDisconnected()
+            }
+        })
+    }
+
+    fun appendText(text: String) {
+        if (text.isBlank()) {
+            return
+        }
+        val event = JSONObject()
+        event.put("event_id", nextEventId())
+        event.put("type", "input_text_buffer.append")
+        event.put("text", text)
+        sendEvent(event)
+    }
+
+    fun finish() {
+        val event = JSONObject()
+        event.put("event_id", nextEventId())
+        event.put("type", "session.finish")
+        sendEvent(event)
+    }
+
+    fun disconnect() {
+        webSocket?.close(1000, "bye")
+        webSocket = null
+        connected.set(false)
+    }
+
+    private fun sendSessionUpdate() {
+        val session = JSONObject()
+        session.put("voice", sessionOptions.voice)
+        session.put("mode", "server_commit")
+        session.put("response_format", "pcm")
+        session.put("sample_rate", DashScopeConfig.AUDIO_SAMPLE_RATE)
+
+        val event = JSONObject()
+        event.put("event_id", nextEventId())
+        event.put("type", "session.update")
+        event.put("session", session)
+        sendEvent(event)
+    }
+
+    private fun handleMessage(text: String) {
+        try {
+            val event = JSONObject(text)
+            when (event.optString("type")) {
+                "response.audio.delta" -> {
+                    val audioBase64 = event.optString("delta")
+                    if (audioBase64.isNotBlank()) {
+                        callback.onAudioDelta(Base64.decode(audioBase64, Base64.NO_WRAP))
+                    }
+                }
+                "response.audio.done" -> callback.onAudioDone()
+                "response.done" -> callback.onResponseDone()
+                "session.finished" -> callback.onSessionFinished()
+                "error" -> {
+                    val error = event.optJSONObject("error")
+                    callback.onError(error?.optString("message") ?: "语音合成失败")
+                }
+            }
+        } catch (e: Exception) {
+            callback.onError(e.message ?: "语音合成失败")
+        }
+    }
+
+    private fun sendEvent(event: JSONObject): Boolean {
+        return connected.get() && webSocket?.send(event.toString()) == true
+    }
+
+    private fun nextEventId(): String {
+        return "event_" + UUID.randomUUID().toString().replace("-", "")
+    }
+
+    data class SessionOptions(
+        val voice: String,
+        val model: String = DashScopeConfig.VOICE_CLONE_TTS_MODEL,
+        val apiKey: String = DashScopeConfig.DEFAULT_API_KEY,
+        val baseUrl: String = DashScopeConfig.REALTIME_BASE_URL
+    )
+
+    interface Callback {
+        fun onConnected()
+
+        fun onDisconnected()
+
+        fun onAudioDelta(pcmData: ByteArray)
+
+        fun onAudioDone()
+
+        fun onResponseDone()
+
+        fun onSessionFinished()
+
+        fun onError(message: String)
+    }
+}

test/src/main/java/ai/guiji/duix/test/service/DashScopeConfig.java

@@ -0,0 +1,14 @@
+package ai.guiji.duix.test.service;
+
+public final class DashScopeConfig {
+
+    public static final String DEFAULT_API_KEY = "sk-7c32083ace99472e97209e611b9ecbb1";
+    public static final String REALTIME_BASE_URL = "wss://dashscope.aliyuncs.com/api-ws/v1/realtime";
+    public static final String VOICE_CLONE_URL = "https://dashscope.aliyuncs.com/api/v1/services/audio/tts/customization";
+    public static final String VOICE_CLONE_MODEL = "qwen-voice-enrollment";
+    public static final String VOICE_CLONE_TTS_MODEL = "qwen3-tts-vc-realtime-2026-01-15";
+    public static final int AUDIO_SAMPLE_RATE = 24000;
+
+    private DashScopeConfig() {
+    }
+}

test/src/main/java/ai/guiji/duix/test/service/VoiceCloneProfileStore.kt

@@ -0,0 +1,53 @@
+package ai.guiji.duix.test.service
+
+import ai.guiji.duix.test.model.VoiceCloneProfile
+import android.content.Context
+import org.json.JSONObject
+
+class VoiceCloneProfileStore(context: Context) {
+
+    private val prefs = context.getSharedPreferences(PREFS_NAME, Context.MODE_PRIVATE)
+
+    fun load(): VoiceCloneProfile? {
+        val raw = prefs.getString(KEY_PROFILE, null) ?: return null
+        return try {
+            val json = JSONObject(raw)
+            VoiceCloneProfile(
+                voice = json.optString("voice"),
+                targetModel = json.optString("targetModel"),
+                preferredName = json.optString("preferredName"),
+                samplePath = json.optString("samplePath"),
+                avatarName = json.optString("avatarName"),
+                avatarVoice = json.optString("avatarVoice"),
+                avatarModelUrl = json.optString("avatarModelUrl"),
+                avatarRes = json.optInt("avatarRes"),
+                createdAt = json.optLong("createdAt")
+            )
+        } catch (_: Exception) {
+            null
+        }
+    }
+
+    fun save(profile: VoiceCloneProfile) {
+        val json = JSONObject()
+        json.put("voice", profile.voice)
+        json.put("targetModel", profile.targetModel)
+        json.put("preferredName", profile.preferredName)
+        json.put("samplePath", profile.samplePath)
+        json.put("avatarName", profile.avatarName)
+        json.put("avatarVoice", profile.avatarVoice)
+        json.put("avatarModelUrl", profile.avatarModelUrl)
+        json.put("avatarRes", profile.avatarRes)
+        json.put("createdAt", profile.createdAt)
+        prefs.edit().putString(KEY_PROFILE, json.toString()).apply()
+    }
+
+    fun clear() {
+        prefs.edit().remove(KEY_PROFILE).apply()
+    }
+
+    companion object {
+        private const val PREFS_NAME = "voice_clone_profile"
+        private const val KEY_PROFILE = "active_profile"
+    }
+}

