@omote/avatar 0.1.1

package/dist/index.cjs

@@ -0,0 +1,288 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+
+// src/index.ts
+var index_exports = {};
+__export(index_exports, {
+  OmoteAvatarCore: () => OmoteAvatarCore
+});
+module.exports = __toCommonJS(index_exports);
+
+// src/OmoteAvatarCore.ts
+var import_core = require("@omote/core");
+var OmoteAvatarCore = class {
+  constructor() {
+    // Voice primitives
+    this.ttsSpeaker = null;
+    this.speechListener = null;
+    this.voiceOrchestrator = null;
+    // Frame source wiring
+    this.frameSourceUnsub = null;
+    this.stateUnsub = null;
+    // Callbacks set by the adapter
+    this._onFrame = null;
+    this._onStateChange = null;
+    // State
+    this._isSpeaking = false;
+    this._state = "idle";
+  }
+  get isSpeaking() {
+    return this._isSpeaking;
+  }
+  get state() {
+    return this._state;
+  }
+  get speaker() {
+    return this.ttsSpeaker ?? this.voiceOrchestrator?.speaker ?? null;
+  }
+  get listener() {
+    return this.speechListener ?? this.voiceOrchestrator?.listener ?? null;
+  }
+  /**
+   * Set the frame handler. Called by the adapter to receive blendshape frames.
+   * The adapter writes these to its renderer (morphTargets, refs, etc.).
+   */
+  set onFrame(handler) {
+    this._onFrame = handler;
+  }
+  /**
+   * Set the state change handler. Called by the adapter to sync conversational state.
+   */
+  set onStateChange(handler) {
+    this._onStateChange = handler;
+  }
+  // ---------------------------------------------------------------------------
+  // Speaker (TTS → lip sync)
+  // ---------------------------------------------------------------------------
+  /** Warm up AudioContext for iOS/Safari autoplay policy. Call from user gesture. */
+  async warmup() {
+    if (this.ttsSpeaker) await this.ttsSpeaker.warmup();
+  }
+  async connectSpeaker(tts, config) {
+    await this.disconnectSpeaker();
+    const speaker = new import_core.TTSSpeaker();
+    await speaker.connect(tts, config);
+    this.ttsSpeaker = speaker;
+    this.wireSpeakerFrameSource(speaker);
+  }
+  async disconnectSpeaker() {
+    this.frameSourceUnsub?.();
+    this.frameSourceUnsub = null;
+    if (this.ttsSpeaker) {
+      await this.ttsSpeaker.dispose();
+      this.ttsSpeaker = null;
+    }
+  }
+  async speak(text, options) {
+    if (this.voiceOrchestrator) {
+      await this.voiceOrchestrator.speak(text, options);
+      return;
+    }
+    if (!this.ttsSpeaker) throw new Error("No speaker connected. Call connectSpeaker() first.");
+    this._isSpeaking = true;
+    this.updateState("speaking");
+    try {
+      await this.ttsSpeaker.speak(text, options);
+    } finally {
+      this._isSpeaking = false;
+      if (this._state === "speaking") this.updateState("idle");
+    }
+  }
+  async streamText(options) {
+    if (this.voiceOrchestrator) {
+      return this.voiceOrchestrator.streamText(options);
+    }
+    if (!this.ttsSpeaker) throw new Error("No speaker connected. Call connectSpeaker() first.");
+    this._isSpeaking = true;
+    this.updateState("speaking");
+    const stream = await this.ttsSpeaker.streamText(options ?? {});
+    return {
+      push: stream.push,
+      end: async () => {
+        try {
+          await stream.end();
+        } finally {
+          this._isSpeaking = false;
+          if (this._state === "speaking") this.updateState("idle");
+        }
+      }
+    };
+  }
+  stopSpeaking() {
+    if (this.voiceOrchestrator) {
+      this.voiceOrchestrator.stopSpeaking();
+      return;
+    }
+    this.ttsSpeaker?.stop();
+  }
+  // ---------------------------------------------------------------------------
+  // Listener (mic → VAD → ASR → transcript)
+  // ---------------------------------------------------------------------------
+  async connectListener(config) {
+    if (this.speechListener) {
+      await this.speechListener.dispose();
+    }
+    const listener = new import_core.SpeechListener(config);
+    this.speechListener = listener;
+    await listener.loadModels();
+  }
+  async disconnectListener() {
+    if (this.speechListener) {
+      await this.speechListener.dispose();
+      this.speechListener = null;
+    }
+  }
+  async startListening() {
+    if (this.voiceOrchestrator) {
+      await this.voiceOrchestrator.startListening();
+      return;
+    }
+    if (!this.speechListener) throw new Error("No listener connected. Call connectListener() first.");
+    this.updateState("listening");
+    await this.speechListener.start();
+  }
+  stopListening() {
+    if (this.voiceOrchestrator) {
+      this.voiceOrchestrator.stopListening();
+      return;
+    }
+    this.speechListener?.stop();
+    if (this._state === "listening") this.updateState("idle");
+  }
+  // Event subscriptions (require listener or orchestrator)
+  onTranscript(callback) {
+    const listener = this.speechListener ?? this.voiceOrchestrator?.listener;
+    if (!listener) throw new Error("No listener connected. Call connectListener() or connectVoice() first.");
+    listener.on("transcript", callback);
+    return () => {
+      listener.off?.("transcript", callback);
+    };
+  }
+  onTranscriptEvent(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("transcript", callback);
+  }
+  onVoiceStateChange(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("state", callback);
+  }
+  onLoadingProgress(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("loading:progress", callback);
+  }
+  onError(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("error", callback);
+  }
+  onAudioLevel(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("audio:level", callback);
+  }
+  // ---------------------------------------------------------------------------
+  // Voice (combined speaker + listener + interruption)
+  // ---------------------------------------------------------------------------
+  async connectVoice(config) {
+    await this.disconnectVoice();
+    const orchestrator = new import_core.VoiceOrchestrator();
+    this.voiceOrchestrator = orchestrator;
+    await orchestrator.connect(config);
+    if (orchestrator.frameSource) {
+      this.wireFrameSource(orchestrator.frameSource);
+    }
+    this.stateUnsub = orchestrator.on("state", (state) => {
+      this._state = state;
+      this._isSpeaking = state === "speaking";
+      this._onStateChange?.(state);
+    });
+  }
+  async disconnectVoice() {
+    if (this.voiceOrchestrator) {
+      this.frameSourceUnsub?.();
+      this.frameSourceUnsub = null;
+      this.stateUnsub?.();
+      this.stateUnsub = null;
+      await this.voiceOrchestrator.disconnect();
+      this.voiceOrchestrator = null;
+    }
+  }
+  // ---------------------------------------------------------------------------
+  // Frame source
+  // ---------------------------------------------------------------------------
+  connectFrameSource(source) {
+    this.disconnectFrameSource();
+    this.wireFrameSource(source);
+  }
+  disconnectFrameSource() {
+    this.frameSourceUnsub?.();
+    this.frameSourceUnsub = null;
