@mastra/voice-google 0.14.0 → 0.14.1

package/dist/index.js

@@ -1,542 +1,559 @@
-import { PassThrough } from 'stream';
-import { SpeechClient, v2 } from '@google-cloud/speech';
-import { TextToSpeechClient } from '@google-cloud/text-to-speech';
-
-// src/index.ts
-
-// ../../packages/_internal-core/dist/chunk-3M4SEWMI.js
-var RegisteredLogger = {
-  LLM: "LLM"};
-var LogLevel = {
-  DEBUG: "debug",
-  INFO: "info",
-  WARN: "warn",
-  ERROR: "error"};
+import { PassThrough } from "stream";
+import { SpeechClient, v2 } from "@google-cloud/speech";
+import { TextToSpeechClient } from "@google-cloud/text-to-speech";
+//#region ../../packages/_internal-core/dist/logger/index.js
+const RegisteredLogger = {
+	AGENT: "AGENT",
+	OBSERVABILITY: "OBSERVABILITY",
+	AUTH: "AUTH",
+	BROWSER: "BROWSER",
+	NETWORK: "NETWORK",
+	WORKFLOW: "WORKFLOW",
+	LLM: "LLM",
+	TTS: "TTS",
+	VOICE: "VOICE",
+	VECTOR: "VECTOR",
+	BUNDLER: "BUNDLER",
+	DEPLOYER: "DEPLOYER",
+	MEMORY: "MEMORY",
+	STORAGE: "STORAGE",
+	EMBEDDINGS: "EMBEDDINGS",
+	MCP_SERVER: "MCP_SERVER",
+	SERVER_CACHE: "SERVER_CACHE",
+	SERVER: "SERVER",
+	WORKSPACE: "WORKSPACE",
+	CHANNEL: "CHANNEL"
+};
+const LogLevel = {
+	DEBUG: "debug",
+	INFO: "info",
+	WARN: "warn",
+	ERROR: "error",
+	NONE: "silent"
+};
 var MastraLogger = class {
-  name;
-  level;
-  transports;
-  constructor(options = {}) {
-    this.name = options.name || "Mastra";
-    this.level = options.level || LogLevel.ERROR;
-    this.transports = new Map(Object.entries(options.transports || {}));
-  }
-  getTransports() {
-    return this.transports;
-  }
-  trackException(_error, _metadata) {
-  }
-  async listLogs(transportId, params) {
-    if (!transportId || !this.transports.has(transportId)) {
-      return { logs: [], total: 0, page: params?.page ?? 1, perPage: params?.perPage ?? 100, hasMore: false };
-    }
-    return this.transports.get(transportId).listLogs?.(params) ?? {
-      logs: [],
-      total: 0,
-      page: params?.page ?? 1,
-      perPage: params?.perPage ?? 100,
-      hasMore: false
-    };
-  }
-  async listLogsByRunId({
-    transportId,
-    runId,
-    fromDate,
-    toDate,
-    logLevel,
-    filters,
-    page,
-    perPage
-  }) {
-    if (!transportId || !this.transports.has(transportId) || !runId) {
-      return { logs: [], total: 0, page: page ?? 1, perPage: perPage ?? 100, hasMore: false };
-    }
-    return this.transports.get(transportId).listLogsByRunId?.({ runId, fromDate, toDate, logLevel, filters, page, perPage }) ?? {
-      logs: [],
-      total: 0,
-      page: page ?? 1,
-      perPage: perPage ?? 100,
-      hasMore: false
-    };
-  }
+	name;
+	level;
+	transports;
+	constructor(options = {}) {
+		this.name = options.name || "Mastra";
+		this.level = options.level || LogLevel.ERROR;
+		this.transports = new Map(Object.entries(options.transports || {}));
+	}
+	getTransports() {
+		return this.transports;
+	}
+	trackException(_error, _metadata) {}
+	async listLogs(transportId, params) {
+		if (!transportId || !this.transports.has(transportId)) return {
+			logs: [],
+			total: 0,
+			page: params?.page ?? 1,
+			perPage: params?.perPage ?? 100,
+			hasMore: false
+		};
+		return this.transports.get(transportId).listLogs?.(params) ?? {
+			logs: [],
+			total: 0,
+			page: params?.page ?? 1,
+			perPage: params?.perPage ?? 100,
+			hasMore: false
+		};
+	}
+	async listLogsByRunId({ transportId, runId, fromDate, toDate, logLevel, filters, page, perPage }) {
+		if (!transportId || !this.transports.has(transportId) || !runId) return {
+			logs: [],
+			total: 0,
+			page: page ?? 1,
+			perPage: perPage ?? 100,
+			hasMore: false
+		};
+		return this.transports.get(transportId).listLogsByRunId?.({
+			runId,
+			fromDate,
+			toDate,
+			logLevel,
+			filters,
+			page,
+			perPage
+		}) ?? {
+			logs: [],
+			total: 0,
+			page: page ?? 1,
+			perPage: perPage ?? 100,
+			hasMore: false
+		};
+	}
 };
-var ConsoleLogger = class _ConsoleLogger extends MastraLogger {
-  component;
-  filter;
-  constructor(options = {}) {
-    super(options);
-    this.component = options.component;
-    this.filter = options.filter;
-  }
-  child(componentOrBindings) {
-    const component = typeof componentOrBindings === "string" ? componentOrBindings : componentOrBindings?.component ?? this.component;
-    return new _ConsoleLogger({
-      name: this.name,
-      level: this.level,
-      component,
-      filter: this.filter