test/src/main/java/ai/guiji/duix/test/service/VoiceCloneService.kt

@@ -0,0 +1,151 @@
+package ai.guiji.duix.test.service
+
+import android.util.Base64
+import okhttp3.Call
+import okhttp3.Callback
+import okhttp3.MediaType.Companion.toMediaType
+import okhttp3.OkHttpClient
+import okhttp3.Request
+import okhttp3.RequestBody.Companion.toRequestBody
+import okhttp3.Response
+import org.json.JSONObject
+import java.io.File
+import java.io.IOException
+import java.util.Locale
+import java.util.concurrent.TimeUnit
+
+class VoiceCloneService {
+
+    private val client = OkHttpClient.Builder()
+        .connectTimeout(30, TimeUnit.SECONDS)
+        .readTimeout(60, TimeUnit.SECONDS)
+        .writeTimeout(60, TimeUnit.SECONDS)
+        .build()
+
+    fun createVoice(
+        sampleFile: File,
+        preferredName: String,
+        targetModel: String = DashScopeConfig.VOICE_CLONE_TTS_MODEL,
+        callback: (Result<VoiceCloneResult>) -> Unit
+    ) {
+        if (!sampleFile.exists()) {
+            callback(Result.failure(IllegalStateException("录音文件不存在")))
+            return
+        }
+        val payload = JSONObject()
+        val input = JSONObject()
+        val audio = JSONObject()
+        audio.put("data", buildDataUri(sampleFile))
+        input.put("action", "create")
+        input.put("target_model", targetModel)
+        input.put("preferred_name", normalizePreferredName(preferredName))
+        input.put("audio", audio)
+        payload.put("model", DashScopeConfig.VOICE_CLONE_MODEL)
+        payload.put("input", input)
+
+        val request = Request.Builder()
+            .url(DashScopeConfig.VOICE_CLONE_URL)
+            .addHeader("Authorization", "Bearer " + DashScopeConfig.DEFAULT_API_KEY)
+            .addHeader("Content-Type", "application/json")
+            .post(payload.toString().toRequestBody(JSON_MEDIA_TYPE))
+            .build()
+
+        client.newCall(request).enqueue(object : Callback {
+            override fun onFailure(call: Call, e: IOException) {
+                callback(Result.failure(e))
+            }
+
+            override fun onResponse(call: Call, response: Response) {
+                response.use {
+                    val body = response.body?.string().orEmpty()
+                    if (!response.isSuccessful) {
+                        callback(Result.failure(IllegalStateException(parseErrorMessage(body).ifEmpty { "保存熟悉声音失败" })))
+                        return
+                    }
+                    try {
+                        val json = JSONObject(body)
+                        val output = json.getJSONObject("output")
+                        val voice = output.optString("voice")
+                        if (voice.isBlank()) {
+                            callback(Result.failure(IllegalStateException("没有拿到可用音色")))
+                            return
+                        }
+                        callback(
+                            Result.success(
+                                VoiceCloneResult(
+                                    voice = voice,
+                                    targetModel = output.optString("target_model", targetModel)
+                                )
+                            )
+                        )
+                    } catch (e: Exception) {
+                        callback(Result.failure(e))
+                    }
+                }
+            }
+        })
+    }
+
+    fun deleteVoice(voice: String) {
+        if (voice.isBlank()) {
+            return
+        }
+        val payload = JSONObject()
+        val input = JSONObject()
+        input.put("action", "delete")
+        input.put("voice", voice)
+        payload.put("model", DashScopeConfig.VOICE_CLONE_MODEL)
+        payload.put("input", input)
+
+        val request = Request.Builder()
+            .url(DashScopeConfig.VOICE_CLONE_URL)
+            .addHeader("Authorization", "Bearer " + DashScopeConfig.DEFAULT_API_KEY)
+            .addHeader("Content-Type", "application/json")
+            .post(payload.toString().toRequestBody(JSON_MEDIA_TYPE))
+            .build()
+
+        client.newCall(request).enqueue(object : Callback {
+            override fun onFailure(call: Call, e: IOException) {
+            }
+
+            override fun onResponse(call: Call, response: Response) {
+                response.close()
+            }
+        })
+    }
+
+    private fun buildDataUri(file: File): String {
+        val mimeType = when (file.extension.lowercase(Locale.US)) {
+            "wav" -> "audio/wav"
+            "mp3" -> "audio/mpeg"
+            else -> "audio/mp4"
+        }
+        val encoded = Base64.encodeToString(file.readBytes(), Base64.NO_WRAP)
+        return "data:$mimeType;base64,$encoded"
+    }
+
+    private fun normalizePreferredName(source: String): String {
+        val normalized = source.lowercase(Locale.US)
+            .replace("[^a-z0-9_]".toRegex(), "_")
+            .trim('_')
+            .ifBlank { "family_voice" }
+        return normalized.take(16)
+    }
+
+    private fun parseErrorMessage(body: String): String {
+        return try {
+            JSONObject(body).optJSONObject("error")?.optString("message").orEmpty()
+        } catch (_: Exception) {
+            ""
+        }
+    }
+
+    data class VoiceCloneResult(
+        val voice: String,
+        val targetModel: String
+    )
+
+    companion object {
+        private val JSON_MEDIA_TYPE = "application/json; charset=utf-8".toMediaType()
+    }
+}

test/src/main/java/ai/guiji/duix/test/ui/activity/CallActivity.kt

@@ -11,6 +11,8 @@ import ai.guiji.duix.test.audio.StreamingAudioRecorder
 import ai.guiji.duix.test.camera.CameraFrameCapture
 import ai.guiji.duix.test.databinding.ActivityCallBinding
 import ai.guiji.duix.test.realtime.OmniRealtimeClient
+import ai.guiji.duix.test.realtime.QwenTtsRealtimeClient
+import ai.guiji.duix.test.service.DashScopeConfig
 import ai.guiji.duix.test.ui.adapter.MotionAdapter
 import ai.guiji.duix.test.ui.dialog.AudioRecordDialog
 import ai.guiji.duix.test.ui.widget.AudioSpectrumView
@@ -73,6 +75,10 @@ class CallActivity : BaseActivity() {
     private var assistantTextBuffer = StringBuilder()
     private var hasStartedPushForResponse = false
     private val enableWebSearch = true
+    private var isCloneMode = false
+    private var clonedVoice = ""
+    private var cloneTtsModel = DashScopeConfig.VOICE_CLONE_TTS_MODEL
+    private var ttsRealtimeClient: QwenTtsRealtimeClient? = null
 
     override fun onCreate(savedInstanceState: Bundle?) {
         super.onCreate(savedInstanceState)
@@ -86,6 +92,9 @@ class CallActivity : BaseActivity() {
         modelUrl = intent.getStringExtra("modelUrl") ?: ""
         voice = intent.getStringExtra("voice") ?: "Ethan"
         debug = intent.getBooleanExtra("debug", false)
+        isCloneMode = intent.getBooleanExtra("cloneMode", false)
+        clonedVoice = intent.getStringExtra("clonedVoice") ?: ""
+        cloneTtsModel = intent.getStringExtra("cloneTtsModel") ?: DashScopeConfig.VOICE_CLONE_TTS_MODEL
 
         Glide.with(mContext).load("file:///android_asset/bg/bg1.png").into(binding.ivBg)
 
@@ -249,7 +258,12 @@ class CallActivity : BaseActivity() {
     }
 
     private fun buildRealtimeSessionOptions(): OmniRealtimeClient.SessionOptions {
-        return OmniRealtimeClient.SessionOptions(voice, enableWebSearch)
+        val outputMode = if (isCloneMode) {
+            OmniRealtimeClient.SessionOptions.OutputMode.TEXT_ONLY
+        } else {
+            OmniRealtimeClient.SessionOptions.OutputMode.AUDIO_AND_TEXT
+        }
+        return OmniRealtimeClient.SessionOptions(voice, enableWebSearch, outputMode)
     }
 
     private fun isImageAppendError(message: String): Boolean {
@@ -334,6 +348,10 @@ class CallActivity : BaseActivity() {
             Toast.makeText(mContext, "数字人未就绪", Toast.LENGTH_SHORT).show()
             return
         }
+        if (isCloneMode && clonedVoice.isBlank()) {
+            Toast.makeText(mContext, R.string.voice_clone_unavailable, Toast.LENGTH_SHORT).show()
+            return
+        }
         applyMessage("AI对话: 连接中...")
         binding.tvSubtitle.visibility = View.VISIBLE
         binding.tvSubtitle.text = getString(R.string.ai_call_connecting)
@@ -348,11 +366,11 @@ class CallActivity : BaseActivity() {
             override fun onConnected() {
                 mainHandler.post {
                     binding.tvSubtitle.text = getString(R.string.ai_call_connected)
-                    binding.btnCameraToggle.visibility = View.VISIBLE
+                    binding.btnCameraToggle.visibility = if (isCloneMode) View.GONE else View.VISIBLE
                     binding.btnCameraToggle.setBackgroundResource(R.drawable.shape_fab_camera_off)
                     binding.btnCameraToggle.contentDescription = getString(R.string.camera_off)
                     isCameraEnabled = false
-                    applyMessage("AI对话: 已连接, model=${OmniRealtimeClient.DEFAULT_MODEL}, search=on")
+                    applyMessage("AI对话: 已连接, model=${OmniRealtimeClient.DEFAULT_MODEL}, search=on, cloneMode=$isCloneMode")
                     streamingRecorder = StreamingAudioRecorder(mContext, object : StreamingAudioRecorder.StreamingCallback {
                         override fun onAudioData(pcmData: ByteArray) {
                             omniClient?.appendAudio(pcmData)
@@ -385,6 +403,9 @@ class CallActivity : BaseActivity() {
             }
 
