qt-human 4.1.2 → 4.3.0

package/dist/index.d.ts

@@ -1880,6 +1880,13 @@ interface VoiceConfig {
   /** 是否要求唤醒词说两遍（如"小德小德"），默认 false（向后兼容） */
   doubleWakeWord?: boolean
 
+  /**
+   * 默认 false。true=唤醒词门控强模式:IDLE 下每轮必须带唤醒词前缀,旁人语音忽略。
+   * 门控强弱(单次 vs 唤醒词说两遍)由 `doubleWakeWord` 治理,且其默认值各模块不同:
+   * VoiceConversationE2E 默认 doubleWakeWord=true,VoiceConversation 默认 false。
+   */
+  wakePerTurn?: boolean
+
   /** 唤醒确认语，如"在呢"。唤醒后通过直接 TTS 播报（不走 LLM，不入历史） */
   wakeConfirmText?: string
 
@@ -2959,164 +2966,6 @@ declare class CameraSDK {
     destroy(): void;
 }
 
-declare class Human extends Emittery implements IHuman {
-    name: string;
-    core: ThreeJSApp | null;
-    private config;
-    loading?: Loading;
-    private audio;
-    private container;
-    recorder: any | null;
-    private recorderMessageHandler;
-    private voiceToTextManager;
-    private status;
-    private chatOrchestrator;
-    private onAnyEvent;
-    private characterData;
-    private isNotRender;
-    private mode;
-    private gatewayService;
-    framePlayer: FramePlayer | null;
-    camera: CameraSDK | null;
-    get sessionId(): string;
-    enableProactiveTTSPlayback(enable?: boolean): void;
-    private _proactiveAudioCtx;
-    private voiceInteractionConfig;
-    private voiceInteractionState;
-    private voiceInteractionExitTimer;
-    private exitPromptFired;
-    private secondExitTimer;
-    private exitTimerTurnId;
-    private voiceInteractionConversationId;
-    private voiceInteractionProcessing;
-    private voiceInteractionLastText;
-    private voiceInteractionLastTextTime;
-    private voiceInteractionTTSMode;
-    private voiceInteractionWebRTCDisabled;
-    private ttsState;
-    private eventBusHandlers;
-    private eventBusModule;
-    private _destroyed;
-    private updateTokenWarnedOnce;
-    private degradedNotice;
-    private static readonly IDLE_LOOP_MODE;
-    private static readonly BARGE_IN_MIN_TEXT_LENGTH;
-    private static readonly BARGE_IN_COOLDOWN_MS;
-    private static readonly INTERRUPT_INTENT_WORDS;
-    private static readonly GREETING_ONLY_WORDS;
-    private static readonly CHAR_HOMOPHONES;
-    private static generateNameVariants;
-    private static normalizeASRText;
-    private isDoubleWakeWord;
-    private speakDirect;
-    private lastBargeInTime;
-    private bargeInInProgress;
-    private voiceInteractionAskId;
-    private isAudioDataDiscarding;
-    private audioDiscardingTimer;
-    private static readonly AUDIO_DISCARDING_TIMEOUT_MS;
-    private isFirstUserInput;
-    private static readonly CONTINUATION_WINDOW_MS;
-    private static readonly MIN_VALID_TEXT_LENGTH;
-    private static readonly NOISE_WORDS;
-    private asrTextBuffer;
-    private asrSilenceTimer;
-    private pendingASRFinal;
-    private static readonly ASR_SILENCE_TIMEOUT_MS;
-    private static readonly DUPLICATE_TEXT_WINDOW_MS;
-    private pendingMergeText;
-    private textMergeTimer;
-    private textMergeStartTime;
-    private static readonly TEXT_MERGE_DELAY_MS;
-    private static readonly TEXT_MERGE_MAX_DELAY_MS;
-    private ttsPlaySafetyTimer;
-    private static readonly DEFAULT_TTS_PLAY_SAFETY_TIMEOUT_MS;
-    private framePlayerPlayEndHandler;
-    private audioDebugger;
-    private webrtcLipSync;
-    constructor(config: HumanConfig);
-    private mergeConfig;
-    private setupContainer;
-    private initFramePlayer;