+  }
+  // ---------------------------------------------------------------------------
+  // State helpers
+  // ---------------------------------------------------------------------------
+  setState(state) {
+    this._state = state;
+    this._onStateChange?.(state);
+  }
+  setSpeaking(speaking) {
+    this._isSpeaking = speaking;
+  }
+  reset() {
+    this._isSpeaking = false;
+    this._state = "idle";
+    this._onStateChange?.("idle");
+  }
+  // ---------------------------------------------------------------------------
+  // Cleanup
+  // ---------------------------------------------------------------------------
+  async dispose() {
+    await this.disconnectVoice();
+    await this.disconnectSpeaker();
+    await this.disconnectListener();
+    this._onFrame = null;
+    this._onStateChange = null;
+  }
+  // ---------------------------------------------------------------------------
+  // Internal
+  // ---------------------------------------------------------------------------
+  wireFrameSource(source) {
+    const handler = (frame) => {
+      this._onFrame?.(frame);
+    };
+    source.on("frame", handler);
+    this.frameSourceUnsub = () => {
+      source.off?.("frame", handler);
+    };
+  }
+  wireSpeakerFrameSource(speaker) {
+    const source = speaker.frameSource;
+    if (source) {
+      this.wireFrameSource(source);
+    }
+  }
+  updateState(state) {
+    this._state = state;
+    this._onStateChange?.(state);
+  }
+};
+//# sourceMappingURL=index.cjs.map

package/dist/index.cjs.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/index.ts","../src/OmoteAvatarCore.ts"],"sourcesContent":["export { OmoteAvatarCore } from './OmoteAvatarCore';\nexport type { SpeakOptions, StreamTextSink, FrameHandler, StateChangeHandler } from './types';\n","/**\n * OmoteAvatarCore — Renderer-agnostic voice composition for OmoteAvatar adapters.\n *\n * Holds all voice state (TTSSpeaker, SpeechListener, VoiceOrchestrator) and\n * the imperative voice API. Adapters instantiate this and delegate voice\n * methods, keeping only renderer-specific code (animation, blendshape writes).\n *\n * Eliminates ~150 lines of identical voice wiring duplicated across\n * Three.js, Babylon.js, and R3F adapters.\n *\n * @category Avatar\n */\n\nimport {\n  TTSSpeaker,\n  SpeechListener,\n  VoiceOrchestrator,\n} from '@omote/core';\nimport type {\n  TTSBackend,\n  TTSSpeakerConfig,\n  SpeechListenerConfig,\n  TranscriptResult,\n  LoadingProgress,\n  ConversationalState,\n  FrameSource,\n  VoiceOrchestratorConfig,\n} from '@omote/core';\nimport type { SpeakOptions, StreamTextSink, FrameHandler, StateChangeHandler } from './types';\n\n/**\n * Renderer-agnostic voice composition class.\n *\n * Owns: TTSSpeaker, SpeechListener, VoiceOrchestrator, frame source wiring.\n * Does NOT own: CharacterController, animation, scene graph, morph targets.\n */\nexport class OmoteAvatarCore {\n  // Voice primitives\n  private ttsSpeaker: TTSSpeaker | null = null;\n  private speechListener: SpeechListener | null = null;\n  private voiceOrchestrator: VoiceOrchestrator | null = null;\n\n  // Frame source wiring\n  private frameSourceUnsub: (() => void) | null = null;\n  private stateUnsub: (() => void) | null = null;\n\n  // Callbacks set by the adapter\n  private _onFrame: FrameHandler | null = null;\n  private _onStateChange: StateChangeHandler | null = null;\n\n  // State\n  private _isSpeaking = false;\n  private _state: ConversationalState = 'idle';\n\n  get isSpeaking(): boolean { return this._isSpeaking; }\n  get state(): ConversationalState { return this._state; }\n  get speaker(): TTSSpeaker | null { return this.ttsSpeaker ?? this.voiceOrchestrator?.speaker ?? null; }\n  get listener(): SpeechListener | null { return this.speechListener ?? this.voiceOrchestrator?.listener ?? null; }\n\n  /**\n   * Set the frame handler. Called by the adapter to receive blendshape frames.\n   * The adapter writes these to its renderer (morphTargets, refs, etc.).\n   */\n  set onFrame(handler: FrameHandler | null) { this._onFrame = handler; }\n\n  /**\n   * Set the state change handler. Called by the adapter to sync conversational state.\n   */\n  set onStateChange(handler: StateChangeHandler | null) { this._onStateChange = handler; }\n\n  // ---------------------------------------------------------------------------\n  // Speaker (TTS → lip sync)\n  // ---------------------------------------------------------------------------\n\n  /** Warm up AudioContext for iOS/Safari autoplay policy. Call from user gesture. */\n  async warmup(): Promise<void> {\n    if (this.ttsSpeaker) await this.ttsSpeaker.warmup();\n  }\n\n  async connectSpeaker(tts: TTSBackend, config?: TTSSpeakerConfig): Promise<void> {\n    await this.disconnectSpeaker();\n    const speaker = new TTSSpeaker();\n    await speaker.connect(tts, config);\n    this.ttsSpeaker = speaker;\n    this.wireSpeakerFrameSource(speaker);\n  }\n\n  async disconnectSpeaker(): Promise<void> {\n    this.frameSourceUnsub?.();\n    this.frameSourceUnsub = null;\n    if (this.ttsSpeaker) {\n      await this.ttsSpeaker.dispose();\n      this.ttsSpeaker = null;\n    }\n  }\n\n  async speak(text: string, options?: SpeakOptions): Promise<void> {\n    if (this.voiceOrchestrator) {\n      await this.voiceOrchestrator.speak(text, options);\n      return;\n    }\n    if (!this.ttsSpeaker) throw new Error('No speaker connected. Call connectSpeaker() first.');\n    this._isSpeaking = true;\n    this.updateState('speaking');\n    try {\n      await this.ttsSpeaker.speak(text, options);\n    } finally {\n      this._isSpeaking = false;\n      if (this._state === 'speaking') this.updateState('idle');\n    }\n  }\n\n  async streamText(options?: SpeakOptions): Promise<StreamTextSink> {\n    if (this.voiceOrchestrator) {\n      return this.voiceOrchestrator.streamText(options);\n    }\n    if (!this.ttsSpeaker) throw new Error('No speaker connected. Call connectSpeaker() first.');\n    this._isSpeaking = true;\n    this.updateState('speaking');\n    const stream = await this.ttsSpeaker.streamText(options ?? {});\n    return {\n      push: stream.push,\n      end: async () => {\n        try { await stream.end(); }\n        finally {\n          this._isSpeaking = false;\n          if (this._state === 'speaking') this.updateState('idle');\n        }\n      },\n    };\n  }\n\n  stopSpeaking(): void {\n    if (this.voiceOrchestrator) {\n      this.voiceOrchestrator.stopSpeaking();\n      return;\n    }\n    this.ttsSpeaker?.stop();\n  }\n\n  // ---------------------------------------------------------------------------\n  // Listener (mic → VAD → ASR → transcript)\n  // ---------------------------------------------------------------------------\n\n  async connectListener(config?: SpeechListenerConfig): Promise<void> {\n    if (this.speechListener) {\n      await this.speechListener.dispose();\n    }\n    const listener = new SpeechListener(config);\n    this.speechListener = listener;\n    await listener.loadModels();\n  }\n\n  async disconnectListener(): Promise<void> {\n    if (this.speechListener) {\n      await this.speechListener.dispose();\n      this.speechListener = null;\n    }\n  }\n\n  async startListening(): Promise<void> {\n    if (this.voiceOrchestrator) {\n      await this.voiceOrchestrator.startListening();\n      return;\n    }\n    if (!this.speechListener) throw new Error('No listener connected. Call connectListener() first.');\n    this.updateState('listening');\n    await this.speechListener.start();\n  }\n\n  stopListening(): void {\n    if (this.voiceOrchestrator) {\n      this.voiceOrchestrator.stopListening();\n      return;\n    }\n    this.speechListener?.stop();\n    if (this._state === 'listening') this.updateState('idle');\n  }\n\n  // Event subscriptions (require listener or orchestrator)\n  onTranscript(callback: (result: TranscriptResult) => void): () => void {\n    const listener = this.speechListener ?? this.voiceOrchestrator?.listener;\n    if (!listener) throw new Error('No listener connected. Call connectListener() or connectVoice() first.');\n    listener.on('transcript', callback);\n    return () => { listener.off?.('transcript', callback); };\n  }\n\n  onTranscriptEvent(callback: (result: TranscriptResult) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('transcript', callback);\n  }\n\n  onVoiceStateChange(callback: (state: ConversationalState) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('state', callback);\n  }\n\n  onLoadingProgress(callback: (progress: LoadingProgress) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('loading:progress', callback);\n  }\n\n  onError(callback: (error: Error) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('error', callback);\n  }\n\n  onAudioLevel(callback: (level: { rms: number; peak: number }) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('audio:level', callback);\n  }\n\n  // ---------------------------------------------------------------------------\n  // Voice (combined speaker + listener + interruption)\n  // ---------------------------------------------------------------------------\n\n  async connectVoice(config: VoiceOrchestratorConfig): Promise<void> {\n    await this.disconnectVoice();\n    const orchestrator = new VoiceOrchestrator();\n    this.voiceOrchestrator = orchestrator;\n    await orchestrator.connect(config);\n\n    // Wire frame source from orchestrator\n    if (orchestrator.frameSource) {\n      this.wireFrameSource(orchestrator.frameSource);\n    }\n\n    // Sync state from orchestrator\n    this.stateUnsub = orchestrator.on('state', (state) => {\n      this._state = state;\n      this._isSpeaking = state === 'speaking';\n      this._onStateChange?.(state);\n    });\n  }\n\n  async disconnectVoice(): Promise<void> {\n    if (this.voiceOrchestrator) {\n      this.frameSourceUnsub?.();\n      this.frameSourceUnsub = null;\n      this.stateUnsub?.();\n      this.stateUnsub = null;\n      await this.voiceOrchestrator.disconnect();\n      this.voiceOrchestrator = null;\n    }\n  }\n\n  // ---------------------------------------------------------------------------\n  // Frame source\n  // ---------------------------------------------------------------------------\n\n  connectFrameSource(source: FrameSource): void {\n    this.disconnectFrameSource();\n    this.wireFrameSource(source);\n  }\n\n  disconnectFrameSource(): void {\n    this.frameSourceUnsub?.();\n    this.frameSourceUnsub = null;\n  }\n\n  // ---------------------------------------------------------------------------\n  // State helpers\n  // ---------------------------------------------------------------------------\n\n  setState(state: ConversationalState): void {\n    this._state = state;\n    this._onStateChange?.(state);\n  }\n\n  setSpeaking(speaking: boolean): void {\n    this._isSpeaking = speaking;\n  }\n\n  reset(): void {\n    this._isSpeaking = false;\n    this._state = 'idle';\n    this._onStateChange?.('idle');\n  }\n\n  // ---------------------------------------------------------------------------\n  // Cleanup\n  // ---------------------------------------------------------------------------\n\n  async dispose(): Promise<void> {\n    await this.disconnectVoice();\n    await this.disconnectSpeaker();\n    await this.disconnectListener();\n    this._onFrame = null;\n    this._onStateChange = null;\n  }\n\n  // ---------------------------------------------------------------------------\n  // Internal\n  // ---------------------------------------------------------------------------\n\n  private wireFrameSource(source: FrameSource): void {\n    const handler = (frame: { blendshapes: Float32Array; emotion?: string }) => {\n      this._onFrame?.(frame);\n    };\n    source.on('frame', handler);\n    this.frameSourceUnsub = () => { source.off?.('frame', handler); };\n  }\n\n  private wireSpeakerFrameSource(speaker: TTSSpeaker): void {\n    const source = speaker.frameSource;\n    if (source) {\n      this.wireFrameSource(source);\n    }\n  }\n\n  private updateState(state: ConversationalState): void {\n    this._state = state;\n    this._onStateChange?.(state);\n  }\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;;;ACaA,kBAIO;AAmBA,IAAM,kBAAN,MAAsB;AAAA,EAAtB;AAEL;AAAA,SAAQ,aAAgC;AACxC,SAAQ,iBAAwC;AAChD,SAAQ,oBAA8C;AAGtD;AAAA,SAAQ,mBAAwC;AAChD,SAAQ,aAAkC;AAG1C;AAAA,SAAQ,WAAgC;AACxC,SAAQ,iBAA4C;AAGpD;AAAA,SAAQ,cAAc;AACtB,SAAQ,SAA8B;AAAA;AAAA,EAEtC,IAAI,aAAsB;AAAE,WAAO,KAAK;AAAA,EAAa;AAAA,EACrD,IAAI,QAA6B;AAAE,WAAO,KAAK;AAAA,EAAQ;AAAA,EACvD,IAAI,UAA6B;AAAE,WAAO,KAAK,cAAc,KAAK,mBAAmB,WAAW;AAAA,EAAM;AAAA,EACtG,IAAI,WAAkC;AAAE,WAAO,KAAK,kBAAkB,KAAK,mBAAmB,YAAY;AAAA,EAAM;AAAA;AAAA;AAAA;AAAA;AAAA,EAMhH,IAAI,QAAQ,SAA8B;AAAE,SAAK,WAAW;AAAA,EAAS;AAAA;AAAA;AAAA;AAAA,EAKrE,IAAI,cAAc,SAAoC;AAAE,SAAK,iBAAiB;AAAA,EAAS;AAAA;AAAA;AAAA;AAAA;AAAA,EAOvF,MAAM,SAAwB;AAC5B,QAAI,KAAK,WAAY,OAAM,KAAK,WAAW,OAAO;AAAA,EACpD;AAAA,EAEA,MAAM,eAAe,KAAiB,QAA0C;AAC9E,UAAM,KAAK,kBAAkB;AAC7B,UAAM,UAAU,IAAI,uBAAW;AAC/B,UAAM,QAAQ,QAAQ,KAAK,MAAM;AACjC,SAAK,aAAa;AAClB,SAAK,uBAAuB,OAAO;AAAA,EACrC;AAAA,EAEA,MAAM,oBAAmC;AACvC,SAAK,mBAAmB;AACxB,SAAK,mBAAmB;AACxB,QAAI,KAAK,YAAY;AACnB,YAAM,KAAK,WAAW,QAAQ;AAC9B,WAAK,aAAa;AAAA,IACpB;AAAA,EACF;AAAA,EAEA,MAAM,MAAM,MAAc,SAAuC;AAC/D,QAAI,KAAK,mBAAmB;AAC1B,YAAM,KAAK,kBAAkB,MAAM,MAAM,OAAO;AAChD;AAAA,IACF;AACA,QAAI,CAAC,KAAK,WAAY,OAAM,IAAI,MAAM,oDAAoD;AAC1F,SAAK,cAAc;AACnB,SAAK,YAAY,UAAU;AAC3B,QAAI;AACF,YAAM,KAAK,WAAW,MAAM,MAAM,OAAO;AAAA,IAC3C,UAAE;AACA,WAAK,cAAc;AACnB,UAAI,KAAK,WAAW,WAAY,MAAK,YAAY,MAAM;AAAA,IACzD;AAAA,EACF;AAAA,EAEA,MAAM,WAAW,SAAiD;AAChE,QAAI,KAAK,mBAAmB;AAC1B,aAAO,KAAK,kBAAkB,WAAW,OAAO;AAAA,IAClD;AACA,QAAI,CAAC,KAAK,WAAY,OAAM,IAAI,MAAM,oDAAoD;AAC1F,SAAK,cAAc;AACnB,SAAK,YAAY,UAAU;AAC3B,UAAM,SAAS,MAAM,KAAK,WAAW,WAAW,WAAW,CAAC,CAAC;AAC7D,WAAO;AAAA,MACL,MAAM,OAAO;AAAA,MACb,KAAK,YAAY;AACf,YAAI;AAAE,gBAAM,OAAO,IAAI;AAAA,QAAG,UAC1B;AACE,eAAK,cAAc;AACnB,cAAI,KAAK,WAAW,WAAY,MAAK,YAAY,MAAM;AAAA,QACzD;AAAA,MACF;AAAA,IACF;AAAA,EACF;AAAA,EAEA,eAAqB;AACnB,QAAI,KAAK,mBAAmB;AAC1B,WAAK,kBAAkB,aAAa;AACpC;AAAA,IACF;AACA,SAAK,YAAY,KAAK;AAAA,EACxB;AAAA;AAAA;AAAA;AAAA,EAMA,MAAM,gBAAgB,QAA8C;AAClE,QAAI,KAAK,gBAAgB;AACvB,YAAM,KAAK,eAAe,QAAQ;AAAA,IACpC;AACA,UAAM,WAAW,IAAI,2BAAe,MAAM;AAC1C,SAAK,iBAAiB;AACtB,UAAM,SAAS,WAAW;AAAA,EAC5B;AAAA,EAEA,MAAM,qBAAoC;AACxC,QAAI,KAAK,gBAAgB;AACvB,YAAM,KAAK,eAAe,QAAQ;AAClC,WAAK,iBAAiB;AAAA,IACxB;AAAA,EACF;AAAA,EAEA,MAAM,iBAAgC;AACpC,QAAI,KAAK,mBAAmB;AAC1B,YAAM,KAAK,kBAAkB,eAAe;AAC5C;AAAA,IACF;AACA,QAAI,CAAC,KAAK,eAAgB,OAAM,IAAI,MAAM,sDAAsD;AAChG,SAAK,YAAY,WAAW;AAC5B,UAAM,KAAK,eAAe,MAAM;AAAA,EAClC;AAAA,EAEA,gBAAsB;AACpB,QAAI,KAAK,mBAAmB;AAC1B,WAAK,kBAAkB,cAAc;AACrC;AAAA,IACF;AACA,SAAK,gBAAgB,KAAK;AAC1B,QAAI,KAAK,WAAW,YAAa,MAAK,YAAY,MAAM;AAAA,EAC1D;AAAA;AAAA,EAGA,aAAa,UAA0D;AACrE,UAAM,WAAW,KAAK,kBAAkB,KAAK,mBAAmB;AAChE,QAAI,CAAC,SAAU,OAAM,IAAI,MAAM,wEAAwE;AACvG,aAAS,GAAG,cAAc,QAAQ;AAClC,WAAO,MAAM;AAAE,eAAS,MAAM,cAAc,QAAQ;AAAA,IAAG;AAAA,EACzD;AAAA,EAEA,kBAAkB,UAA0D;AAC1E,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,cAAc,QAAQ;AAAA,EACvC;AAAA,EAEA,mBAAmB,UAA4D;AAC7E,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,SAAS,QAAQ;AAAA,EAClC;AAAA,EAEA,kBAAkB,UAA2D;AAC3E,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,oBAAoB,QAAQ;AAAA,EAC7C;AAAA,EAEA,QAAQ,UAA8C;AACpD,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,SAAS,QAAQ;AAAA,EAClC;AAAA,EAEA,aAAa,UAAsE;AACjF,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,eAAe,QAAQ;AAAA,EACxC;AAAA;AAAA;AAAA;AAAA,EAMA,MAAM,aAAa,QAAgD;AACjE,UAAM,KAAK,gBAAgB;AAC3B,UAAM,eAAe,IAAI,8BAAkB;AAC3C,SAAK,oBAAoB;AACzB,UAAM,aAAa,QAAQ,MAAM;AAGjC,QAAI,aAAa,aAAa;AAC5B,WAAK,gBAAgB,aAAa,WAAW;AAAA,IAC/C;AAGA,SAAK,aAAa,aAAa,GAAG,SAAS,CAAC,UAAU;AACpD,WAAK,SAAS;AACd,WAAK,cAAc,UAAU;AAC7B,WAAK,iBAAiB,KAAK;AAAA,IAC7B,CAAC;AAAA,EACH;AAAA,EAEA,MAAM,kBAAiC;AACrC,QAAI,KAAK,mBAAmB;AAC1B,WAAK,mBAAmB;AACxB,WAAK,mBAAmB;AACxB,WAAK,aAAa;AAClB,WAAK,aAAa;AAClB,YAAM,KAAK,kBAAkB,WAAW;AACxC,WAAK,oBAAoB;AAAA,IAC3B;AAAA,EACF;AAAA;AAAA;AAAA;AAAA,EAMA,mBAAmB,QAA2B;AAC5C,SAAK,sBAAsB;AAC3B,SAAK,gBAAgB,MAAM;AAAA,EAC7B;AAAA,EAEA,wBAA8B;AAC5B,SAAK,mBAAmB;AACxB,SAAK,mBAAmB;AAAA,EAC1B;AAAA;AAAA;AAAA;AAAA,EAMA,SAAS,OAAkC;AACzC,SAAK,SAAS;AACd,SAAK,iBAAiB,KAAK;AAAA,EAC7B;AAAA,EAEA,YAAY,UAAyB;AACnC,SAAK,cAAc;AAAA,EACrB;AAAA,EAEA,QAAc;AACZ,SAAK,cAAc;AACnB,SAAK,SAAS;AACd,SAAK,iBAAiB,MAAM;AAAA,EAC9B;AAAA;AAAA;AAAA;AAAA,EAMA,MAAM,UAAyB;AAC7B,UAAM,KAAK,gBAAgB;AAC3B,UAAM,KAAK,kBAAkB;AAC7B,UAAM,KAAK,mBAAmB;AAC9B,SAAK,WAAW;AAChB,SAAK,iBAAiB;AAAA,EACxB;AAAA;AAAA;AAAA;AAAA,EAMQ,gBAAgB,QAA2B;AACjD,UAAM,UAAU,CAAC,UAA2D;AAC1E,WAAK,WAAW,KAAK;AAAA,IACvB;AACA,WAAO,GAAG,SAAS,OAAO;AAC1B,SAAK,mBAAmB,MAAM;AAAE,aAAO,MAAM,SAAS,OAAO;AAAA,IAAG;AAAA,EAClE;AAAA,EAEQ,uBAAuB,SAA2B;AACxD,UAAM,SAAS,QAAQ;AACvB,QAAI,QAAQ;AACV,WAAK,gBAAgB,MAAM;AAAA,IAC7B;AAAA,EACF;AAAA,EAEQ,YAAY,OAAkC;AACpD,SAAK,SAAS;AACd,SAAK,iBAAiB,KAAK;AAAA,EAC7B;AACF;","names":[]}

package/dist/index.d.cts

@@ -0,0 +1,104 @@
+import { ConversationalState, TTSSpeaker, SpeechListener, TTSBackend, TTSSpeakerConfig, SpeechListenerConfig, TranscriptResult, LoadingProgress, VoiceOrchestratorConfig, FrameSource } from '@omote/core';
+
+/**
+ * Types for @omote/avatar — renderer-agnostic voice composition.
+ *
+ * @category Avatar
+ */
+
+/** Options passed to `speak()` and `streamText()`. */
+interface SpeakOptions {
+    signal?: AbortSignal;
+    voice?: string;
+    speed?: number;
+    language?: string;
+}
+/** Sink returned by `streamText()` for token-by-token streaming TTS. */
+interface StreamTextSink {
+    push: (token: string) => void;
+    end: () => Promise<void>;
+}
+/** Callback receiving blendshape frames from a FrameSource. */
+type FrameHandler = (frame: {
+    blendshapes: Float32Array;
+    emotion?: string;
+}) => void;
+/** Callback receiving conversational state changes. */
+type StateChangeHandler = (state: ConversationalState) => void;
+
+/**
+ * OmoteAvatarCore — Renderer-agnostic voice composition for OmoteAvatar adapters.
+ *
+ * Holds all voice state (TTSSpeaker, SpeechListener, VoiceOrchestrator) and
+ * the imperative voice API. Adapters instantiate this and delegate voice
+ * methods, keeping only renderer-specific code (animation, blendshape writes).
+ *
+ * Eliminates ~150 lines of identical voice wiring duplicated across
+ * Three.js, Babylon.js, and R3F adapters.
+ *
+ * @category Avatar
+ */
+
+/**
+ * Renderer-agnostic voice composition class.
+ *
+ * Owns: TTSSpeaker, SpeechListener, VoiceOrchestrator, frame source wiring.
+ * Does NOT own: CharacterController, animation, scene graph, morph targets.
+ */
+declare class OmoteAvatarCore {
+    private ttsSpeaker;
+    private speechListener;
+    private voiceOrchestrator;
+    private frameSourceUnsub;
+    private stateUnsub;
+    private _onFrame;
+    private _onStateChange;
+    private _isSpeaking;
+    private _state;
+    get isSpeaking(): boolean;
+    get state(): ConversationalState;
+    get speaker(): TTSSpeaker | null;
+    get listener(): SpeechListener | null;
+    /**
+     * Set the frame handler. Called by the adapter to receive blendshape frames.
+     * The adapter writes these to its renderer (morphTargets, refs, etc.).
+     */
+    set onFrame(handler: FrameHandler | null);
+    /**
+     * Set the state change handler. Called by the adapter to sync conversational state.
+     */
+    set onStateChange(handler: StateChangeHandler | null);
+    /** Warm up AudioContext for iOS/Safari autoplay policy. Call from user gesture. */
+    warmup(): Promise<void>;
+    connectSpeaker(tts: TTSBackend, config?: TTSSpeakerConfig): Promise<void>;
+    disconnectSpeaker(): Promise<void>;
+    speak(text: string, options?: SpeakOptions): Promise<void>;
+    streamText(options?: SpeakOptions): Promise<StreamTextSink>;
+    stopSpeaking(): void;
+    connectListener(config?: SpeechListenerConfig): Promise<void>;
+    disconnectListener(): Promise<void>;
+    startListening(): Promise<void>;
+    stopListening(): void;
+    onTranscript(callback: (result: TranscriptResult) => void): () => void;
+    onTranscriptEvent(callback: (result: TranscriptResult) => void): () => void;
+    onVoiceStateChange(callback: (state: ConversationalState) => void): () => void;
+    onLoadingProgress(callback: (progress: LoadingProgress) => void): () => void;
+    onError(callback: (error: Error) => void): () => void;
+    onAudioLevel(callback: (level: {
+        rms: number;
+        peak: number;
+    }) => void): () => void;
+    connectVoice(config: VoiceOrchestratorConfig): Promise<void>;
+    disconnectVoice(): Promise<void>;
+    connectFrameSource(source: FrameSource): void;
+    disconnectFrameSource(): void;
+    setState(state: ConversationalState): void;
+    setSpeaking(speaking: boolean): void;
+    reset(): void;
+    dispose(): Promise<void>;
+    private wireFrameSource;
+    private wireSpeakerFrameSource;
+    private updateState;
+}
+
+export { type FrameHandler, OmoteAvatarCore, type SpeakOptions, type StateChangeHandler, type StreamTextSink };

package/dist/index.d.ts

@@ -0,0 +1,104 @@
+import { ConversationalState, TTSSpeaker, SpeechListener, TTSBackend, TTSSpeakerConfig, SpeechListenerConfig, TranscriptResult, LoadingProgress, VoiceOrchestratorConfig, FrameSource } from '@omote/core';
+
+/**
+ * Types for @omote/avatar — renderer-agnostic voice composition.