             override fun onUserSpeechStarted() {
+                if (isCloneMode) {
+                    stopCloneSpeech()
+                }
                 cameraFrameCapture?.requestImmediateCapture()
                 mainHandler.post {
                     binding.tvSubtitle.text = getString(R.string.ai_call_listening)
@@ -419,6 +440,9 @@ class CallActivity : BaseActivity() {
             }
 
             override fun onAssistantAudioDelta(pcmData: ByteArray) {
+                if (isCloneMode) {
+                    return
+                }
                 if (!hasStartedPushForResponse) {
                     hasStartedPushForResponse = true
                     duix?.startPush()
@@ -430,6 +454,9 @@ class CallActivity : BaseActivity() {
             }
 
             override fun onAssistantAudioDone() {
+                if (isCloneMode) {
+                    return
+                }
                 if (hasStartedPushForResponse) {
                     duix?.stopPush()
                     hasStartedPushForResponse = false
@@ -438,7 +465,12 @@ class CallActivity : BaseActivity() {
 
             override fun onResponseDone() {
                 mainHandler.post {
-                    if (hasStartedPushForResponse) {
+                    if (isCloneMode) {
+                        val replyText = assistantTextBuffer.toString().trim()
+                        if (replyText.isNotEmpty()) {
+                            speakWithClonedVoice(replyText)
+                        }
+                    } else if (hasStartedPushForResponse) {
                         duix?.stopPush()
                         hasStartedPushForResponse = false
                     }
@@ -454,6 +486,7 @@ class CallActivity : BaseActivity() {
     private fun stopAIConversation() {
         isAIConversationActive = false
         stopCamera()
+        stopCloneSpeech()
         streamingRecorder?.release()
         streamingRecorder = null
         omniClient?.disconnect()
@@ -476,6 +509,80 @@ class CallActivity : BaseActivity() {
         }
     }
 
+    private fun speakWithClonedVoice(text: String) {
+        if (!isAIConversationActive || clonedVoice.isBlank()) {
+            return
+        }
+        stopCloneSpeech()
+        ttsRealtimeClient = QwenTtsRealtimeClient(
+            QwenTtsRealtimeClient.SessionOptions(
+                voice = clonedVoice,
+                model = cloneTtsModel
+            ),
+            object : QwenTtsRealtimeClient.Callback {
+                override fun onConnected() {
+                    ttsRealtimeClient?.appendText(text)
+                    ttsRealtimeClient?.finish()
+                }
+
+                override fun onDisconnected() {
+                    ttsRealtimeClient = null
+                }
+
+                override fun onAudioDelta(pcmData: ByteArray) {
+                    if (!hasStartedPushForResponse) {
+                        hasStartedPushForResponse = true
+                        duix?.startPush()
+                    }
+                    val resampled = AudioResampler.resample24kTo16k(pcmData)
+                    if (resampled.isNotEmpty()) {
+                        duix?.pushPcm(resampled)
+                    }
+                }
+
+                override fun onAudioDone() {
+                    if (hasStartedPushForResponse) {
+                        duix?.stopPush()
+                        hasStartedPushForResponse = false
+                    }
+                }
+
+                override fun onResponseDone() {
+                }
+
+                override fun onSessionFinished() {
+                    if (hasStartedPushForResponse) {
+                        duix?.stopPush()
+                        hasStartedPushForResponse = false
+                    }
+                    ttsRealtimeClient = null
+                }
+
+                override fun onError(message: String) {
+                    mainHandler.post {
+                        if (hasStartedPushForResponse) {
+                            duix?.stopPush()
+                            hasStartedPushForResponse = false
+                        }
+                        Toast.makeText(mContext, toFriendlyErrorMessage(message), Toast.LENGTH_SHORT).show()
+                        applyMessage("复刻声音错误: $message")
+                    }
+                }
+            }
+        )
+        ttsRealtimeClient?.connect()
+    }
+
+    private fun stopCloneSpeech() {
+        ttsRealtimeClient?.disconnect()
+        ttsRealtimeClient = null
+        if (hasStartedPushForResponse) {
+            duix?.stopPush()
+            hasStartedPushForResponse = false
+        }
+        duix?.stopAudio()
+    }
+
     private fun toggleCamera() {
         if (!isAIConversationActive || omniClient == null) return
         isCameraEnabled = !isCameraEnabled

test/src/main/java/ai/guiji/duix/test/ui/activity/MainActivity.kt

@@ -3,6 +3,7 @@ package ai.guiji.duix.test.ui.activity
 import ai.guiji.duix.sdk.client.VirtualModelUtil
 import ai.guiji.duix.test.R
 import ai.guiji.duix.test.databinding.ActivityMainBinding
+import ai.guiji.duix.test.model.AvatarCatalog
 import ai.guiji.duix.test.model.AvatarModel
 import ai.guiji.duix.test.ui.adapter.AvatarAdapter
 import ai.guiji.duix.test.ui.dialog.LoadingDialog
@@ -21,16 +22,6 @@ class MainActivity : BaseActivity() {
 
     private val baseConfigUrl = "https://dl.xn--876a.net/models/gj_dh_res.zip"
 
-    private val avatarModels = listOf(
-        AvatarModel("晨煦", "Ethan", "https://dl.xn--876a.net/models/bendi3_20240518.zip", R.mipmap.ethan),
-        AvatarModel("詹妮弗", "Maia", "https://dl.xn--876a.net/models/Emma.zip", R.mipmap.jennifer),
-        AvatarModel("月白", "Moon", "https://dl.xn--876a.net/models/Kai.zip", R.mipmap.moon),
-        AvatarModel("艾登", "Aiden", "https://dl.xn--876a.net/models/Leo.zip", R.mipmap.leo),
-        AvatarModel("四月", "Serena", "https://dl.xn--876a.net/models/Lily.zip", R.mipmap.lily),
-        AvatarModel("奥利弗", "Andre", "https://dl.xn--876a.net/models/Oliver.zip", R.mipmap.oliver),
-        AvatarModel("索菲亚", "Sonrisa", "https://dl.xn--876a.net/models/Sofia.zip", R.mipmap.sofia),
-    )
-
     private var mSelectedModel: AvatarModel? = null
 
     override fun onCreate(savedInstanceState: Bundle?) {
@@ -38,8 +29,12 @@ class MainActivity : BaseActivity() {
         binding = ActivityMainBinding.inflate(layoutInflater)
         setContentView(binding.root)
 