-    private clearTTSPlaySafetyTimer;
-    private replaceFramePlayerPlayEndHandler;
-    private initGatewayService;
-    private setupGatewayEventForwarding;
-    private refreshGatewayToken;
-    setCharacterData(characterId: string): Promise<boolean>;
-    setCharacter(characterId: string, configuration?: Partial<HumanConfig>): Promise<boolean>;
-    speak(text: string, callback?: Function): Promise<StreamAudioPlayInfo>;
-    askCommand(params: AskParams, callback: (msgs: IResultData<string[]>) => void): Promise<void>;
-    askStream(params: AskParams, callback: (msgs: Result<ChatMessage$1[]>) => void, options?: {
-        skipStop?: boolean;
-    }): Promise<Result<AskStreamResponse>>;
-    ask(params: AskParams, callback: (result: Result<ChatMessage$1>) => void): Promise<Result<AskResponse>>;
-    speakByAudioShapes(uris: string[], shapes: Array<number[][]>): Promise<boolean>;
-    speakByAudioStreamShapes(audios: ArrayBuffer[], shapes: number[][], rate?: number): Promise<boolean>;
-    speakByAudio(uri: string, shapes: number[][]): Promise<boolean>;
-    startVoice2Text(callback?: (data: Voice2TextResult) => void): Promise<IResultData<RecorderPermission>>;
-    stopVoice2Text(callback?: (data: Voice2TextResult) => void): Promise<Result<null>>;
-    sendToolAction(tool: string, action: string, data: object): void;
-    stop(): Promise<boolean>;
-    private stopPlayback;
-    getStatus(): Status;
-    private initStorage;
-    muteAudio(isMute?: boolean): void;
-    muteAction(isMute?: boolean): void;
-    private dealVoice2Text;
-    enableAudio(): Promise<void>;
-    enableAudioDebug(enabled?: boolean): void;
-    getDebugAudioInfo(): {
-        sampleRate: number;
-        totalSize: number;
-        chunks: number;
-    };
-    downloadDebugAudio(filename?: string): void;
-    playDebugAudio(sampleRateOverride?: number): Promise<void>;
-    playDebugAudio2(sampleRateOverride?: number, streamMode?: boolean): Promise<void>;
-    private playIdleAnimation;
-    private finalizeTTSPlayback;
-    destroy(): Promise<void>;
-    refreshAuthToken(token: string): Promise<void>;
-    updateToken(token: string): void;
-    setAngleView({ camera }: AngleViewConfig): void;
-    playAction(code: string, opts?: PlayActionOptions): Promise<void>;
-    getModelInfo(): ModelInfo | null;
-    openVoiceInteraction(config: VoiceConfig): Promise<boolean>;
-    closeVoiceInteraction(): Promise<boolean>;
-    wakeVoiceInteraction(): void;
-    sleepVoiceInteraction(): void;
-    private createVoiceInteractionWebRTC;
-    private handleVoiceInteractionASR;
-    private flushPendingASRIfReady;
-    private processCompleteASRText;
-    private flushMergedText;
-    private sendTextToProcess;
-    private checkKeywords;
-    private containsInterruptIntent;
-    private triggerBargeIn;
-    private isNoiseText;
-    private handleVoiceInteractionWakeup;
-    private handleVoiceInteractionSleep;
-    private handleVoiceInteractionInterrupt;
-    private handleVoiceInteractionAsk;
-    private resetVoiceInteractionExitTimer;
-    offlineRender(files: CharacterModel[]): Promise<boolean>;
-    startRender(): Promise<boolean>;
-    stopRender(): Promise<boolean>;
-    closeRecorder(): Promise<boolean>;
-    openRecorder(): Promise<Result<RecorderPermission>>;
-    openCamera(config: CameraConfig): Promise<boolean>;
-    startFaceRecognition(config?: FaceRecognitionConfig): Promise<boolean>;