+ *
+ * @category Avatar
+ */
+
+/** Options passed to `speak()` and `streamText()`. */
+interface SpeakOptions {
+    signal?: AbortSignal;
+    voice?: string;
+    speed?: number;
+    language?: string;
+}
+/** Sink returned by `streamText()` for token-by-token streaming TTS. */
+interface StreamTextSink {
+    push: (token: string) => void;
+    end: () => Promise<void>;
+}
+/** Callback receiving blendshape frames from a FrameSource. */
+type FrameHandler = (frame: {
+    blendshapes: Float32Array;
+    emotion?: string;
+}) => void;
+/** Callback receiving conversational state changes. */
+type StateChangeHandler = (state: ConversationalState) => void;
+
+/**
+ * OmoteAvatarCore — Renderer-agnostic voice composition for OmoteAvatar adapters.
+ *
+ * Holds all voice state (TTSSpeaker, SpeechListener, VoiceOrchestrator) and
+ * the imperative voice API. Adapters instantiate this and delegate voice
+ * methods, keeping only renderer-specific code (animation, blendshape writes).
+ *
+ * Eliminates ~150 lines of identical voice wiring duplicated across
+ * Three.js, Babylon.js, and R3F adapters.
+ *
+ * @category Avatar
+ */
+
+/**
+ * Renderer-agnostic voice composition class.
+ *
+ * Owns: TTSSpeaker, SpeechListener, VoiceOrchestrator, frame source wiring.
+ * Does NOT own: CharacterController, animation, scene graph, morph targets.
+ */
+declare class OmoteAvatarCore {
+    private ttsSpeaker;
+    private speechListener;
+    private voiceOrchestrator;
+    private frameSourceUnsub;
+    private stateUnsub;
+    private _onFrame;
+    private _onStateChange;
+    private _isSpeaking;
+    private _state;
+    get isSpeaking(): boolean;
+    get state(): ConversationalState;
+    get speaker(): TTSSpeaker | null;
+    get listener(): SpeechListener | null;
+    /**
+     * Set the frame handler. Called by the adapter to receive blendshape frames.
+     * The adapter writes these to its renderer (morphTargets, refs, etc.).
+     */
+    set onFrame(handler: FrameHandler | null);
+    /**
+     * Set the state change handler. Called by the adapter to sync conversational state.
+     */
+    set onStateChange(handler: StateChangeHandler | null);
+    /** Warm up AudioContext for iOS/Safari autoplay policy. Call from user gesture. */
+    warmup(): Promise<void>;
+    connectSpeaker(tts: TTSBackend, config?: TTSSpeakerConfig): Promise<void>;
+    disconnectSpeaker(): Promise<void>;
+    speak(text: string, options?: SpeakOptions): Promise<void>;
+    streamText(options?: SpeakOptions): Promise<StreamTextSink>;
+    stopSpeaking(): void;
+    connectListener(config?: SpeechListenerConfig): Promise<void>;
+    disconnectListener(): Promise<void>;
+    startListening(): Promise<void>;
+    stopListening(): void;
+    onTranscript(callback: (result: TranscriptResult) => void): () => void;
+    onTranscriptEvent(callback: (result: TranscriptResult) => void): () => void;
+    onVoiceStateChange(callback: (state: ConversationalState) => void): () => void;
+    onLoadingProgress(callback: (progress: LoadingProgress) => void): () => void;
+    onError(callback: (error: Error) => void): () => void;
+    onAudioLevel(callback: (level: {
+        rms: number;
+        peak: number;
+    }) => void): () => void;
+    connectVoice(config: VoiceOrchestratorConfig): Promise<void>;
+    disconnectVoice(): Promise<void>;
+    connectFrameSource(source: FrameSource): void;
+    disconnectFrameSource(): void;
+    setState(state: ConversationalState): void;
+    setSpeaking(speaking: boolean): void;
+    reset(): void;
+    dispose(): Promise<void>;
+    private wireFrameSource;
+    private wireSpeakerFrameSource;
+    private updateState;
+}
+
+export { type FrameHandler, OmoteAvatarCore, type SpeakOptions, type StateChangeHandler, type StreamTextSink };

package/dist/index.js

@@ -0,0 +1,269 @@
+// src/OmoteAvatarCore.ts
+import {
+  TTSSpeaker,
+  SpeechListener,
+  VoiceOrchestrator
+} from "@omote/core";
+var OmoteAvatarCore = class {
+  constructor() {
+    // Voice primitives
+    this.ttsSpeaker = null;
+    this.speechListener = null;
+    this.voiceOrchestrator = null;
+    // Frame source wiring
+    this.frameSourceUnsub = null;
+    this.stateUnsub = null;
+    // Callbacks set by the adapter
+    this._onFrame = null;
+    this._onStateChange = null;
+    // State
+    this._isSpeaking = false;
+    this._state = "idle";
+  }
+  get isSpeaking() {
+    return this._isSpeaking;
+  }
+  get state() {
+    return this._state;
+  }
+  get speaker() {
+    return this.ttsSpeaker ?? this.voiceOrchestrator?.speaker ?? null;
+  }
+  get listener() {
+    return this.speechListener ?? this.voiceOrchestrator?.listener ?? null;
+  }
+  /**
+   * Set the frame handler. Called by the adapter to receive blendshape frames.
+   * The adapter writes these to its renderer (morphTargets, refs, etc.).
+   */
+  set onFrame(handler) {
+    this._onFrame = handler;
+  }
+  /**
+   * Set the state change handler. Called by the adapter to sync conversational state.
+   */
+  set onStateChange(handler) {
+    this._onStateChange = handler;
+  }
+  // ---------------------------------------------------------------------------
+  // Speaker (TTS → lip sync)
+  // ---------------------------------------------------------------------------
+  /** Warm up AudioContext for iOS/Safari autoplay policy. Call from user gesture. */
+  async warmup() {
+    if (this.ttsSpeaker) await this.ttsSpeaker.warmup();
+  }
+  async connectSpeaker(tts, config) {
+    await this.disconnectSpeaker();
+    const speaker = new TTSSpeaker();
+    await speaker.connect(tts, config);
+    this.ttsSpeaker = speaker;
+    this.wireSpeakerFrameSource(speaker);
+  }
+  async disconnectSpeaker() {
+    this.frameSourceUnsub?.();
+    this.frameSourceUnsub = null;
+    if (this.ttsSpeaker) {
+      await this.ttsSpeaker.dispose();
+      this.ttsSpeaker = null;
+    }
+  }
+  async speak(text, options) {
+    if (this.voiceOrchestrator) {
+      await this.voiceOrchestrator.speak(text, options);
+      return;
+    }
+    if (!this.ttsSpeaker) throw new Error("No speaker connected. Call connectSpeaker() first.");
+    this._isSpeaking = true;
+    this.updateState("speaking");
+    try {
+      await this.ttsSpeaker.speak(text, options);
+    } finally {
+      this._isSpeaking = false;
+      if (this._state === "speaking") this.updateState("idle");
+    }
+  }
+  async streamText(options) {
+    if (this.voiceOrchestrator) {
+      return this.voiceOrchestrator.streamText(options);
+    }
+    if (!this.ttsSpeaker) throw new Error("No speaker connected. Call connectSpeaker() first.");
+    this._isSpeaking = true;
+    this.updateState("speaking");
+    const stream = await this.ttsSpeaker.streamText(options ?? {});
+    return {
+      push: stream.push,
+      end: async () => {
+        try {
+          await stream.end();
+        } finally {
+          this._isSpeaking = false;
+          if (this._state === "speaking") this.updateState("idle");
+        }
+      }
+    };
+  }
+  stopSpeaking() {
+    if (this.voiceOrchestrator) {