-    });
-  }
-  shouldLog(level, message, args) {
-    if (!this.filter) return true;
-    try {
-      return this.filter({ component: this.component, level, message, args });
-    } catch (e) {
-      console.error(`[Logger] Filter error for component=${this.component} level=${level}:`, e);
-      return true;
-    }
-  }
-  prefix() {
-    return this.component ? `[${this.component}] ` : "";
-  }
-  debug(message, ...args) {
-    if (this.level === LogLevel.DEBUG && this.shouldLog(LogLevel.DEBUG, message, args)) {
-      console.info(`${this.prefix()}${message}`, ...args);
-    }
-  }
-  info(message, ...args) {
-    if ((this.level === LogLevel.INFO || this.level === LogLevel.DEBUG) && this.shouldLog(LogLevel.INFO, message, args)) {
-      console.info(`${this.prefix()}${message}`, ...args);
-    }
-  }
-  warn(message, ...args) {
-    if ((this.level === LogLevel.WARN || this.level === LogLevel.INFO || this.level === LogLevel.DEBUG) && this.shouldLog(LogLevel.WARN, message, args)) {
-      console.warn(`${this.prefix()}${message}`, ...args);
-    }
-  }
-  error(message, ...args) {
-    if ((this.level === LogLevel.ERROR || this.level === LogLevel.WARN || this.level === LogLevel.INFO || this.level === LogLevel.DEBUG) && this.shouldLog(LogLevel.ERROR, message, args)) {
-      console.error(`${this.prefix()}${message}`, ...args);
-    }
-  }
-  async listLogs(_transportId, _params) {
-    return { logs: [], total: 0, page: _params?.page ?? 1, perPage: _params?.perPage ?? 100, hasMore: false };
-  }
-  async listLogsByRunId(_args) {
-    return { logs: [], total: 0, page: _args.page ?? 1, perPage: _args.perPage ?? 100, hasMore: false };
-  }
+var ConsoleLogger = class ConsoleLogger extends MastraLogger {
+	component;
+	filter;
+	constructor(options = {}) {
+		super(options);
+		this.component = options.component;
+		this.filter = options.filter;
+	}
+	child(componentOrBindings) {
+		const component = typeof componentOrBindings === "string" ? componentOrBindings : componentOrBindings?.component ?? this.component;
+		return new ConsoleLogger({
+			name: this.name,
+			level: this.level,
+			component,
+			filter: this.filter
+		});
+	}
+	shouldLog(level, message, args) {
+		if (!this.filter) return true;
+		try {
+			return this.filter({
+				component: this.component,
+				level,
+				message,
+				args
+			});
+		} catch (e) {
+			console.error(`[Logger] Filter error for component=${this.component} level=${level}:`, e);
+			return true;
+		}
+	}
+	prefix() {
+		return this.component ? `[${this.component}] ` : "";
+	}
+	debug(message, ...args) {
+		if (this.level === LogLevel.DEBUG && this.shouldLog(LogLevel.DEBUG, message, args)) console.info(`${this.prefix()}${message}`, ...args);
+	}
+	info(message, ...args) {
+		if ((this.level === LogLevel.INFO || this.level === LogLevel.DEBUG) && this.shouldLog(LogLevel.INFO, message, args)) console.info(`${this.prefix()}${message}`, ...args);
+	}
+	warn(message, ...args) {
+		if ((this.level === LogLevel.WARN || this.level === LogLevel.INFO || this.level === LogLevel.DEBUG) && this.shouldLog(LogLevel.WARN, message, args)) console.warn(`${this.prefix()}${message}`, ...args);
+	}
+	error(message, ...args) {
+		if ((this.level === LogLevel.ERROR || this.level === LogLevel.WARN || this.level === LogLevel.INFO || this.level === LogLevel.DEBUG) && this.shouldLog(LogLevel.ERROR, message, args)) console.error(`${this.prefix()}${message}`, ...args);
+	}
+	async listLogs(_transportId, _params) {
+		return {
+			logs: [],
+			total: 0,
+			page: _params?.page ?? 1,
+			perPage: _params?.perPage ?? 100,
+			hasMore: false
+		};
+	}
+	async listLogsByRunId(_args) {
+		return {
+			logs: [],
+			total: 0,
+			page: _args.page ?? 1,
+			perPage: _args.perPage ?? 100,
+			hasMore: false
+		};
+	}
 };
-
-// ../../packages/_internal-core/dist/base/index.js
+//#endregion
+//#region ../../packages/_internal-core/dist/base/index.js
 var MastraBase = class {
-  component = RegisteredLogger.LLM;
-  logger;
-  name;
-  #rawConfig;
-  constructor({
-    component,
-    name,
-    rawConfig
-  }) {
-    this.component = component || RegisteredLogger.LLM;
-    this.name = name;
-    this.#rawConfig = rawConfig;
-    this.logger = new ConsoleLogger({ name: `${this.component} - ${this.name}` });
-  }
-  /**
-   * Returns the raw storage configuration this primitive was created from,
-   * or undefined if it was created from code.