-    stopFaceRecognition(): Promise<boolean>;
-    startBodyDetect(): Promise<boolean>;
-    stopBodyDetect(): Promise<boolean>;
-    startSmileDetection(config?: SmileDetectionConfig): Promise<boolean>;
-    stopSmileDetection(): boolean;
-    closeCamera(): void;
-}
-//# sourceMappingURL=human.d.ts.map
-
 interface GatewayConfig {
     url: string;
     token: string;
@@ -3418,6 +3267,9 @@ interface VoiceConversationEvents {
     wakeConfirm: {
         text: string;
     };
+    audioLevel: {
+        level: number;
+    };
 }
 interface VoiceConversationConfig {
     gateway: GatewayService;
@@ -3440,6 +3292,7 @@ interface VoiceConversationConfig {
     enableInterrupt?: boolean;
     cooldownDuration?: number;
     doubleWakeWord?: boolean;
+    wakePerTurn?: boolean;
     wakeConfirmText?: string;
     syntheticTexts?: string[];
     gracefulExitEnabled?: boolean;
@@ -3468,7 +3321,7 @@ interface VoiceConversationConfig {
     onResponse?: (text: string) => void;
     onError?: (error: VoiceError$1) => void;
 }
-interface InternalConfig extends Required<Omit<VoiceConversationConfig, 'onStateChange' | 'onTranscript' | 'onResponse' | 'onError' | 'animations' | 'interruptAck' | 'interruptFeedback' | 'enableInterrupt' | 'cooldownDuration' | 'doubleWakeWord' | 'wakeConfirmText' | 'syntheticTexts' | 'gracefulExitEnabled' | 'gracefulExitPrompt' | 'gracefulExitPhase2Prompt' | 'gracefulExitPhase2Silence' | 'gracefulExitPhase3Silence' | 'bargeInStrictMode' | 'continuousWindowMs' | 'ttsTransport' | 'voiceEngine' | 'bargeInCooldownMs' | 'llmTimeoutMs' | 'llmMaxRetry' | 'webrtcMaxReconnectAttempts' | 'webrtcReconnectInterval' | 'vadGateWindowMs' | 'vadGraceMs' | 'interruptFollowUpWindowMs' | 'interruptWordFilterMs'>> {
+interface InternalConfig extends Required<Omit<VoiceConversationConfig, 'onStateChange' | 'onTranscript' | 'onResponse' | 'onError' | 'animations' | 'interruptAck' | 'interruptFeedback' | 'enableInterrupt' | 'cooldownDuration' | 'doubleWakeWord' | 'wakePerTurn' | 'wakeConfirmText' | 'syntheticTexts' | 'gracefulExitEnabled' | 'gracefulExitPrompt' | 'gracefulExitPhase2Prompt' | 'gracefulExitPhase2Silence' | 'gracefulExitPhase3Silence' | 'bargeInStrictMode' | 'continuousWindowMs' | 'ttsTransport' | 'voiceEngine' | 'bargeInCooldownMs' | 'llmTimeoutMs' | 'llmMaxRetry' | 'webrtcMaxReconnectAttempts' | 'webrtcReconnectInterval' | 'vadGateWindowMs' | 'vadGraceMs' | 'interruptFollowUpWindowMs' | 'interruptWordFilterMs'>> {
     onStateChange?: (state: VoiceState$1, prevState: VoiceState$1) => void;
     onTranscript?: (text: string, isFinal: boolean) => void;
     onResponse?: (text: string) => void;
@@ -3482,6 +3335,7 @@ interface InternalConfig extends Required<Omit<VoiceConversationConfig, 'onState
     interruptFeedback?: string;
     enableInterrupt: boolean;
     doubleWakeWord: boolean;
+    wakePerTurn: boolean;
     wakeConfirmText?: string;
     syntheticTexts: string[];
     gracefulExitEnabled: boolean;
@@ -3609,6 +3463,7 @@ declare class VoiceConversation extends Emittery {
     sleep(): void;
     suspendASR(): void;
     resumeASR(): void;
+    interrupt(): Promise<void>;
     private isGracefulExitPromptContext;
     private shouldCaptureGracefulExitFollowUp;
     private bufferGracefulExitFollowUp;
@@ -3695,6 +3550,167 @@ declare class VoiceConversation extends Emittery {
     private emitError;
 }
 