+      this.voiceOrchestrator.stopSpeaking();
+      return;
+    }
+    this.ttsSpeaker?.stop();
+  }
+  // ---------------------------------------------------------------------------
+  // Listener (mic → VAD → ASR → transcript)
+  // ---------------------------------------------------------------------------
+  async connectListener(config) {
+    if (this.speechListener) {
+      await this.speechListener.dispose();
+    }
+    const listener = new SpeechListener(config);
+    this.speechListener = listener;
+    await listener.loadModels();
+  }
+  async disconnectListener() {
+    if (this.speechListener) {
+      await this.speechListener.dispose();
+      this.speechListener = null;
+    }
+  }
+  async startListening() {
+    if (this.voiceOrchestrator) {
+      await this.voiceOrchestrator.startListening();
+      return;
+    }
+    if (!this.speechListener) throw new Error("No listener connected. Call connectListener() first.");
+    this.updateState("listening");
+    await this.speechListener.start();
+  }
+  stopListening() {
+    if (this.voiceOrchestrator) {
+      this.voiceOrchestrator.stopListening();
+      return;
+    }
+    this.speechListener?.stop();
+    if (this._state === "listening") this.updateState("idle");
+  }
+  // Event subscriptions (require listener or orchestrator)
+  onTranscript(callback) {
+    const listener = this.speechListener ?? this.voiceOrchestrator?.listener;
+    if (!listener) throw new Error("No listener connected. Call connectListener() or connectVoice() first.");
+    listener.on("transcript", callback);
+    return () => {
+      listener.off?.("transcript", callback);
+    };
+  }
+  onTranscriptEvent(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("transcript", callback);
+  }
+  onVoiceStateChange(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("state", callback);
+  }
+  onLoadingProgress(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("loading:progress", callback);
+  }
+  onError(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("error", callback);
+  }
+  onAudioLevel(callback) {
+    const orch = this.voiceOrchestrator;
+    if (!orch) throw new Error("No voice connected. Call connectVoice() first.");
+    return orch.on("audio:level", callback);
+  }
+  // ---------------------------------------------------------------------------
+  // Voice (combined speaker + listener + interruption)
+  // ---------------------------------------------------------------------------
+  async connectVoice(config) {
+    await this.disconnectVoice();
+    const orchestrator = new VoiceOrchestrator();
+    this.voiceOrchestrator = orchestrator;
+    await orchestrator.connect(config);
+    if (orchestrator.frameSource) {
+      this.wireFrameSource(orchestrator.frameSource);
+    }
+    this.stateUnsub = orchestrator.on("state", (state) => {
+      this._state = state;
+      this._isSpeaking = state === "speaking";
+      this._onStateChange?.(state);
+    });
+  }
+  async disconnectVoice() {
+    if (this.voiceOrchestrator) {
+      this.frameSourceUnsub?.();
+      this.frameSourceUnsub = null;
+      this.stateUnsub?.();
+      this.stateUnsub = null;
+      await this.voiceOrchestrator.disconnect();
+      this.voiceOrchestrator = null;
+    }
+  }
+  // ---------------------------------------------------------------------------
+  // Frame source
+  // ---------------------------------------------------------------------------
+  connectFrameSource(source) {
+    this.disconnectFrameSource();
+    this.wireFrameSource(source);
+  }
+  disconnectFrameSource() {
+    this.frameSourceUnsub?.();
+    this.frameSourceUnsub = null;
+  }
+  // ---------------------------------------------------------------------------
+  // State helpers
+  // ---------------------------------------------------------------------------
+  setState(state) {
+    this._state = state;
+    this._onStateChange?.(state);
+  }
+  setSpeaking(speaking) {
+    this._isSpeaking = speaking;
+  }
+  reset() {
+    this._isSpeaking = false;
+    this._state = "idle";
+    this._onStateChange?.("idle");
+  }
+  // ---------------------------------------------------------------------------
+  // Cleanup
+  // ---------------------------------------------------------------------------
+  async dispose() {
+    await this.disconnectVoice();
+    await this.disconnectSpeaker();
+    await this.disconnectListener();
+    this._onFrame = null;
+    this._onStateChange = null;
+  }
+  // ---------------------------------------------------------------------------
+  // Internal
+  // ---------------------------------------------------------------------------
+  wireFrameSource(source) {
+    const handler = (frame) => {
+      this._onFrame?.(frame);
+    };
+    source.on("frame", handler);
+    this.frameSourceUnsub = () => {
+      source.off?.("frame", handler);
+    };
+  }
+  wireSpeakerFrameSource(speaker) {
+    const source = speaker.frameSource;
+    if (source) {
+      this.wireFrameSource(source);
+    }
+  }
+  updateState(state) {
+    this._state = state;
+    this._onStateChange?.(state);
+  }
+};
+export {
+  OmoteAvatarCore
+};
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/OmoteAvatarCore.ts"],"sourcesContent":["/**\n * OmoteAvatarCore — Renderer-agnostic voice composition for OmoteAvatar adapters.\n *\n * Holds all voice state (TTSSpeaker, SpeechListener, VoiceOrchestrator) and\n * the imperative voice API. Adapters instantiate this and delegate voice\n * methods, keeping only renderer-specific code (animation, blendshape writes).\n *\n * Eliminates ~150 lines of identical voice wiring duplicated across\n * Three.js, Babylon.js, and R3F adapters.\n *\n * @category Avatar\n */\n\nimport {\n  TTSSpeaker,\n  SpeechListener,\n  VoiceOrchestrator,\n} from '@omote/core';\nimport type {\n  TTSBackend,\n  TTSSpeakerConfig,\n  SpeechListenerConfig,\n  TranscriptResult,\n  LoadingProgress,\n  ConversationalState,\n  FrameSource,\n  VoiceOrchestratorConfig,\n} from '@omote/core';\nimport type { SpeakOptions, StreamTextSink, FrameHandler, StateChangeHandler } from './types';\n\n/**\n * Renderer-agnostic voice composition class.\n *\n * Owns: TTSSpeaker, SpeechListener, VoiceOrchestrator, frame source wiring.\n * Does NOT own: CharacterController, animation, scene graph, morph targets.\n */\nexport class OmoteAvatarCore {\n  // Voice primitives\n  private ttsSpeaker: TTSSpeaker | null = null;\n  private speechListener: SpeechListener | null = null;\n  private voiceOrchestrator: VoiceOrchestrator | null = null;\n\n  // Frame source wiring\n  private frameSourceUnsub: (() => void) | null = null;\n  private stateUnsub: (() => void) | null = null;\n\n  // Callbacks set by the adapter\n  private _onFrame: FrameHandler | null = null;\n  private _onStateChange: StateChangeHandler | null = null;\n\n  // State\n  private _isSpeaking = false;\n  private _state: ConversationalState = 'idle';\n\n  get isSpeaking(): boolean { return this._isSpeaking; }\n  get state(): ConversationalState { return this._state; }\n  get speaker(): TTSSpeaker | null { return this.ttsSpeaker ?? this.voiceOrchestrator?.speaker ?? null; }\n  get listener(): SpeechListener | null { return this.speechListener ?? this.voiceOrchestrator?.listener ?? null; }\n\n  /**\n   * Set the frame handler. Called by the adapter to receive blendshape frames.