-   */
-  toRawConfig() {
-    return this.#rawConfig;
-  }
-  /**
-   * Sets the raw storage configuration for this primitive.
-   * @internal
-   */
-  __setRawConfig(rawConfig) {
-    this.#rawConfig = rawConfig;
-  }
-  /**
-   * Set the logger for the agent
-   * @param logger
-   */
-  __setLogger(logger) {
-    this.logger = "child" in logger && typeof logger.child === "function" ? logger.child({ component: this.component }) : logger;
-  }
+	component = RegisteredLogger.LLM;
+	logger;
+	name;
+	#rawConfig;
+	constructor({ component, name, rawConfig }) {
+		this.component = component || RegisteredLogger.LLM;
+		this.name = name;
+		this.#rawConfig = rawConfig;
+		this.logger = new ConsoleLogger({ name: `${this.component} - ${this.name}` });
+	}
+	/**
+	* Returns the raw storage configuration this primitive was created from,
+	* or undefined if it was created from code.
+	*/
+	toRawConfig() {
+		return this.#rawConfig;
+	}
+	/**
+	* Sets the raw storage configuration for this primitive.
+	* @internal
+	*/
+	__setRawConfig(rawConfig) {
+		this.#rawConfig = rawConfig;
+	}
+	/**
+	* Set the logger for the agent
+	* @param logger
+	*/
+	__setLogger(logger) {
+		this.logger = "child" in logger && typeof logger.child === "function" ? logger.child({ component: this.component }) : logger;
+	}
 };
-
-// ../../packages/_internals/voice/dist/chunk-NWNKSBZV.js
+//#endregion
+//#region ../../packages/_internals/voice/dist/aisdk-C_K-VCwq.js
 var MastraVoice = class extends MastraBase {
-  listeningModel;
-  speechModel;
-  speaker;
-  realtimeConfig;
-  constructor({ listeningModel, speechModel, speaker, realtimeConfig, name } = {}) {
-    super({
-      component: "VOICE",
-      name
-    });
-    this.listeningModel = listeningModel;
-    this.speechModel = speechModel;
-    this.speaker = speaker;
-    this.realtimeConfig = realtimeConfig;
-  }
-  /**
-   * Custom serialization for tracing/observability spans.
-   * Excludes `apiKey` from listeningModel / speechModel / realtimeConfig
-   * and any provider-specific state held by subclasses. Subclasses that
-   * need to expose additional non-sensitive fields can override.
-   */
-  serializeForSpan() {
-    return {
-      component: "VOICE",
-      name: this.name,
-      speaker: this.speaker,
-      listeningModel: this.listeningModel ? { name: this.listeningModel.name } : void 0,
-      speechModel: this.speechModel ? { name: this.speechModel.name } : void 0,
-      realtimeModel: this.realtimeConfig?.model
-    };
-  }
-  updateConfig(_options) {
-    this.logger.debug("updateConfig not implemented by this voice provider");
-  }
-  /**
-   * Initializes a WebSocket or WebRTC connection for real-time communication
-   * @returns Promise that resolves when the connection is established
-   */
-  async connect(_options) {
-    this.logger.debug("connect not implemented by this voice provider");
-  }
-  /**
-   * Relay audio data to the voice provider for real-time processing
-   * @param audioData Audio data to relay
-   */
-  async send(_audioData) {
-    this.logger.debug("relay not implemented by this voice provider");
-  }
-  /**
-   * Trigger voice providers to respond
-   */
-  async answer(_options) {
-    this.logger.debug("answer not implemented by this voice provider");
-  }
-  /**
-   * Equip the voice provider with instructions
-   * @param instructions Instructions to add
-   */
-  addInstructions(_instructions) {
-  }
-  /**
-   * Equip the voice provider with tools
-   * @param tools Array of tools to add
-   */
-  addTools(_tools) {
-  }
-  /**
-   * Disconnect from the WebSocket or WebRTC connection
-   */
-  close() {
-    this.logger.debug("close not implemented by this voice provider");
-  }
-  /**
-   * Register an event listener
-   * @param event Event name (e.g., 'speaking', 'writing', 'error')
-   * @param callback Callback function that receives event data
-   */
-  on(_event, _callback) {
-    this.logger.debug("on not implemented by this voice provider");
-  }
-  /**
-   * Remove an event listener
-   * @param event Event name (e.g., 'speaking', 'writing', 'error')
-   * @param callback Callback function to remove
-   */
-  off(_event, _callback) {
-    this.logger.debug("off not implemented by this voice provider");
-  }
-  /**
-   * Get available speakers/voices
-   * @returns Array of available voice IDs and their metadata
-   */
-  getSpeakers() {
-    this.logger.debug("getSpeakers not implemented by this voice provider");
-    return Promise.resolve([]);
-  }
-  /**
-   * Get available speakers/voices
-   * @returns Array of available voice IDs and their metadata
-   */
-  getListener() {
-    this.logger.debug("getListener not implemented by this voice provider");
-    return Promise.resolve({ enabled: false });
-  }
+	listeningModel;
+	speechModel;
+	speaker;
+	realtimeConfig;
+	constructor({ listeningModel, speechModel, speaker, realtimeConfig, name } = {}) {
+		super({
+			component: "VOICE",
+			name
+		});
+		this.listeningModel = listeningModel;
+		this.speechModel = speechModel;
+		this.speaker = speaker;
+		this.realtimeConfig = realtimeConfig;
+	}
+	/**
+	* Custom serialization for tracing/observability spans.