+declare class Human extends Emittery implements IHuman {
+    name: string;
+    core: ThreeJSApp | null;
+    private config;
+    loading?: Loading;
+    private audio;
+    private container;
+    recorder: any | null;
+    private recorderMessageHandler;
+    private voiceToTextManager;
+    private status;
+    private chatOrchestrator;
+    private onAnyEvent;
+    private characterData;
+    private isNotRender;
+    private mode;
+    private gatewayService;
+    framePlayer: FramePlayer | null;
+    camera: CameraSDK | null;
+    get sessionId(): string;
+    enableProactiveTTSPlayback(enable?: boolean): void;
+    private _proactiveAudioCtx;
+    private voiceInteractionConfig;
+    private voiceInteractionState;
+    private voiceInteractionExitTimer;
+    private exitPromptFired;
+    private secondExitTimer;
+    private exitTimerTurnId;
+    private voiceInteractionConversationId;
+    private voiceInteractionProcessing;
+    private voiceInteractionLastText;
+    private voiceInteractionLastTextTime;
+    private voiceInteractionTTSMode;
+    private voiceInteractionWebRTCDisabled;
+    private ttsState;
+    private eventBusHandlers;
+    private eventBusModule;
+    private _destroyed;
+    private updateTokenWarnedOnce;
+    private degradedNotice;
+    private static readonly IDLE_LOOP_MODE;
+    private static readonly BARGE_IN_MIN_TEXT_LENGTH;
+    private static readonly BARGE_IN_COOLDOWN_MS;
+    private static readonly INTERRUPT_INTENT_WORDS;
+    private static readonly GREETING_ONLY_WORDS;
+    private static readonly CHAR_HOMOPHONES;
+    private static generateNameVariants;
+    private static normalizeASRText;
+    private isDoubleWakeWord;
+    private speakDirect;
+    private lastBargeInTime;
+    private bargeInInProgress;
+    private voiceInteractionAskId;
+    private isAudioDataDiscarding;
+    private audioDiscardingTimer;
+    private static readonly AUDIO_DISCARDING_TIMEOUT_MS;
+    private isFirstUserInput;
+    private static readonly CONTINUATION_WINDOW_MS;
+    private static readonly MIN_VALID_TEXT_LENGTH;
+    private static readonly NOISE_WORDS;
+    private asrTextBuffer;
+    private asrSilenceTimer;
+    private pendingASRFinal;
+    private static readonly ASR_SILENCE_TIMEOUT_MS;
+    private static readonly DUPLICATE_TEXT_WINDOW_MS;
+    private pendingMergeText;
+    private textMergeTimer;
+    private textMergeStartTime;
+    private static readonly TEXT_MERGE_DELAY_MS;
+    private static readonly TEXT_MERGE_MAX_DELAY_MS;
+    private ttsPlaySafetyTimer;
+    private static readonly DEFAULT_TTS_PLAY_SAFETY_TIMEOUT_MS;
+    private framePlayerPlayEndHandler;
+    private audioDebugger;
+    private webrtcLipSync;
+    constructor(config: HumanConfig);
+    private mergeConfig;
+    private setupContainer;
+    private initFramePlayer;
+    private clearTTSPlaySafetyTimer;
+    private replaceFramePlayerPlayEndHandler;
+    private initGatewayService;
+    private setupGatewayEventForwarding;
+    private refreshGatewayToken;
+    setCharacterData(characterId: string): Promise<boolean>;
+    setCharacter(characterId: string, configuration?: Partial<HumanConfig>): Promise<boolean>;
+    speak(text: string, callback?: Function): Promise<StreamAudioPlayInfo>;
+    askCommand(params: AskParams, callback: (msgs: IResultData<string[]>) => void): Promise<void>;
+    askStream(params: AskParams, callback: (msgs: Result<ChatMessage$1[]>) => void, options?: {
+        skipStop?: boolean;
+    }): Promise<Result<AskStreamResponse>>;
+    ask(params: AskParams, callback: (result: Result<ChatMessage$1>) => void): Promise<Result<AskResponse>>;
+    speakByAudioShapes(uris: string[], shapes: Array<number[][]>): Promise<boolean>;
+    speakByAudioStreamShapes(audios: ArrayBuffer[], shapes: number[][], rate?: number): Promise<boolean>;
+    speakByAudio(uri: string, shapes: number[][]): Promise<boolean>;
+    startVoice2Text(callback?: (data: Voice2TextResult) => void): Promise<IResultData<RecorderPermission>>;