\n   * The adapter writes these to its renderer (morphTargets, refs, etc.).\n   */\n  set onFrame(handler: FrameHandler | null) { this._onFrame = handler; }\n\n  /**\n   * Set the state change handler. Called by the adapter to sync conversational state.\n   */\n  set onStateChange(handler: StateChangeHandler | null) { this._onStateChange = handler; }\n\n  // ---------------------------------------------------------------------------\n  // Speaker (TTS → lip sync)\n  // ---------------------------------------------------------------------------\n\n  /** Warm up AudioContext for iOS/Safari autoplay policy. Call from user gesture. */\n  async warmup(): Promise<void> {\n    if (this.ttsSpeaker) await this.ttsSpeaker.warmup();\n  }\n\n  async connectSpeaker(tts: TTSBackend, config?: TTSSpeakerConfig): Promise<void> {\n    await this.disconnectSpeaker();\n    const speaker = new TTSSpeaker();\n    await speaker.connect(tts, config);\n    this.ttsSpeaker = speaker;\n    this.wireSpeakerFrameSource(speaker);\n  }\n\n  async disconnectSpeaker(): Promise<void> {\n    this.frameSourceUnsub?.();\n    this.frameSourceUnsub = null;\n    if (this.ttsSpeaker) {\n      await this.ttsSpeaker.dispose();\n      this.ttsSpeaker = null;\n    }\n  }\n\n  async speak(text: string, options?: SpeakOptions): Promise<void> {\n    if (this.voiceOrchestrator) {\n      await this.voiceOrchestrator.speak(text, options);\n      return;\n    }\n    if (!this.ttsSpeaker) throw new Error('No speaker connected. Call connectSpeaker() first.');\n    this._isSpeaking = true;\n    this.updateState('speaking');\n    try {\n      await this.ttsSpeaker.speak(text, options);\n    } finally {\n      this._isSpeaking = false;\n      if (this._state === 'speaking') this.updateState('idle');\n    }\n  }\n\n  async streamText(options?: SpeakOptions): Promise<StreamTextSink> {\n    if (this.voiceOrchestrator) {\n      return this.voiceOrchestrator.streamText(options);\n    }\n    if (!this.ttsSpeaker) throw new Error('No speaker connected. Call connectSpeaker() first.');\n    this._isSpeaking = true;\n    this.updateState('speaking');\n    const stream = await this.ttsSpeaker.streamText(options ?? {});\n    return {\n      push: stream.push,\n      end: async () => {\n        try { await stream.end(); }\n        finally {\n          this._isSpeaking = false;\n          if (this._state === 'speaking') this.updateState('idle');\n        }\n      },\n    };\n  }\n\n  stopSpeaking(): void {\n    if (this.voiceOrchestrator) {\n      this.voiceOrchestrator.stopSpeaking();\n      return;\n    }\n    this.ttsSpeaker?.stop();\n  }\n\n  // ---------------------------------------------------------------------------\n  // Listener (mic → VAD → ASR → transcript)\n  // ---------------------------------------------------------------------------\n\n  async connectListener(config?: SpeechListenerConfig): Promise<void> {\n    if (this.speechListener) {\n      await this.speechListener.dispose();\n    }\n    const listener = new SpeechListener(config);\n    this.speechListener = listener;\n    await listener.loadModels();\n  }\n\n  async disconnectListener(): Promise<void> {\n    if (this.speechListener) {\n      await this.speechListener.dispose();\n      this.speechListener = null;\n    }\n  }\n\n  async startListening(): Promise<void> {\n    if (this.voiceOrchestrator) {\n      await this.voiceOrchestrator.startListening();\n      return;\n    }\n    if (!this.speechListener) throw new Error('No listener connected. Call connectListener() first.');\n    this.updateState('listening');\n    await this.speechListener.start();\n  }\n\n  stopListening(): void {\n    if (this.voiceOrchestrator) {\n      this.voiceOrchestrator.stopListening();\n      return;\n    }\n    this.speechListener?.stop();\n    if (this._state === 'listening') this.updateState('idle');\n  }\n\n  // Event subscriptions (require listener or orchestrator)\n  onTranscript(callback: (result: TranscriptResult) => void): () => void {\n    const listener = this.speechListener ?? this.voiceOrchestrator?.listener;\n    if (!listener) throw new Error('No listener connected. Call connectListener() or connectVoice() first.');\n    listener.on('transcript', callback);\n    return () => { listener.off?.('transcript', callback); };\n  }\n\n  onTranscriptEvent(callback: (result: TranscriptResult) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('transcript', callback);\n  }\n\n  onVoiceStateChange(callback: (state: ConversationalState) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('state', callback);\n  }\n\n  onLoadingProgress(callback: (progress: LoadingProgress) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('loading:progress', callback);\n  }\n\n  onError(callback: (error: Error) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('error', callback);\n  }\n\n  onAudioLevel(callback: (level: { rms: number; peak: number }) => void): () => void {\n    const orch = this.voiceOrchestrator;\n    if (!orch) throw new Error('No voice connected. Call connectVoice() first.');\n    return orch.on('audio:level', callback);\n  }\n\n  // ---------------------------------------------------------------------------\n  // Voice (combined speaker + listener + interruption)\n  // ---------------------------------------------------------------------------\n\n  async connectVoice(config: VoiceOrchestratorConfig): Promise<void> {\n    await this.disconnectVoice();\n    const orchestrator = new VoiceOrchestrator();\n    this.voiceOrchestrator = orchestrator;\n    await orchestrator.connect(config);\n\n    // Wire frame source from orchestrator\n    if (orchestrator.frameSource) {\n      this.wireFrameSource(orchestrator.frameSource);\n    }\n\n    // Sync state from orchestrator\n    this.stateUnsub = orchestrator.on('state', (state) => {\n      this._state = state;\n      this._isSpeaking = state === 'speaking';\n      this._onStateChange?.(state);\n    });\n  }\n\n  async disconnectVoice(): Promise<void> {\n    if (this.voiceOrchestrator) {\n      this.frameSourceUnsub?.();\n      this.frameSourceUnsub = null;\n      this.stateUnsub?.();\n      this.stateUnsub = null;\n      await this.voiceOrchestrator.disconnect();\n      this.voiceOrchestrator = null;\n    }\n  }\n\n  // ---------------------------------------------------------------------------\n  // Frame source\n  // ---------------------------------------------------------------------------\n\n  connectFrameSource(source: FrameSource): void {\n    this.disconnectFrameSource();\n    this.wireFrameSource(source);\n  }\n\n  disconnectFrameSource(): void {\n    this.frameSourceUnsub?.