+	* Excludes `apiKey` from listeningModel / speechModel / realtimeConfig
+	* and any provider-specific state held by subclasses. Subclasses that
+	* need to expose additional non-sensitive fields can override.
+	*/
+	serializeForSpan() {
+		return {
+			component: "VOICE",
+			name: this.name,
+			speaker: this.speaker,
+			listeningModel: this.listeningModel ? { name: this.listeningModel.name } : void 0,
+			speechModel: this.speechModel ? { name: this.speechModel.name } : void 0,
+			realtimeModel: this.realtimeConfig?.model
+		};
+	}
+	updateConfig(_options) {
+		this.logger.debug("updateConfig not implemented by this voice provider");
+	}
+	/**
+	* Initializes a WebSocket or WebRTC connection for real-time communication
+	* @returns Promise that resolves when the connection is established
+	*/
+	async connect(_options) {
+		this.logger.debug("connect not implemented by this voice provider");
+	}
+	/**
+	* Relay audio data to the voice provider for real-time processing
+	* @param audioData Audio data to relay
+	*/
+	async send(_audioData) {
+		this.logger.debug("relay not implemented by this voice provider");
+	}
+	/**
+	* Trigger voice providers to respond
+	*/
+	async answer(_options) {
+		this.logger.debug("answer not implemented by this voice provider");
+	}
+	/**
+	* Equip the voice provider with instructions
+	* @param instructions Instructions to add
+	*/
+	addInstructions(_instructions) {}
+	/**
+	* Equip the voice provider with tools
+	* @param tools Array of tools to add
+	*/
+	addTools(_tools) {}
+	/**
+	* Disconnect from the WebSocket or WebRTC connection
+	*/
+	close() {
+		this.logger.debug("close not implemented by this voice provider");
+	}
+	/**
+	* Register an event listener
+	* @param event Event name (e.g., 'speaking', 'writing', 'error')
+	* @param callback Callback function that receives event data
+	*/
+	on(_event, _callback) {
+		this.logger.debug("on not implemented by this voice provider");
+	}
+	/**
+	* Remove an event listener
+	* @param event Event name (e.g., 'speaking', 'writing', 'error')
+	* @param callback Callback function to remove
+	*/
+	off(_event, _callback) {
+		this.logger.debug("off not implemented by this voice provider");
+	}
+	/**
+	* Get available speakers/voices
+	* @returns Array of available voice IDs and their metadata
+	*/
+	getSpeakers() {
+		this.logger.debug("getSpeakers not implemented by this voice provider");
+		return Promise.resolve([]);
+	}
+	/**
+	* Get available speakers/voices
+	* @returns Array of available voice IDs and their metadata
+	*/
+	getListener() {
+		this.logger.debug("getListener not implemented by this voice provider");
+		return Promise.resolve({ enabled: false });
+	}
 };
-
-// src/index.ts
-var resolveAuthConfig = (modelConfig, fallback, vertexConfig) => {
-  const resolved = {};
-  if (vertexConfig?.vertexAI) {
-    const projectId = vertexConfig.project || process.env.GOOGLE_CLOUD_PROJECT;
-    if (projectId) {
-      resolved.projectId = projectId;
-    }
-  }
-  const apiKey = modelConfig?.apiKey ?? fallback.apiKey;
-  if (apiKey && !vertexConfig?.vertexAI) {
-    resolved.apiKey = apiKey;
-  }
-  const keyFilename = modelConfig?.keyFilename ?? fallback.keyFilename;
-  if (keyFilename) {
-    resolved.keyFilename = keyFilename;
-  }
-  const credentials = modelConfig?.credentials ?? fallback.credentials;
-  if (credentials) {
-    resolved.credentials = credentials;
-  }
-  return resolved;
+//#endregion
+//#region src/index.ts
+const resolveAuthConfig = (modelConfig, fallback, vertexConfig) => {
+	const resolved = {};
+	if (vertexConfig?.vertexAI) {
+		const projectId = vertexConfig.project || process.env.GOOGLE_CLOUD_PROJECT;
+		if (projectId) resolved.projectId = projectId;
+	}
+	const apiKey = modelConfig?.apiKey ?? fallback.apiKey;
+	if (apiKey && !vertexConfig?.vertexAI) resolved.apiKey = apiKey;
+	const keyFilename = modelConfig?.keyFilename ?? fallback.keyFilename;
+	if (keyFilename) resolved.keyFilename = keyFilename;
+	const credentials = modelConfig?.credentials ?? fallback.credentials;
+	if (credentials) resolved.credentials = credentials;
+	return resolved;
 };
-var buildAuthOptions = (config, vertexConfig) => {
-  const options = {};
-  if (config.credentials) {
-    options.credentials = config.credentials;
-  }
-  if (config.keyFilename) {
-    options.keyFilename = config.keyFilename;
-  }
-  if (config.apiKey && !vertexConfig?.vertexAI) {
-    options.apiKey = config.apiKey;
-  }
-  if (config.projectId) {
-    options.projectId = config.projectId;
-  }
-  return options;
+const buildAuthOptions = (config, vertexConfig) => {
+	const options = {};
+	if (config.credentials) options.credentials = config.credentials;
+	if (config.keyFilename) options.keyFilename = config.keyFilename;
+	if (config.apiKey && !vertexConfig?.vertexAI) options.apiKey = config.apiKey;
+	if (config.projectId) options.projectId = config.projectId;
+	return options;
 };
-var DEFAULT_VOICE = "en-US-Casual-K";
+const DEFAULT_VOICE = "en-US-Casual-K";
+/**
+* GoogleVoice class provides Text-to-Speech and Speech-to-Text capabilities using Google Cloud services.