+    stopVoice2Text(callback?: (data: Voice2TextResult) => void): Promise<Result<null>>;
+    sendToolAction(tool: string, action: string, data: object): void;
+    stop(): Promise<boolean>;
+    private stopPlayback;
+    getStatus(): Status;
+    private initStorage;
+    muteAudio(isMute?: boolean): void;
+    muteAction(isMute?: boolean): void;
+    private dealVoice2Text;
+    enableAudio(): Promise<void>;
+    enableAudioDebug(enabled?: boolean): void;
+    getDebugAudioInfo(): {
+        sampleRate: number;
+        totalSize: number;
+        chunks: number;
+    };
+    downloadDebugAudio(filename?: string): void;
+    playDebugAudio(sampleRateOverride?: number): Promise<void>;
+    playDebugAudio2(sampleRateOverride?: number, streamMode?: boolean): Promise<void>;
+    private playIdleAnimation;
+    private finalizeTTSPlayback;
+    destroy(): Promise<void>;
+    refreshAuthToken(token: string): Promise<void>;
+    updateToken(token: string): void;
+    setAngleView({ camera }: AngleViewConfig): void;
+    playAction(code: string, opts?: PlayActionOptions): Promise<void>;
+    getModelInfo(): ModelInfo | null;
+    createVoiceConversation(config?: Omit<VoiceConversationConfig, 'gateway' | 'renderer' | 'framePlayer' | 'characterId'> & {
+        characterId?: string;
+    }): Promise<VoiceConversation>;
+    openVoiceInteraction(config: VoiceConfig): Promise<boolean>;
+    closeVoiceInteraction(): Promise<boolean>;
+    wakeVoiceInteraction(): void;
+    sleepVoiceInteraction(): void;
+    private createVoiceInteractionWebRTC;
+    private handleVoiceInteractionASR;
+    private flushPendingASRIfReady;
+    private processCompleteASRText;
+    private flushMergedText;
+    private sendTextToProcess;
+    private checkKeywords;
+    private containsInterruptIntent;
+    private triggerBargeIn;
+    private isNoiseText;
+    private handleVoiceInteractionWakeup;
+    private handleVoiceInteractionSleep;
+    private handleVoiceInteractionInterrupt;
+    private handleVoiceInteractionAsk;
+    private resetVoiceInteractionExitTimer;
+    offlineRender(files: CharacterModel[]): Promise<boolean>;
+    startRender(): Promise<boolean>;
+    stopRender(): Promise<boolean>;
+    closeRecorder(): Promise<boolean>;
+    openRecorder(): Promise<Result<RecorderPermission>>;
+    openCamera(config: CameraConfig): Promise<boolean>;
+    startFaceRecognition(config?: FaceRecognitionConfig): Promise<boolean>;
+    stopFaceRecognition(): Promise<boolean>;
+    startBodyDetect(): Promise<boolean>;
+    stopBodyDetect(): Promise<boolean>;
+    startSmileDetection(config?: SmileDetectionConfig): Promise<boolean>;
+    stopSmileDetection(): boolean;
+    closeCamera(): void;
+}
+//# sourceMappingURL=human.d.ts.map
+
 type VoiceState = 'SLEEPING' | 'IDLE' | 'LISTENING' | 'THINKING' | 'SPEAKING';
 interface VoiceError {
     code: 'WEBRTC_FAILED' | 'ASR_ERROR' | 'LLM_TIMEOUT' | 'LLM_ERROR' | 'TTS_FAILED' | 'GATEWAY_DISCONNECTED';
@@ -3713,6 +3729,7 @@ interface VoiceConversationE2EConfig {
     autoWake?: boolean;
     echoDetection?: boolean;
     doubleWakeWord?: boolean;
+    wakePerTurn?: boolean;
     bargeInStrictMode?: boolean;
     continuousWindowMs?: number;
     gracefulExitEnabled?: boolean;

package/dist/types/types/config.d.ts

@@ -293,6 +293,13 @@ export interface VoiceConfig {
   /** 是否要求唤醒词说两遍（如"小德小德"），默认 false（向后兼容） */
   doubleWakeWord?: boolean
 
+  /**
+   * 默认 false。true=唤醒词门控强模式:IDLE 下每轮必须带唤醒词前缀,旁人语音忽略。
+   * 门控强弱(单次 vs 唤醒词说两遍)由 `doubleWakeWord` 治理,且其默认值各模块不同:
+   * VoiceConversationE2E 默认 doubleWakeWord=true,VoiceConversation 默认 false。
+   */
+  wakePerTurn?: boolean
+
   /** 唤醒确认语，如"在呢"。唤醒后通过直接 TTS 播报（不走 LLM，不入历史） */
   wakeConfirmText?: string
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "qt-human",
-  "version": "4.1.2",
+  "version": "4.3.0",
   "description": "A JavaScript SDK for 3D digital human rendering and interaction",
   "publishConfig": {
     "registry": "https://registry.npmjs.org/"