+        binding.btnVoiceClone.setOnClickListener {
+            startActivity(Intent(mContext, VoiceCloneSetupActivity::class.java))
+        }
+
         binding.rvAvatars.layoutManager = GridLayoutManager(this, 2)
-        binding.rvAvatars.adapter = AvatarAdapter(avatarModels, object : AvatarAdapter.Callback {
+        binding.rvAvatars.adapter = AvatarAdapter(AvatarCatalog.models, object : AvatarAdapter.Callback {
             override fun onSelect(model: AvatarModel) {
                 mSelectedModel = model
                 play(model)

test/src/main/java/ai/guiji/duix/test/ui/activity/VoiceCloneSetupActivity.kt

@@ -0,0 +1,276 @@
+package ai.guiji.duix.test.ui.activity
+
+import ai.guiji.duix.test.R
+import ai.guiji.duix.test.audio.VoiceCloneRecorder
+import ai.guiji.duix.test.databinding.ActivityVoiceCloneSetupBinding
+import ai.guiji.duix.test.model.AvatarCatalog
+import ai.guiji.duix.test.model.AvatarModel
+import ai.guiji.duix.test.model.VoiceCloneProfile
+import ai.guiji.duix.test.service.DashScopeConfig
+import ai.guiji.duix.test.service.VoiceCloneProfileStore
+import ai.guiji.duix.test.service.VoiceCloneService
+import ai.guiji.duix.test.ui.adapter.AvatarAdapter
+import android.Manifest
+import android.content.Intent
+import android.os.Bundle
+import android.os.Handler
+import android.os.Looper
+import android.os.SystemClock
+import android.view.View
+import android.widget.Toast
+import androidx.recyclerview.widget.GridLayoutManager
+import java.io.File
+import java.util.Locale
+
+class VoiceCloneSetupActivity : BaseActivity() {
+
+    private lateinit var binding: ActivityVoiceCloneSetupBinding
+    private lateinit var avatarAdapter: AvatarAdapter
+    private val mainHandler = Handler(Looper.getMainLooper())
+    private val profileStore by lazy { VoiceCloneProfileStore(this) }
+    private val voiceCloneService = VoiceCloneService()
+    private val recorder = VoiceCloneRecorder()
+
+    private var selectedAvatar: AvatarModel = AvatarCatalog.models.first()
+    private var activeProfile: VoiceCloneProfile? = null
+    private var currentRecording: VoiceCloneRecorder.VoiceRecording? = null
+    private var isRecording = false
+    private var isCreatingVoice = false
+    private var recordStartedAtMs = 0L
+
+    private val recordTicker = object : Runnable {
+        override fun run() {
+            if (!isRecording) {
+                return
+            }
+            val elapsed = (SystemClock.elapsedRealtime() - recordStartedAtMs).coerceAtLeast(0L)
+            binding.tvRecordTime.text = getString(R.string.voice_clone_record_time, formatDuration(elapsed))
+            if (elapsed >= MAX_RECORDING_MS) {
+                stopRecording(true)
+                return
+            }
+            mainHandler.postDelayed(this, 500L)
+        }
+    }
+
+    override fun onCreate(savedInstanceState: Bundle?) {
+        super.onCreate(savedInstanceState)
+        keepScreenOn()
+        binding = ActivityVoiceCloneSetupBinding.inflate(layoutInflater)
+        setContentView(binding.root)
+
+        activeProfile = profileStore.load()
+        selectedAvatar = activeProfile?.toAvatarModel() ?: AvatarCatalog.models.first()
+
+        avatarAdapter = AvatarAdapter(
+            AvatarCatalog.models,
+            object : AvatarAdapter.Callback {
+                override fun onSelect(model: AvatarModel) {
+                    selectedAvatar = model
+                    avatarAdapter.setSelectedModel(model)
+                    updateUi()
+                }
+            },
+            enableSelection = true
+        )
+
+        binding.rvAvatars.layoutManager = GridLayoutManager(this, 2)
+        binding.rvAvatars.adapter = avatarAdapter
+        avatarAdapter.setSelectedModel(selectedAvatar)
+
+        binding.btnPrimaryAction.setOnClickListener {
+            when {
+                isCreatingVoice -> Unit
+                isRecording -> stopRecording(false)
+                currentRecording != null -> createVoiceFromSample()
+                activeProfile != null -> startVoiceCloneConversation()
+                else -> requestPermission(arrayOf(Manifest.permission.RECORD_AUDIO), REQUEST_RECORD_AUDIO)
+            }
+        }
+
+        binding.btnSecondaryAction.setOnClickListener {
+            if (isCreatingVoice || isRecording) {
+                return@setOnClickListener
+            }
+            if (currentRecording != null) {
+                currentRecording?.file?.delete()
+                currentRecording = null
+                updateUi()
+            } else {
+                requestPermission(arrayOf(Manifest.permission.RECORD_AUDIO), REQUEST_RECORD_AUDIO)
+            }
+        }
+
+        binding.tvRecordScript.text = getString(R.string.voice_clone_record_script)
+        binding.tvRecordTime.text = getString(R.string.voice_clone_record_time, formatDuration(0L))
+        updateUi()
+    }
+
+    override fun onDestroy() {
+        super.onDestroy()
+        mainHandler.removeCallbacks(recordTicker)
+        recorder.release()
+    }
+
+    override fun permissionsGet(get: Boolean, code: Int) {
+        super.permissionsGet(get, code)
+        if (code == REQUEST_RECORD_AUDIO) {
+            if (get) {
+                startRecording()
+            } else {
+                Toast.makeText(mContext, R.string.need_permission_continue, Toast.LENGTH_SHORT).show()
+            }
+        }
+    }
+
+    private fun startRecording() {
+        val outputFile = File(filesDir, "voice_clone/sample_${System.currentTimeMillis()}.m4a")
+        try {
+            recorder.start(outputFile)
+            currentRecording = null
+            isRecording = true
+            recordStartedAtMs = SystemClock.elapsedRealtime()
+            binding.tvRecordTime.text = getString(R.string.voice_clone_record_time, formatDuration(0L))
+            mainHandler.removeCallbacks(recordTicker)
+            mainHandler.post(recordTicker)
+            updateUi()
+        } catch (e: Exception) {
+            Toast.makeText(mContext, e.message ?: getString(R.string.voice_clone_record_failed), Toast.LENGTH_SHORT).show()
+            recorder.release()
+            isRecording = false
+            updateUi()
+        }
+    }
+
+    private fun stopRecording(autoStopped: Boolean) {
+        if (!isRecording) {
+            return
+        }
+        isRecording = false
+        mainHandler.removeCallbacks(recordTicker)
+        try {
+            val recording = recorder.stop()
+            if (recording.durationMs < MIN_RECORDING_MS) {
+                recording.file.delete()
+                currentRecording = null
+                Toast.makeText(mContext, R.string.voice_clone_record_too_short, Toast.LENGTH_SHORT).show()
+            } else {
+                currentRecording = recording
+                if (autoStopped) {
+                    Toast.makeText(mContext, R.string.voice_clone_record_auto_stopped, Toast.LENGTH_SHORT).show()
+                } else if (recording.durationMs < RECOMMENDED_RECORDING_MS) {
+                    Toast.makeText(mContext, R.string.voice_clone_record_recommendation, Toast.LENGTH_SHORT).show()
+                }
+            }
+        } catch (e: Exception) {
+            currentRecording = null
+            Toast.makeText(mContext, e.message ?: getString(R.string.voice_clone_record_failed), Toast.LENGTH_SHORT).show()
+        }
+        updateUi()
+    }
+
+    private fun createVoiceFromSample() {
+        val recording = currentRecording ?: return
+        val oldVoice = activeProfile?.voice
+        isCreatingVoice = true
+        updateUi()
+        voiceCloneService.createVoice(
+            sampleFile = recording.file,
+            preferredName = "family_${selectedAvatar.voice}",
+            targetModel = DashScopeConfig.VOICE_CLONE_TTS_MODEL
+        ) { result ->
+            runOnUiThread {
+                isCreatingVoice = false
+                result.onSuccess { cloneResult ->
+                    val profile = VoiceCloneProfile(
+                        voice = cloneResult.voice,
+                        targetModel = cloneResult.targetModel,
+                        preferredName = "family_${selectedAvatar.voice}",
+                        samplePath = recording.file.absolutePath,
+                        avatarName = selectedAvatar.name,
+                        avatarVoice = selectedAvatar.voice,
+                        avatarModelUrl = selectedAvatar.modelUrl,
+                        avatarRes = selectedAvatar.avatarRes,
+                        createdAt = System.currentTimeMillis()
+                    )
+                    profileStore.save(profile)
+                    activeProfile = profile
+                    currentRecording = null
+                    if (!oldVoice.isNullOrBlank() && oldVoice != profile.voice) {
+                        voiceCloneService.deleteVoice(oldVoice)
+                    }
+                    Toast.makeText(mContext, R.string.voice_clone_saved, Toast.LENGTH_SHORT).show()
+                    updateUi()
+                }.onFailure { throwable ->
+                    Toast.makeText(mContext, throwable.message ?: getString(R.string.voice_clone_create_failed), Toast.LENGTH_SHORT).show()
+                    updateUi()
+                }
+            }
+        }
+    }
+
+    private fun startVoiceCloneConversation() {
+        val profile = (activeProfile ?: return).copy(
+            avatarName = selectedAvatar.name,
+            avatarVoice = selectedAvatar.voice,
+            avatarModelUrl = selectedAvatar.modelUrl,
+            avatarRes = selectedAvatar.avatarRes
+        )
+        profileStore.save(profile)
+        activeProfile = profile
+
+        val intent = Intent(mContext, CallActivity::class.java)
+        intent.putExtra("modelUrl", selectedAvatar.modelUrl)
+        intent.putExtra("voice", selectedAvatar.voice)
+        intent.putExtra("debug", false)
+        intent.putExtra("cloneMode", true)
+        intent.putExtra("clonedVoice", profile.voice)
+        intent.putExtra("cloneTtsModel", profile.targetModel)
+        startActivity(intent)
+    }
+
+    private fun updateUi() {
+        val hasReadyProfile = activeProfile != null
+        val hasSample = currentRecording != null
+
+        binding.tvStatus.text = when {
+            isCreatingVoice -> getString(R.string.voice_clone_status_creating)
+            isRecording -> getString(R.string.voice_clone_status_recording, selectedAvatar.name)
+            hasSample && hasReadyProfile -> getString(R.string.voice_clone_status_sample_replace, selectedAvatar.name)
+            hasSample -> getString(R.string.voice_clone_status_sample_ready, selectedAvatar.name)
+            hasReadyProfile -> getString(R.string.voice_clone_status_ready, selectedAvatar.name)
+            else -> getString(R.string.voice_clone_status_idle, selectedAvatar.name)
+        }
+
+        binding.btnPrimaryAction.text = when {
+            isCreatingVoice -> getString(R.string.voice_clone_primary_creating)
+            isRecording -> getString(R.string.voice_clone_primary_stop_record)
+            hasSample -> getString(R.string.voice_clone_primary_create)
+            hasReadyProfile -> getString(R.string.voice_clone_primary_start_chat)
+            else -> getString(R.string.voice_clone_primary_start_record)
+        }
+
+        binding.btnPrimaryAction.isEnabled = !isCreatingVoice
+        binding.btnSecondaryAction.visibility = if (isCreatingVoice || isRecording) View.INVISIBLE else View.VISIBLE
+        binding.btnSecondaryAction.text = if (hasSample) {
+            getString(R.string.voice_clone_secondary_cancel_sample)
+        } else {
+            getString(R.string.voice_clone_secondary_rerecord)
+        }
+        binding.tvRecordHint.text = getString(R.string.voice_clone_record_hint)
+    }
+
+    private fun formatDuration(durationMs: Long): String {
+        val totalSeconds = (durationMs / 1000L).toInt()
+        val minutes = totalSeconds / 60
+        val seconds = totalSeconds % 60
+        return String.format(Locale.getDefault(), "%02d:%02d", minutes, seconds)
+    }
+
+    companion object {
+        private const val REQUEST_RECORD_AUDIO = 11
+        private const val MIN_RECORDING_MS = 3_000L
+        private const val RECOMMENDED_RECORDING_MS = 10_000L
+        private const val MAX_RECORDING_MS = 60_000L
+    }
+}