+* Supports both standard Google Cloud API authentication and Vertex AI mode for enterprise deployments.
+*
+* @class GoogleVoice
+* @extends MastraVoice
+*
+* @example Standard usage with API key
+* ```typescript
+* const voice = new GoogleVoice({
+*   speechModel: { apiKey: 'your-api-key' },
+*   speaker: 'en-US-Studio-O',
+* });
+* ```
+*
+* @example Vertex AI mode (recommended for production)
+* ```typescript
+* const voice = new GoogleVoice({
+*   vertexAI: true,
+*   project: 'your-gcp-project',
+*   location: 'us-central1',
+*   speaker: 'en-US-Studio-O',
+* });
+* ```
+*
+* @example Vertex AI with service account
+* ```typescript
+* const voice = new GoogleVoice({
+*   vertexAI: true,
+*   project: 'your-gcp-project',
+*   speechModel: {
+*     keyFilename: '/path/to/service-account.json',
+*   },
+* });
+* ```
+*/
 var GoogleVoice = class extends MastraVoice {
-  ttsClient;
-  speechClient;
-  speechClientV2;
-  speechOptionsV2;
-  vertexAI;
-  project;
-  location;
-  /**
-   * Creates an instance of GoogleVoice
-   * @param {GoogleVoiceConfig} config - Configuration options
-   * @param {GoogleModelConfig} [config.speechModel] - Configuration for speech synthesis
-   * @param {GoogleModelConfig} [config.listeningModel] - Configuration for speech recognition
-   * @param {string} [config.speaker] - Default voice ID to use for speech synthesis
-   * @param {boolean} [config.vertexAI] - Enable Vertex AI mode
-   * @param {string} [config.project] - Google Cloud project ID (required for Vertex AI)
-   * @param {string} [config.location] - Google Cloud region (default: 'us-central1')
-   */
-  constructor({ listeningModel, speechModel, speaker, vertexAI = false, project, location } = {}) {
-    const defaultApiKey = process.env.GOOGLE_API_KEY;
-    const defaultKeyFilename = process.env.GOOGLE_APPLICATION_CREDENTIALS;
-    const defaultSpeaker = DEFAULT_VOICE;
-    const resolvedProject = project || process.env.GOOGLE_CLOUD_PROJECT;
-    const resolvedLocation = location || process.env.GOOGLE_CLOUD_LOCATION || "us-central1";
-    if (vertexAI && !resolvedProject) {
-      throw new Error(
-        "Google Cloud project ID is required when using Vertex AI. Set GOOGLE_CLOUD_PROJECT environment variable or pass project to constructor."
-      );
-    }
-    const vertexConfig = { vertexAI, project: resolvedProject };
-    const sharedFallback = {
-      apiKey: defaultApiKey ?? speechModel?.apiKey ?? listeningModel?.apiKey,
-      keyFilename: defaultKeyFilename ?? speechModel?.keyFilename ?? listeningModel?.keyFilename,
-      credentials: speechModel?.credentials ?? listeningModel?.credentials};
-    const speechAuthConfig = resolveAuthConfig(speechModel, sharedFallback, vertexConfig);
-    const listeningAuthConfig = resolveAuthConfig(listeningModel, sharedFallback, vertexConfig);
-    super({
-      speechModel: {
-        name: "",
-        apiKey: speechAuthConfig.apiKey ?? defaultApiKey
-      },
-      listeningModel: {
-        name: "",
-        apiKey: listeningAuthConfig.apiKey ?? defaultApiKey
-      },
-      speaker: speaker ?? defaultSpeaker
-    });
-    this.vertexAI = vertexAI;
-    this.project = resolvedProject;
-    this.location = resolvedLocation;
-    const ttsOptions = buildAuthOptions(speechAuthConfig, { vertexAI});
-    const speechOptions = buildAuthOptions(listeningAuthConfig, { vertexAI});
-    this.ttsClient = new TextToSpeechClient(ttsOptions);
-    this.speechClient = new SpeechClient(speechOptions);
-    this.speechOptionsV2 = speechOptions;
-  }
-  getV2SpeechClient() {
-    if (!this.speechClientV2) {
-      this.speechClientV2 = new v2.SpeechClient(this.speechOptionsV2);
-    }
-    return this.speechClientV2;
-  }
-  /**
-   * Check if Vertex AI mode is enabled
-   * @returns {boolean} True if using Vertex AI
-   */
-  isUsingVertexAI() {
-    return this.vertexAI;
-  }
-  /**
-   * Get the configured Google Cloud project ID
-   * @returns {string | undefined} The project ID or undefined if not set
-   */
-  getProject() {
-    return this.project;
-  }
-  /**
-   * Get the configured Google Cloud location/region
-   * @returns {string} The location (default: 'us-central1')
-   */
-  getLocation() {
-    return this.location;
-  }
-  /**
-   * Gets a list of available voices
-   * @returns {Promise<Array<{voiceId: string, languageCodes: string[]}>>} List of available voices and their supported languages. Default language is en-US.