();\n    this.frameSourceUnsub = null;\n  }\n\n  // ---------------------------------------------------------------------------\n  // State helpers\n  // ---------------------------------------------------------------------------\n\n  setState(state: ConversationalState): void {\n    this._state = state;\n    this._onStateChange?.(state);\n  }\n\n  setSpeaking(speaking: boolean): void {\n    this._isSpeaking = speaking;\n  }\n\n  reset(): void {\n    this._isSpeaking = false;\n    this._state = 'idle';\n    this._onStateChange?.('idle');\n  }\n\n  // ---------------------------------------------------------------------------\n  // Cleanup\n  // ---------------------------------------------------------------------------\n\n  async dispose(): Promise<void> {\n    await this.disconnectVoice();\n    await this.disconnectSpeaker();\n    await this.disconnectListener();\n    this._onFrame = null;\n    this._onStateChange = null;\n  }\n\n  // ---------------------------------------------------------------------------\n  // Internal\n  // ---------------------------------------------------------------------------\n\n  private wireFrameSource(source: FrameSource): void {\n    const handler = (frame: { blendshapes: Float32Array; emotion?: string }) => {\n      this._onFrame?.(frame);\n    };\n    source.on('frame', handler);\n    this.frameSourceUnsub = () => { source.off?.('frame', handler); };\n  }\n\n  private wireSpeakerFrameSource(speaker: TTSSpeaker): void {\n    const source = speaker.frameSource;\n    if (source) {\n      this.wireFrameSource(source);\n    }\n  }\n\n  private updateState(state: ConversationalState): void {\n    this._state = state;\n    this._onStateChange?.(state);\n  }\n}\n"],"mappings":";AAaA;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,OACK;AAmBA,IAAM,kBAAN,MAAsB;AAAA,EAAtB;AAEL;AAAA,SAAQ,aAAgC;AACxC,SAAQ,iBAAwC;AAChD,SAAQ,oBAA8C;AAGtD;AAAA,SAAQ,mBAAwC;AAChD,SAAQ,aAAkC;AAG1C;AAAA,SAAQ,WAAgC;AACxC,SAAQ,iBAA4C;AAGpD;AAAA,SAAQ,cAAc;AACtB,SAAQ,SAA8B;AAAA;AAAA,EAEtC,IAAI,aAAsB;AAAE,WAAO,KAAK;AAAA,EAAa;AAAA,EACrD,IAAI,QAA6B;AAAE,WAAO,KAAK;AAAA,EAAQ;AAAA,EACvD,IAAI,UAA6B;AAAE,WAAO,KAAK,cAAc,KAAK,mBAAmB,WAAW;AAAA,EAAM;AAAA,EACtG,IAAI,WAAkC;AAAE,WAAO,KAAK,kBAAkB,KAAK,mBAAmB,YAAY;AAAA,EAAM;AAAA;AAAA;AAAA;AAAA;AAAA,EAMhH,IAAI,QAAQ,SAA8B;AAAE,SAAK,WAAW;AAAA,EAAS;AAAA;AAAA;AAAA;AAAA,EAKrE,IAAI,cAAc,SAAoC;AAAE,SAAK,iBAAiB;AAAA,EAAS;AAAA;AAAA;AAAA;AAAA;AAAA,EAOvF,MAAM,SAAwB;AAC5B,QAAI,KAAK,WAAY,OAAM,KAAK,WAAW,OAAO;AAAA,EACpD;AAAA,EAEA,MAAM,eAAe,KAAiB,QAA0C;AAC9E,UAAM,KAAK,kBAAkB;AAC7B,UAAM,UAAU,IAAI,WAAW;AAC/B,UAAM,QAAQ,QAAQ,KAAK,MAAM;AACjC,SAAK,aAAa;AAClB,SAAK,uBAAuB,OAAO;AAAA,EACrC;AAAA,EAEA,MAAM,oBAAmC;AACvC,SAAK,mBAAmB;AACxB,SAAK,mBAAmB;AACxB,QAAI,KAAK,YAAY;AACnB,YAAM,KAAK,WAAW,QAAQ;AAC9B,WAAK,aAAa;AAAA,IACpB;AAAA,EACF;AAAA,EAEA,MAAM,MAAM,MAAc,SAAuC;AAC/D,QAAI,KAAK,mBAAmB;AAC1B,YAAM,KAAK,kBAAkB,MAAM,MAAM,OAAO;AAChD;AAAA,IACF;AACA,QAAI,CAAC,KAAK,WAAY,OAAM,IAAI,MAAM,oDAAoD;AAC1F,SAAK,cAAc;AACnB,SAAK,YAAY,UAAU;AAC3B,QAAI;AACF,YAAM,KAAK,WAAW,MAAM,MAAM,OAAO;AAAA,IAC3C,UAAE;AACA,WAAK,cAAc;AACnB,UAAI,KAAK,WAAW,WAAY,MAAK,YAAY,MAAM;AAAA,IACzD;AAAA,EACF;AAAA,EAEA,MAAM,WAAW,SAAiD;AAChE,QAAI,KAAK,mBAAmB;AAC1B,aAAO,KAAK,kBAAkB,WAAW,OAAO;AAAA,IAClD;AACA,QAAI,CAAC,KAAK,WAAY,OAAM,IAAI,MAAM,oDAAoD;AAC1F,SAAK,cAAc;AACnB,SAAK,YAAY,UAAU;AAC3B,UAAM,SAAS,MAAM,KAAK,WAAW,WAAW,WAAW,CAAC,CAAC;AAC7D,WAAO;AAAA,MACL,MAAM,OAAO;AAAA,MACb,KAAK,YAAY;AACf,YAAI;AAAE,gBAAM,OAAO,IAAI;AAAA,QAAG,UAC1B;AACE,eAAK,cAAc;AACnB,cAAI,KAAK,WAAW,WAAY,MAAK,YAAY,MAAM;AAAA,QACzD;AAAA,MACF;AAAA,IACF;AAAA,EACF;AAAA,EAEA,eAAqB;AACnB,QAAI,KAAK,mBAAmB;AAC1B,WAAK,kBAAkB,aAAa;AACpC;AAAA,IACF;AACA,SAAK,YAAY,KAAK;AAAA,EACxB;AAAA;AAAA;AAAA;AAAA,EAMA,MAAM,gBAAgB,QAA8C;AAClE,QAAI,KAAK,gBAAgB;AACvB,YAAM,KAAK,eAAe,QAAQ;AAAA,IACpC;AACA,UAAM,WAAW,IAAI,eAAe,MAAM;AAC1C,SAAK,iBAAiB;AACtB,UAAM,SAAS,WAAW;AAAA,EAC5B;AAAA,EAEA,MAAM,qBAAoC;AACxC,QAAI,KAAK,gBAAgB;AACvB,YAAM,KAAK,eAAe,QAAQ;AAClC,WAAK,iBAAiB;AAAA,IACxB;AAAA,EACF;AAAA,EAEA,MAAM,iBAAgC;AACpC,QAAI,KAAK,mBAAmB;AAC1B,YAAM,KAAK,kBAAkB,eAAe;AAC5C;AAAA,IACF;AACA,QAAI,CAAC,KAAK,eAAgB,OAAM,IAAI,MAAM,sDAAsD;AAChG,SAAK,YAAY,WAAW;AAC5B,UAAM,KAAK,eAAe,MAAM;AAAA,EAClC;AAAA,EAEA,gBAAsB;AACpB,QAAI,KAAK,mBAAmB;AAC1B,WAAK,kBAAkB,cAAc;AACrC;AAAA,IACF;AACA,SAAK,gBAAgB,KAAK;AAC1B,QAAI,KAAK,WAAW,YAAa,MAAK,YAAY,MAAM;AAAA,EAC1D;AAAA;AAAA,EAGA,aAAa,UAA0D;AACrE,UAAM,WAAW,KAAK,kBAAkB,KAAK,mBAAmB;AAChE,QAAI,CAAC,SAAU,OAAM,IAAI,MAAM,wEAAwE;AACvG,aAAS,GAAG,cAAc,QAAQ;AAClC,WAAO,MAAM;AAAE,eAAS,MAAM,cAAc,QAAQ;AAAA,IAAG;AAAA,EACzD;AAAA,EAEA,kBAAkB,UAA0D;AAC1E,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,cAAc,QAAQ;AAAA,EACvC;AAAA,EAEA,mBAAmB,UAA4D;AAC7E,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,SAAS,QAAQ;AAAA,EAClC;AAAA,EAEA,kBAAkB,UAA2D;AAC3E,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,oBAAoB,QAAQ;AAAA,EAC7C;AAAA,EAEA,QAAQ,UAA8C;AACpD,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,SAAS,QAAQ;AAAA,EAClC;AAAA,EAEA,aAAa,UAAsE;AACjF,UAAM,OAAO,KAAK;AAClB,QAAI,CAAC,KAAM,OAAM,IAAI,MAAM,gDAAgD;AAC3E,WAAO,KAAK,GAAG,eAAe,QAAQ;AAAA,EACxC;AAAA;AAAA;AAAA;AAAA,EAMA,MAAM,aAAa,QAAgD;AACjE,UAAM,KAAK,gBAAgB;AAC3B,UAAM,eAAe,IAAI,kBAAkB;AAC3C,SAAK,oBAAoB;AACzB,UAAM,aAAa,QAAQ,MAAM;AAGjC,QAAI,aAAa,aAAa;AAC5B,WAAK,gBAAgB,aAAa,WAAW;AAAA,IAC/C;AAGA,SAAK,aAAa,aAAa,GAAG,SAAS,CAAC,UAAU;AACpD,WAAK,SAAS;AACd,WAAK,cAAc,UAAU;AAC7B,WAAK,iBAAiB,KAAK;AAAA,IAC7B,CAAC;AAAA,EACH;AAAA,EAEA,MAAM,kBAAiC;AACrC,QAAI,KAAK,mBAAmB;AAC1B,WAAK,mBAAmB;AACxB,WAAK,mBAAmB;AACxB,WAAK,aAAa;AAClB,WAAK,aAAa;AAClB,YAAM,KAAK,kBAAkB,WAAW;AACxC,WAAK,oBAAoB;AAAA,IAC3B;AAAA,EACF;AAAA;AAAA;AAAA;AAAA,EAMA,mBAAmB,QAA2B;AAC5C,SAAK,sBAAsB;AAC3B,SAAK,gBAAgB,MAAM;AAAA,EAC7B;AAAA,EAEA,wBAA8B;AAC5B,SAAK,mBAAmB;AACxB,SAAK,mBAAmB;AAAA,EAC1B;AAAA;AAAA;AAAA;AAAA,EAMA,SAAS,OAAkC;AACzC,SAAK,SAAS;AACd,SAAK,iBAAiB,KAAK;AAAA,EAC7B;AAAA,EAEA,YAAY,UAAyB;AACnC,SAAK,cAAc;AAAA,EACrB;AAAA,EAEA,QAAc;AACZ,SAAK,cAAc;AACnB,SAAK,SAAS;AACd,SAAK,iBAAiB,MAAM;AAAA,EAC9B;AAAA;AAAA;AAAA;AAAA,EAMA,MAAM,UAAyB;AAC7B,UAAM,KAAK,gBAAgB;AAC3B,UAAM,KAAK,kBAAkB;AAC7B,UAAM,KAAK,mBAAmB;AAC9B,SAAK,WAAW;AAChB,SAAK,iBAAiB;AAAA,EACxB;AAAA;AAAA;AAAA;AAAA,EAMQ,gBAAgB,QAA2B;AACjD,UAAM,UAAU,CAAC,UAA2D;AAC1E,WAAK,WAAW,KAAK;AAAA,IACvB;AACA,WAAO,GAAG,SAAS,OAAO;AAC1B,SAAK,mBAAmB,MAAM;AAAE,aAAO,MAAM,SAAS,OAAO;AAAA,IAAG;AAAA,EAClE;AAAA,EAEQ,uBAAuB,SAA2B;AACxD,UAAM,SAAS,QAAQ;AACvB,QAAI,QAAQ;AACV,WAAK,gBAAgB,MAAM;AAAA,IAC7B;AAAA,EACF;AAAA,EAEQ,YAAY,OAAkC;AACpD,SAAK,SAAS;AACd,SAAK,iBAAiB,KAAK;AAAA,EAC7B;AACF;","names":[]}

package/package.json

@@ -0,0 +1,40 @@
+{
+  "name": "@omote/avatar",
+  "version": "0.1.1",
+  "description": "Renderer-agnostic avatar voice composition for Omote SDK",
+  "type": "module",
+  "main": "./dist/index.cjs",
+  "module": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "import": "./dist/index.js",
+      "require": "./dist/index.cjs"
+    }
+  },
+  "scripts": {
+    "build": "tsup",
+    "typecheck": "tsc --noEmit"
+  },
+  "peerDependencies": {
+    "@omote/core": ">=0.10.0"
+  },
+  "devDependencies": {
+    "@omote/core": "file:../core",
+    "tsup": "^8.0.0",
+    "typescript": "^5.3.0"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "files": [
+    "dist"
+  ],
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/omoteai/omote.git",
+    "directory": "packages/avatar"
+  }
+}