test/src/main/java/ai/guiji/duix/test/ui/adapter/AvatarAdapter.kt

@@ -4,13 +4,17 @@ import ai.guiji.duix.test.databinding.ItemAvatarCardBinding
 import ai.guiji.duix.test.model.AvatarModel
 import android.view.LayoutInflater
 import android.view.ViewGroup
+import androidx.core.content.ContextCompat
 import androidx.recyclerview.widget.RecyclerView
 
 class AvatarAdapter(
     private val list: List<AvatarModel>,
-    private val callback: Callback
+    private val callback: Callback,
+    private val enableSelection: Boolean = false
 ) : RecyclerView.Adapter<AvatarAdapter.ViewHolder>() {
 
+    private var selectedModelUrl: String? = null
+
     class ViewHolder(val binding: ItemAvatarCardBinding) :
         RecyclerView.ViewHolder(binding.root)
 
@@ -25,13 +29,32 @@ class AvatarAdapter(
         val model = list[position]
         holder.binding.ivAvatar.setImageResource(model.avatarRes)
         holder.binding.tvName.text = model.name
+        val selected = enableSelection && model.modelUrl == selectedModelUrl
+        holder.binding.cardAvatar.strokeWidth = if (selected) 3 else 0
+        holder.binding.cardAvatar.cardElevation = if (selected) 8f else 3f
+        holder.binding.cardAvatar.alpha = if (selected) 1f else 0.96f
+        holder.binding.cardAvatar.setStrokeColor(
+            if (selected) {
+                ContextCompat.getColor(holder.binding.root.context, ai.guiji.duix.test.R.color.warm_accent)
+            } else {
+                ContextCompat.getColor(holder.binding.root.context, android.R.color.transparent)
+            }
+        )
         holder.binding.cardAvatar.setOnClickListener {
+            if (enableSelection) {
+                setSelectedModel(model)
+            }
             callback.onSelect(model)
         }
     }
 
     override fun getItemCount(): Int = list.size
 
+    fun setSelectedModel(model: AvatarModel?) {
+        selectedModelUrl = model?.modelUrl
+        notifyDataSetChanged()
+    }
+
     interface Callback {
         fun onSelect(model: AvatarModel)
     }

test/src/main/res/color/text_warm_primary_button.xml

@@ -0,0 +1,5 @@
+<?xml version="1.0" encoding="utf-8"?>
+<selector xmlns:android="http://schemas.android.com/apk/res/android">
+    <item android:color="#FFF7EDE6" android:state_enabled="false" />
+    <item android:color="@android:color/white" />
+</selector>

test/src/main/res/color/text_warm_secondary_button.xml

@@ -0,0 +1,5 @@
+<?xml version="1.0" encoding="utf-8"?>
+<selector xmlns:android="http://schemas.android.com/apk/res/android">
+    <item android:color="@color/warm_text_hint" android:state_enabled="false" />
+    <item android:color="@color/warm_text_primary" />
+</selector>

test/src/main/res/drawable/selector_warm_primary_button.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="utf-8"?>
+<selector xmlns:android="http://schemas.android.com/apk/res/android">
+    <item android:drawable="@drawable/shape_warm_primary_button_disabled" android:state_enabled="false" />
+    <item android:drawable="@drawable/shape_warm_primary_button_pressed" android:state_pressed="true" />
+    <item android:drawable="@drawable/shape_warm_primary_button" />
+</selector>

test/src/main/res/drawable/selector_warm_secondary_button.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="utf-8"?>
+<selector xmlns:android="http://schemas.android.com/apk/res/android">
+    <item android:drawable="@drawable/shape_warm_secondary_button_disabled" android:state_enabled="false" />
+    <item android:drawable="@drawable/shape_warm_secondary_button_pressed" android:state_pressed="true" />
+    <item android:drawable="@drawable/shape_warm_secondary_button" />
+</selector>