-   */
-  async getSpeakers({ languageCode = "en-US" } = {}) {
-    const [response] = await this.ttsClient.listVoices({ languageCode });
-    return (response?.voices || []).filter((voice) => voice.name && voice.languageCodes).map((voice) => ({
-      voiceId: voice.name,
-      languageCodes: voice.languageCodes
-    }));
-  }
-  async streamToString(stream) {
-    const chunks = [];
-    for await (const chunk of stream) {
-      if (typeof chunk === "string") {
-        chunks.push(Buffer.from(chunk));
-      } else {
-        chunks.push(chunk);
-      }
-    }
-    return Buffer.concat(chunks).toString("utf-8");
-  }
-  /**
-   * Converts text to speech.
-   *
-   * When `input` is a string or stream, builds a text-only request (existing behaviour).
-   * Pass `options.input` to send richer proto fields (ssml, markup, prompt,
-   * customPronunciations, multiSpeakerMarkup) and `options.voice` for fields
-   * like modelName or multiSpeakerVoiceConfig.
-   *
-   * @param {string | NodeJS.ReadableStream} input - Text or stream to convert to speech
-   * @param {GoogleSpeakOptions} [options] - Speech synthesis options
-   * @returns {Promise<NodeJS.ReadableStream>} Stream of synthesised audio. Default encoding is LINEAR16.
-   */
-  async speak(input, options) {
-    const defaultVoiceName = options?.speaker || this.speaker;
-    const defaultLanguageCode = options?.languageCode || defaultVoiceName?.split("-").slice(0, 2).join("-") || "en-US";
-    const requestInput = options?.input ? { ...options.input } : { text: typeof input === "string" ? input : await this.streamToString(input) };
-    if (options?.input && !options.input.text && !options.input.ssml && !options.input.markup && !options.input.multiSpeakerMarkup) {
-      requestInput.text = typeof input === "string" ? input : await this.streamToString(input);
-    }
-    const request = {
-      input: requestInput,
-      voice: {
-        name: defaultVoiceName,
-        languageCode: defaultLanguageCode,
-        ...options?.voice
-      },
-      audioConfig: options?.audioConfig || { audioEncoding: "LINEAR16" }
-    };
-    const [response] = await this.ttsClient.synthesizeSpeech(request);
-    if (!response.audioContent) {
-      throw new Error("No audio content returned.");
-    }
-    if (typeof response.audioContent === "string") {
-      throw new Error("Audio content is a string.");
-    }
-    const stream = new PassThrough();
-    stream.end(Buffer.from(response.audioContent));
-    return stream;
-  }
-  /**
-   * Checks if listening capabilities are enabled.
-   *
-   * @returns {Promise<{ enabled: boolean }>}
-   */
-  async getListener() {
-    return { enabled: true };
-  }
-  /**
-   * Converts speech to text using Cloud Speech-to-Text v1 or v2.
-   *
-   * Pass `{ v2: true }` in options to use the v2 API, which supports additional
-   * audio formats like AAC-in-MP4 (iOS Safari) via `autoDecodingConfig` or
-   * `explicitDecodingConfig`. The v1 path remains the default.
-   *
-   * @param {NodeJS.ReadableStream} audioStream - Audio stream to transcribe. Default encoding is LINEAR16.