test/src/main/res/drawable/shape_warm_feature_card.xml

@@ -0,0 +1,7 @@
+<shape xmlns:android="http://schemas.android.com/apk/res/android">
+    <corners android:radius="22dp" />
+    <solid android:color="@color/warm_surface_alt" />
+    <stroke
+        android:width="1dp"
+        android:color="#FFDAB89E" />
+</shape>

test/src/main/res/drawable/shape_warm_primary_button.xml

@@ -0,0 +1,4 @@
+<shape xmlns:android="http://schemas.android.com/apk/res/android">
+    <corners android:radius="28dp" />
+    <solid android:color="@color/warm_button_primary" />
+</shape>

test/src/main/res/drawable/shape_warm_primary_button_disabled.xml

@@ -0,0 +1,4 @@
+<shape xmlns:android="http://schemas.android.com/apk/res/android">
+    <corners android:radius="28dp" />
+    <solid android:color="@color/warm_button_primary_disabled" />
+</shape>

test/src/main/res/drawable/shape_warm_primary_button_pressed.xml

@@ -0,0 +1,4 @@
+<shape xmlns:android="http://schemas.android.com/apk/res/android">
+    <corners android:radius="28dp" />
+    <solid android:color="@color/warm_button_primary_pressed" />
+</shape>

test/src/main/res/drawable/shape_warm_secondary_button.xml

@@ -0,0 +1,7 @@
+<shape xmlns:android="http://schemas.android.com/apk/res/android">
+    <corners android:radius="28dp" />
+    <solid android:color="@color/warm_button_secondary_bg" />
+    <stroke
+        android:width="1dp"
+        android:color="@color/warm_button_secondary_stroke" />
+</shape>

test/src/main/res/drawable/shape_warm_secondary_button_disabled.xml

@@ -0,0 +1,7 @@
+<shape xmlns:android="http://schemas.android.com/apk/res/android">
+    <corners android:radius="28dp" />
+    <solid android:color="@color/warm_button_secondary_bg_disabled" />
+    <stroke
+        android:width="1dp"
+        android:color="#FFE6D3C7" />
+</shape>

test/src/main/res/drawable/shape_warm_secondary_button_pressed.xml

@@ -0,0 +1,7 @@
+<shape xmlns:android="http://schemas.android.com/apk/res/android">
+    <corners android:radius="28dp" />
+    <solid android:color="@color/warm_button_secondary_bg_pressed" />
+    <stroke
+        android:width="1dp"
+        android:color="@color/warm_button_secondary_stroke" />
+</shape>

test/src/main/res/layout/activity_main.xml

@@ -1,58 +1,118 @@
 <?xml version="1.0" encoding="utf-8"?>
-<androidx.constraintlayout.widget.ConstraintLayout xmlns:android="http://schemas.android.com/apk/res/android"
+<androidx.core.widget.NestedScrollView xmlns:android="http://schemas.android.com/apk/res/android"
-    xmlns:app="http://schemas.android.com/apk/res-auto"
     android:layout_width="match_parent"
     android:layout_height="match_parent"
-    android:fitsSystemWindows="true"
+    android:background="@color/warm_bg"
-    android:background="@color/warm_bg">
+    android:fillViewport="true">
 
-    <!-- 顶部标题区域 -->
+    <LinearLayout
-    <TextView
-        android:id="@+id/tv_title"
         android:layout_width="match_parent"
         android:layout_height="wrap_content"
-        android:gravity="center"
+        android:orientation="vertical"
-        android:paddingTop="48dp"
+        android:paddingTop="40dp"
-        android:paddingBottom="4dp"
+        android:paddingBottom="20dp">
-        android:text="@string/app_title"
-        android:textColor="@color/warm_text_primary"
-        android:textSize="22sp"
-        android:textStyle="bold"
-        app:layout_constraintTop_toTopOf="parent" />
 
-    <TextView
+        <TextView
-        android:id="@+id/tv_subtitle"
+            android:id="@+id/tv_title"
-        android:layout_width="match_parent"
+            android:layout_width="match_parent"
-        android:layout_height="wrap_content"
+            android:layout_height="wrap_content"
-        android:gravity="center"
+            android:gravity="center"
-        android:paddingBottom="16dp"
+            android:paddingStart="24dp"
-        android:text="@string/app_subtitle"
+            android:paddingEnd="24dp"
-        android:textColor="@color/warm_text_secondary"
+            android:text="@string/app_title"
-        android:textSize="14sp"
+            android:textColor="@color/warm_text_primary"
-        app:layout_constraintTop_toBottomOf="@+id/tv_title" />
+            android:textSize="24sp"
+            android:textStyle="bold" />
 
-    <!-- 角色列表 -->
+        <TextView
-    <androidx.recyclerview.widget.RecyclerView
+            android:id="@+id/tv_subtitle"
-        android:id="@+id/rvAvatars"
+            android:layout_width="match_parent"
-        android:layout_width="match_parent"
+            android:layout_height="wrap_content"
-        android:layout_height="0dp"
+            android:layout_marginTop="8dp"
-        android:clipToPadding="false"
+            android:gravity="center"
-        android:paddingStart="10dp"
+            android:paddingStart="28dp"
-        android:paddingEnd="10dp"
+            android:paddingEnd="28dp"
-        android:paddingBottom="16dp"
+            android:text="@string/app_subtitle"
-        app:layout_constraintBottom_toTopOf="@+id/tvBottomHint"
+            android:textColor="@color/warm_text_secondary"
-        app:layout_constraintTop_toBottomOf="@+id/tv_subtitle" />
+            android:textSize="15sp" />
 
-    <TextView
+        <LinearLayout
-        android:id="@+id/tvBottomHint"
+            android:id="@+id/cardVoiceCloneEntry"
-        android:layout_width="match_parent"
+            android:layout_width="match_parent"
-        android:layout_height="wrap_content"
+            android:layout_height="wrap_content"
-        android:gravity="center"
+            android:layout_marginStart="16dp"
-        android:paddingTop="8dp"
+            android:layout_marginTop="18dp"
-        android:paddingBottom="16dp"
+            android:layout_marginEnd="16dp"
-        android:text="@string/choose_companion_hint"
+            android:background="@drawable/shape_warm_feature_card"
-        android:textColor="@color/warm_text_hint"
+            android:orientation="vertical"
-        android:textSize="12sp"
+            android:padding="18dp">
-        app:layout_constraintBottom_toBottomOf="parent" />
 
-</androidx.constraintlayout.widget.ConstraintLayout>
+            <TextView
+                android:layout_width="wrap_content"
+                android:layout_height="wrap_content"
+                android:text="@string/voice_clone_entry_title"
+                android:textColor="@color/warm_text_primary"
+                android:textSize="19sp"
+                android:textStyle="bold" />
+
+            <TextView
+                android:layout_width="match_parent"
+                android:layout_height="wrap_content"
+                android:layout_marginTop="8dp"
+                android:lineSpacingExtra="4dp"
+                android:text="@string/voice_clone_entry_desc"
+                android:textColor="@color/warm_text_secondary"
+                android:textSize="14sp" />
+
+            <androidx.appcompat.widget.AppCompatButton
+                android:id="@+id/btnVoiceClone"
+                android:layout_width="match_parent"
+                android:layout_height="50dp"
+                android:layout_marginTop="16dp"
+                android:background="@drawable/selector_warm_primary_button"
+                android:elevation="0dp"
+                android:stateListAnimator="@null"
+                android:text="@string/voice_clone_entry_action"
+                android:textAllCaps="false"
+                android:textColor="@color/text_warm_primary_button"
+                android:textSize="17sp"
+                android:textStyle="bold" />
+
+        </LinearLayout>
+
+        <TextView
+            android:layout_width="match_parent"
+            android:layout_height="wrap_content"
+            android:layout_marginTop="20dp"
+            android:paddingStart="18dp"
+            android:paddingEnd="18dp"
+            android:text="@string/choose_companion_hint"
+            android:textColor="@color/warm_text_secondary"
+            android:textSize="15sp" />
+
+        <androidx.recyclerview.widget.RecyclerView
+            android:id="@+id/rvAvatars"
+            android:layout_width="match_parent"
+            android:layout_height="wrap_content"
+            android:clipToPadding="false"
+            android:nestedScrollingEnabled="false"
+            android:paddingStart="10dp"
+            android:paddingTop="8dp"
+            android:paddingEnd="10dp"
+            android:paddingBottom="12dp" />
+
+        <TextView
+            android:id="@+id/tvBottomHint"
+            android:layout_width="match_parent"
+            android:layout_height="wrap_content"
+            android:gravity="center"
+            android:paddingStart="24dp"
+            android:paddingTop="6dp"
+            android:paddingEnd="24dp"
+            android:paddingBottom="10dp"
+            android:text="@string/choose_companion_hint"
+            android:textColor="@color/warm_text_hint"
+            android:textSize="12sp" />
+
+    </LinearLayout>
+</androidx.core.widget.NestedScrollView>

test/src/main/res/layout/activity_voice_clone_setup.xml

@@ -0,0 +1,127 @@
+<?xml version="1.0" encoding="utf-8"?>
+<androidx.core.widget.NestedScrollView xmlns:android="http://schemas.android.com/apk/res/android"
+    android:layout_width="match_parent"
+    android:layout_height="match_parent"
+    android:background="@color/warm_bg"
+    android:clipToPadding="false"
+    android:fillViewport="false">
+
+    <LinearLayout
+        android:layout_width="match_parent"
+        android:layout_height="wrap_content"
+        android:orientation="vertical"
+        android:paddingStart="18dp"
+        android:paddingTop="32dp"
+        android:paddingEnd="18dp"
+        android:paddingBottom="28dp">
+
+    <TextView
+        android:layout_width="wrap_content"
+        android:layout_height="wrap_content"
+        android:text="@string/voice_clone_title"
+        android:textColor="@color/warm_text_primary"
+        android:textSize="25sp"
+        android:textStyle="bold" />
+
+    <TextView
+        android:layout_width="wrap_content"
+        android:layout_height="wrap_content"
+        android:layout_marginTop="8dp"
+        android:text="@string/voice_clone_subtitle"
+        android:textColor="@color/warm_text_secondary"
+        android:textSize="15sp" />
+
+    <TextView
+        android:id="@+id/tvStatus"
+        android:layout_width="match_parent"
+        android:layout_height="wrap_content"
+        android:layout_marginTop="18dp"
+        android:background="@drawable/shape_warm_feature_card"
+        android:lineSpacingExtra="4dp"
+        android:padding="18dp"
+        android:textColor="@color/warm_text_primary"
+        android:textSize="18sp" />
+
+        <TextView
+            android:layout_width="wrap_content"
+            android:layout_height="wrap_content"
+            android:layout_marginTop="20dp"
+            android:text="@string/voice_clone_choose_avatar"
+            android:textColor="@color/warm_text_primary"
+            android:textSize="18sp"
+            android:textStyle="bold" />
+
+        <androidx.recyclerview.widget.RecyclerView
+            android:id="@+id/rvAvatars"
+            android:layout_width="match_parent"
+            android:layout_height="wrap_content"
+            android:layout_marginTop="10dp"
+            android:clipToPadding="false"
+            android:nestedScrollingEnabled="false"
+            android:paddingBottom="12dp" />
+
+    <TextView
+        android:layout_width="wrap_content"
+        android:layout_height="wrap_content"
+        android:layout_marginTop="8dp"
+        android:text="@string/voice_clone_record_title"
+        android:textColor="@color/warm_text_primary"
+        android:textSize="18sp"
+        android:textStyle="bold" />
+
+    <TextView
+        android:id="@+id/tvRecordHint"
+        android:layout_width="match_parent"
+        android:layout_height="wrap_content"
+        android:layout_marginTop="10dp"
+        android:textColor="@color/warm_text_secondary"
+        android:textSize="15sp" />
+
+    <TextView
+        android:id="@+id/tvRecordTime"
+        android:layout_width="wrap_content"
+        android:layout_height="wrap_content"
+        android:layout_gravity="center_horizontal"
+        android:layout_marginTop="12dp"
+        android:textColor="@color/warm_button_primary"
+        android:textSize="24sp"
+        android:textStyle="bold" />
+
+    <TextView
+        android:id="@+id/tvRecordScript"
+        android:layout_width="match_parent"
+        android:layout_height="wrap_content"
+        android:layout_marginTop="14dp"
+        android:background="@drawable/shape_warm_feature_card"
+        android:lineSpacingExtra="5dp"
+        android:padding="18dp"
+        android:textColor="@color/warm_text_primary"
+        android:textSize="16sp" />
+
+        <androidx.appcompat.widget.AppCompatButton
+            android:id="@+id/btnPrimaryAction"
+            android:layout_width="match_parent"
+            android:layout_height="56dp"
+            android:layout_marginTop="20dp"
+            android:background="@drawable/selector_warm_primary_button"
+            android:elevation="0dp"
+            android:stateListAnimator="@null"
+            android:textAllCaps="false"
+            android:textColor="@color/text_warm_primary_button"
+            android:textSize="18sp"
+            android:textStyle="bold" />
+
+        <androidx.appcompat.widget.AppCompatButton
+            android:id="@+id/btnSecondaryAction"
+            android:layout_width="match_parent"
+            android:layout_height="54dp"
+            android:layout_marginTop="12dp"
+            android:background="@drawable/selector_warm_secondary_button"
+            android:elevation="0dp"
+            android:stateListAnimator="@null"
+            android:textAllCaps="false"
+            android:textColor="@color/text_warm_secondary_button"
+            android:textSize="17sp" />
+
+    </LinearLayout>
+</androidx.core.widget.NestedScrollView>

test/src/main/res/layout/item_avatar_card.xml

@@ -4,18 +4,19 @@
     android:id="@+id/cardAvatar"
     android:layout_width="match_parent"
     android:layout_height="wrap_content"
-    android:layout_margin="6dp"
+    android:layout_margin="8dp"
     android:clickable="true"
     android:focusable="true"
     android:foreground="?android:attr/selectableItemBackground"
-    app:cardCornerRadius="16dp"
+    app:cardCornerRadius="18dp"
-    app:cardElevation="2dp"
+    app:cardElevation="3dp"
+    app:strokeColor="@android:color/transparent"
     app:strokeWidth="0dp">
 
     <androidx.constraintlayout.widget.ConstraintLayout
         android:layout_width="match_parent"
         android:layout_height="wrap_content"
-        android:background="@color/white">
+        android:background="@color/warm_surface">
 