-   * @param {GoogleListenOptions} [options] - Recognition options
-   * @returns {Promise<string>} Transcribed text
-   */
-  async listen(audioStream, options) {
-    const chunks = [];
-    for await (const chunk of audioStream) {
-      if (typeof chunk === "string") {
-        chunks.push(Buffer.from(chunk));
-      } else {
-        chunks.push(chunk);
-      }
-    }
-    const buffer = Buffer.concat(chunks);
-    if (options && "v2" in options && options.v2) {
-      return this.recognizeV2(buffer, options);
-    }
-    return this.recognizeV1(buffer, options);
-  }
-  async recognizeV1(buffer, options) {
-    const request = {
-      config: {
-        encoding: "LINEAR16",
-        languageCode: "en-US",
-        ...options?.config
-      },
-      audio: {
-        content: buffer.toString("base64")
-      }
-    };
-    const [response] = await this.speechClient.recognize(request);
-    return this.extractTranscription(response?.results);
-  }
-  async recognizeV2(buffer, options) {
-    const config = { ...options.config };
-    if (!config.autoDecodingConfig && !config.explicitDecodingConfig) {
-      config.autoDecodingConfig = {};
-    }
-    if (!config.languageCodes || config.languageCodes.length === 0) {
-      config.languageCodes = ["en-US"];
-    }
-    if (!config.model) {
-      config.model = "long";
-    }
-    let recognizer = options.recognizer;
-    if (!recognizer) {
-      const project = this.project || await this.getV2SpeechClient().getProjectId();
-      recognizer = `projects/${project}/locations/global/recognizers/_`;
-    }
-    const request = {
-      recognizer,
-      config,
-      content: buffer
-    };
-    const client = this.getV2SpeechClient();
-    const [response] = await client.recognize(request);
-    return this.extractTranscription(response?.results);
-  }
-  extractTranscription(results) {
-    if (!results || results.length === 0) {
-      throw new Error("No transcription results returned");
-    }
-    const transcription = results.map((result) => {
-      if (!result.alternatives || result.alternatives.length === 0) {
-        return "";
-      }
-      return result.alternatives[0].transcript || "";
-    }).filter((text) => text.length > 0).join(" ");
-    if (!transcription) {
-      throw new Error("No valid transcription found in results");
-    }
-    return transcription;
-  }
+	ttsClient;
+	speechClient;
+	speechClientV2;
+	speechOptionsV2;
+	vertexAI;
+	project;
+	location;
+	/**
+	* Creates an instance of GoogleVoice
+	* @param {GoogleVoiceConfig} config - Configuration options
+	* @param {GoogleModelConfig} [config.speechModel] - Configuration for speech synthesis
+	* @param {GoogleModelConfig} [config.listeningModel] - Configuration for speech recognition
+	* @param {string} [config.speaker] - Default voice ID to use for speech synthesis
+	* @param {boolean} [config.vertexAI] - Enable Vertex AI mode
+	* @param {string} [config.project] - Google Cloud project ID (required for Vertex AI)
+	* @param {string} [config.location] - Google Cloud region (default: 'us-central1')
+	*/
+	constructor({ listeningModel, speechModel, speaker, vertexAI = false, project, location } = {}) {
+		const defaultApiKey = process.env.GOOGLE_API_KEY;
+		const defaultKeyFilename = process.env.GOOGLE_APPLICATION_CREDENTIALS;
+		const defaultSpeaker = DEFAULT_VOICE;
+		const resolvedProject = project || process.env.GOOGLE_CLOUD_PROJECT;
+		const resolvedLocation = location || process.env.GOOGLE_CLOUD_LOCATION || "us-central1";
+		if (vertexAI && !resolvedProject) throw new Error("Google Cloud project ID is required when using Vertex AI. Set GOOGLE_CLOUD_PROJECT environment variable or pass project to constructor.");
+		const vertexConfig = {
+			vertexAI,
+			project: resolvedProject
+		};
+		const sharedFallback = {
+			apiKey: defaultApiKey ?? speechModel?.apiKey ?? listeningModel?.apiKey,
+			keyFilename: defaultKeyFilename ?? speechModel?.keyFilename ?? listeningModel?.keyFilename,
+			credentials: speechModel?.credentials ?? listeningModel?.credentials,
+			projectId: resolvedProject
+		};
+		const speechAuthConfig = resolveAuthConfig(speechModel, sharedFallback, vertexConfig);
+		const listeningAuthConfig = resolveAuthConfig(listeningModel, sharedFallback, vertexConfig);
+		super({
+			speechModel: {
+				name: "",
+				apiKey: speechAuthConfig.apiKey ?? defaultApiKey
+			},
+			listeningModel: {
+				name: "",
+				apiKey: listeningAuthConfig.apiKey ?? defaultApiKey
+			},
+			speaker: speaker ?? defaultSpeaker
+		});
+		this.vertexAI = vertexAI;
+		this.project = resolvedProject;
+		this.location = resolvedLocation;
+		const ttsOptions = buildAuthOptions(speechAuthConfig, {
+			vertexAI,
+			location: resolvedLocation
+		});
+		const speechOptions = buildAuthOptions(listeningAuthConfig, {
+			vertexAI,
+			location: resolvedLocation
+		});
+		this.ttsClient = new TextToSpeechClient(ttsOptions);
+		this.speechClient = new SpeechClient(speechOptions);
+		this.speechOptionsV2 = speechOptions;
+	}
+	getV2SpeechClient() {
+		if (!this.speechClientV2) this.speechClientV2 = new v2.SpeechClient(this.speechOptionsV2);
+		return this.speechClientV2;
+	}
+	/**
+	* Check if Vertex AI mode is enabled
+	* @returns {boolean} True if using Vertex AI
+	*/
+	isUsingVertexAI() {
+		return this.vertexAI;
+	}
+	/**
+	* Get the configured Google Cloud project ID
+	* @returns {string | undefined} The project ID or undefined if not set
+	*/
+	getProject() {
+		return this.project;
+	}
+	/**
+	* Get the configured Google Cloud location/region
+	* @returns {string} The location (default: 'us-central1')
+	*/
+	getLocation() {
+		return this.location;
+	}
+	/**
+	* Gets a list of available voices
+	* @returns {Promise<Array<{voiceId: string, languageCodes: string[]}>>} List of available voices and their supported languages. Default language is en-US.