         <ImageView
             android:id="@+id/ivAvatar"
@@ -32,8 +33,8 @@
             android:gravity="center"
             android:paddingTop="10dp"
             android:paddingBottom="12dp"
-            android:textColor="#333333"
+            android:textColor="@color/warm_text_primary"
-            android:textSize="15sp"
+            android:textSize="16sp"
             android:textStyle="bold"
             app:layout_constraintTop_toBottomOf="@+id/ivAvatar" />
 

test/src/main/res/values-zh-rCN/strings.xml

@@ -30,6 +30,36 @@
     <string name="ai_call_user_said">您刚才说：%1$s</string>
     <string name="ai_call_camera_paused">镜头先休息一下，您继续说话就可以</string>
     <string name="ai_call_error_retry">刚刚有点忙，我们稍后再试一次</string>
+    <string name="voice_clone_entry_title">熟悉声音陪伴</string>
+    <string name="voice_clone_entry_desc">先录一段熟悉的声音，保存好以后，陪伴者会尽量用这份声音和您说话。</string>
+    <string name="voice_clone_entry_action">去录熟悉声音</string>
+    <string name="voice_clone_title">熟悉声音陪伴</string>
+    <string name="voice_clone_subtitle">录一小段声音，保存后就能在陪伴模式里使用。</string>
+    <string name="voice_clone_choose_avatar">先选一个形象</string>
+    <string name="voice_clone_record_title">再录一段声音</string>
+    <string name="voice_clone_status_idle">已为您选好 %1$s。接下来录 10 到 20 秒清楚的人声，就可以开始保存了。</string>
+    <string name="voice_clone_status_recording">正在为 %1$s 记住这段声音，您慢慢说就好。</string>
+    <string name="voice_clone_status_sample_ready">这段声音已经记下来了。确认后，%1$s 就会学着用这份声音陪您说话。</string>
+    <string name="voice_clone_status_sample_replace">新的声音样本已经准备好。确认后，会把 %1$s 的熟悉声音更新成这段新的录音。</string>
+    <string name="voice_clone_status_creating">正在保存熟悉声音，请稍候。</string>
+    <string name="voice_clone_status_ready">熟悉声音已经准备好了。现在就可以让 %1$s 开始陪您说话。</string>
+    <string name="voice_clone_record_hint">建议在安静一点的地方录制，保持 10 厘米左右距离，自然说 10 到 20 秒，最长不超过 60 秒。</string>
+    <string name="voice_clone_record_script">可以这样慢慢念：\n您好，今天天气不错，我想和家里人说说话。最近吃得还可以，睡得也还行，希望大家都平平安安，心里踏实。</string>
+    <string name="voice_clone_record_time">录音时长：%1$s</string>
+    <string name="voice_clone_primary_start_record">开始录音</string>
+    <string name="voice_clone_primary_stop_record">结束录音</string>
+    <string name="voice_clone_primary_create">保存熟悉声音</string>
+    <string name="voice_clone_primary_creating">正在保存…</string>
+    <string name="voice_clone_primary_start_chat">开始陪伴</string>
+    <string name="voice_clone_secondary_rerecord">重新录一段</string>
+    <string name="voice_clone_secondary_cancel_sample">放弃这段新录音</string>
+    <string name="voice_clone_record_failed">录音没有成功，请再试一次</string>
+    <string name="voice_clone_record_too_short">这段声音太短了，请再录长一点</string>
+    <string name="voice_clone_record_auto_stopped">已经帮您停下来了，这段录音可以直接使用</string>
+    <string name="voice_clone_record_recommendation">已经录好了。如果能说到 10 秒以上，效果通常会更自然</string>
+    <string name="voice_clone_saved">熟悉声音已经保存好了</string>
+    <string name="voice_clone_create_failed">熟悉声音保存失败，请稍后再试</string>
+    <string name="voice_clone_unavailable">请先在熟悉声音陪伴里保存一段声音</string>
     <string name="main_download_tips">Github地址可能下载失败，您可以考虑使用代理或者将文件缓存到自己的存储服务</string>
     <string name="debug_message">调试信息</string>
 

test/src/main/res/values/colors.xml

@@ -16,4 +16,13 @@
     <color name="warm_text_secondary">#FF8D6E63</color>
     <color name="warm_text_hint">#FFBCAAA4</color>
     <color name="warm_accent">#FFE8926F</color>
+    <color name="warm_surface">#FFFFFBF7</color>
+    <color name="warm_surface_alt">#FFF0E2D4</color>
+    <color name="warm_button_primary">#FFC16845</color>
+    <color name="warm_button_primary_pressed">#FFAF583A</color>
+    <color name="warm_button_primary_disabled">#FFD9B9A8</color>
+    <color name="warm_button_secondary_bg">#FFF8EEE5</color>
+    <color name="warm_button_secondary_bg_pressed">#FFF0E0D2</color>
+    <color name="warm_button_secondary_bg_disabled">#FFF8F1EB</color>
+    <color name="warm_button_secondary_stroke">#FFD7B39B</color>
 </resources>

test/src/main/res/values/strings.xml

@@ -33,6 +33,36 @@
     <string name="ai_call_user_said">You just said: %1$s</string>
     <string name="ai_call_camera_paused">The camera is paused for now, you can keep talking</string>
     <string name="ai_call_error_retry">Something was a little busy just now, please try again shortly</string>
+    <string name="voice_clone_entry_title">Familiar Voice Companion</string>
+    <string name="voice_clone_entry_desc">Record a familiar voice first. After it is saved, the companion will try to speak with that voice.</string>
+    <string name="voice_clone_entry_action">Record Familiar Voice</string>
+    <string name="voice_clone_title">Familiar Voice Companion</string>
+    <string name="voice_clone_subtitle">Record a short voice sample, save it, and use it in companion mode.</string>
+    <string name="voice_clone_choose_avatar">Choose an avatar first</string>
+    <string name="voice_clone_record_title">Then record a voice sample</string>
+    <string name="voice_clone_status_idle">%1$s is ready. Record 10 to 20 seconds of clear speech, then save it.</string>
+    <string name="voice_clone_status_recording">Recording a familiar voice for %1$s. You can speak slowly.</string>
+    <string name="voice_clone_status_sample_ready">This voice sample is ready. After saving, %1$s will try to speak with this voice.</string>
+    <string name="voice_clone_status_sample_replace">A new sample is ready. After saving, %1$s will use this new familiar voice.</string>
+    <string name="voice_clone_status_creating">Saving the familiar voice. Please wait.</string>
+    <string name="voice_clone_status_ready">The familiar voice is ready. %1$s can start the conversation now.</string>
+    <string name="voice_clone_record_hint">Try to record in a quiet place, keep about 10 cm away, speak naturally for 10 to 20 seconds, and do not exceed 60 seconds.</string>
+    <string name="voice_clone_record_script">You may read this slowly:\nHello, today feels calm and comfortable. I would like to talk with my family for a while, and I hope everyone is safe and at ease.</string>
+    <string name="voice_clone_record_time">Recording time: %1$s</string>
+    <string name="voice_clone_primary_start_record">Start Recording</string>
+    <string name="voice_clone_primary_stop_record">Stop Recording</string>
+    <string name="voice_clone_primary_create">Save Familiar Voice</string>
+    <string name="voice_clone_primary_creating">Saving…</string>
+    <string name="voice_clone_primary_start_chat">Start Companion Mode</string>
+    <string name="voice_clone_secondary_rerecord">Record Again</string>
+    <string name="voice_clone_secondary_cancel_sample">Discard New Recording</string>
+    <string name="voice_clone_record_failed">Recording failed, please try again</string>
+    <string name="voice_clone_record_too_short">The recording is too short, please record a little longer</string>
+    <string name="voice_clone_record_auto_stopped">The recording has been stopped automatically and can be used now</string>
+    <string name="voice_clone_record_recommendation">The sample is saved. If it is longer than 10 seconds, the result is usually more natural</string>
+    <string name="voice_clone_saved">The familiar voice has been saved</string>
+    <string name="voice_clone_create_failed">Failed to save the familiar voice, please try again later</string>
+    <string name="voice_clone_unavailable">Please save a familiar voice first</string>
     <string name="main_download_tips">The GitHub url may fail to download, you can consider using a proxy or caching the file to your own storage service</string>
     <string name="debug_message">Debug message</string>
 </resources>