+	*/
+	async getSpeakers({ languageCode = "en-US" } = {}) {
+		const [response] = await this.ttsClient.listVoices({ languageCode });
+		return (response?.voices || []).filter((voice) => voice.name && voice.languageCodes).map((voice) => ({
+			voiceId: voice.name,
+			languageCodes: voice.languageCodes
+		}));
+	}
+	async streamToString(stream) {
+		const chunks = [];
+		for await (const chunk of stream) if (typeof chunk === "string") chunks.push(Buffer.from(chunk));
+		else chunks.push(chunk);
+		return Buffer.concat(chunks).toString("utf-8");
+	}
+	/**
+	* Converts text to speech.
+	*
+	* When `input` is a string or stream, builds a text-only request (existing behaviour).
+	* Pass `options.input` to send richer proto fields (ssml, markup, prompt,
+	* customPronunciations, multiSpeakerMarkup) and `options.voice` for fields
+	* like modelName or multiSpeakerVoiceConfig.
+	*
+	* @param {string | NodeJS.ReadableStream} input - Text or stream to convert to speech
+	* @param {GoogleSpeakOptions} [options] - Speech synthesis options
+	* @returns {Promise<NodeJS.ReadableStream>} Stream of synthesised audio. Default encoding is LINEAR16.
+	*/
+	async speak(input, options) {
+		const defaultVoiceName = options?.speaker || this.speaker;
+		const defaultLanguageCode = options?.languageCode || defaultVoiceName?.split("-").slice(0, 2).join("-") || "en-US";
+		const requestInput = options?.input ? { ...options.input } : { text: typeof input === "string" ? input : await this.streamToString(input) };
+		if (options?.input && !options.input.text && !options.input.ssml && !options.input.markup && !options.input.multiSpeakerMarkup) requestInput.text = typeof input === "string" ? input : await this.streamToString(input);
+		const request = {
+			input: requestInput,
+			voice: {
+				name: defaultVoiceName,
+				languageCode: defaultLanguageCode,
+				...options?.voice
+			},
+			audioConfig: options?.audioConfig || { audioEncoding: "LINEAR16" }
+		};
+		const [response] = await this.ttsClient.synthesizeSpeech(request);
+		if (!response.audioContent) throw new Error("No audio content returned.");
+		if (typeof response.audioContent === "string") throw new Error("Audio content is a string.");
+		const stream = new PassThrough();
+		stream.end(Buffer.from(response.audioContent));
+		return stream;
+	}
+	/**
+	* Checks if listening capabilities are enabled.
+	*
+	* @returns {Promise<{ enabled: boolean }>}
+	*/
+	async getListener() {
+		return { enabled: true };
+	}
+	/**
+	* Converts speech to text using Cloud Speech-to-Text v1 or v2.
+	*
+	* Pass `{ v2: true }` in options to use the v2 API, which supports additional
+	* audio formats like AAC-in-MP4 (iOS Safari) via `autoDecodingConfig` or
+	* `explicitDecodingConfig`. The v1 path remains the default.
+	*
+	* @param {NodeJS.ReadableStream} audioStream - Audio stream to transcribe. Default encoding is LINEAR16.
+	* @param {GoogleListenOptions} [options] - Recognition options
+	* @returns {Promise<string>} Transcribed text
+	*/
+	async listen(audioStream, options) {
+		const chunks = [];
+		for await (const chunk of audioStream) if (typeof chunk === "string") chunks.push(Buffer.from(chunk));
+		else chunks.push(chunk);
+		const buffer = Buffer.concat(chunks);
+		if (options && "v2" in options && options.v2) return this.recognizeV2(buffer, options);
+		return this.recognizeV1(buffer, options);
+	}
+	async recognizeV1(buffer, options) {
+		const request = {
+			config: {
+				encoding: "LINEAR16",
+				languageCode: "en-US",
+				...options?.config
+			},
+			audio: { content: buffer.toString("base64") }
+		};
+		const [response] = await this.speechClient.recognize(request);
+		return this.extractTranscription(response?.results);
+	}
+	async recognizeV2(buffer, options) {
+		const config = { ...options.config };
+		if (!config.autoDecodingConfig && !config.explicitDecodingConfig) config.autoDecodingConfig = {};
+		if (!config.languageCodes || config.languageCodes.length === 0) config.languageCodes = ["en-US"];
+		if (!config.model) config.model = "long";
+		let recognizer = options.recognizer;
+		if (!recognizer) recognizer = `projects/${this.project || await this.getV2SpeechClient().getProjectId()}/locations/global/recognizers/_`;
+		const request = {
+			recognizer,
+			config,
+			content: buffer
+		};
+		const [response] = await this.getV2SpeechClient().recognize(request);
+		return this.extractTranscription(response?.results);
+	}
+	extractTranscription(results) {
+		if (!results || results.length === 0) throw new Error("No transcription results returned");
+		const transcription = results.map((result) => {
+			if (!result.alternatives || result.alternatives.length === 0) return "";
+			return result.alternatives[0].transcript || "";
+		}).filter((text) => text.length > 0).join(" ");
+		if (!transcription) throw new Error("No valid transcription found in results");
+		return transcription;
+	}
 };
-
+//#endregion
 export { GoogleVoice };
-//# sourceMappingURL=index.js.map
+
 //# sourceMappingURL=index.js.map