@amaster.ai/client 1.1.0-beta.50 → 1.1.0-beta.51

package/dist/index.d.cts

@@ -7,7 +7,7 @@ export { ActivityInstanceTree, CamundaVariable, HistoryActivityInstance, History
 import { WorkflowClient } from '@amaster.ai/workflow-client';
 export { WorkflowFile, WorkflowInputValue, WorkflowRunRequest, WorkflowRunResponse } from '@amaster.ai/workflow-client';
 import { ASRClientConfig, ASRClient, ASRHttpClientConfig, ASRHttpClient } from '@amaster.ai/asr-client';
-export { ASRClient, ASRClientConfig, ASRHttpClient, ASRHttpClientConfig } from '@amaster.ai/asr-client';
+export { ASRClient, ASRClientConfig, ASRHttpClient, ASRHttpClientConfig, ASRLanguage } from '@amaster.ai/asr-client';
 import { CopilotClient } from '@amaster.ai/copilot-client';
 export { ChatMessage, ChatOptions, CopilotClient, FileContent, ImageContent, MessageContent, TextContent } from '@amaster.ai/copilot-client';
 import { FunctionClient } from '@amaster.ai/function-client';

package/dist/index.d.ts

@@ -7,7 +7,7 @@ export { ActivityInstanceTree, CamundaVariable, HistoryActivityInstance, History
 import { WorkflowClient } from '@amaster.ai/workflow-client';
 export { WorkflowFile, WorkflowInputValue, WorkflowRunRequest, WorkflowRunResponse } from '@amaster.ai/workflow-client';
 import { ASRClientConfig, ASRClient, ASRHttpClientConfig, ASRHttpClient } from '@amaster.ai/asr-client';
-export { ASRClient, ASRClientConfig, ASRHttpClient, ASRHttpClientConfig } from '@amaster.ai/asr-client';
+export { ASRClient, ASRClientConfig, ASRHttpClient, ASRHttpClientConfig, ASRLanguage } from '@amaster.ai/asr-client';
 import { CopilotClient } from '@amaster.ai/copilot-client';
 export { ChatMessage, ChatOptions, CopilotClient, FileContent, ImageContent, MessageContent, TextContent } from '@amaster.ai/copilot-client';
 import { FunctionClient } from '@amaster.ai/function-client';

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@amaster.ai/client",
-  "version": "1.1.0-beta.50",
+  "version": "1.1.0-beta.51",
   "description": "Unified API client for Amaster platform - All services in one package",
   "type": "module",
   "main": "./dist/index.cjs",
@@ -72,16 +72,16 @@
     "registry": "https://registry.npmjs.org/"
   },
   "dependencies": {
-    "@amaster.ai/asr-client": "1.1.0-beta.50",
-    "@amaster.ai/copilot-client": "1.1.0-beta.50",
-    "@amaster.ai/auth-client": "1.1.0-beta.50",
-    "@amaster.ai/entity-client": "1.1.0-beta.50",
-    "@amaster.ai/function-client": "1.1.0-beta.50",
-    "@amaster.ai/http-client": "1.1.0-beta.50",
-    "@amaster.ai/s3-client": "1.1.0-beta.50",
-    "@amaster.ai/bpm-client": "1.1.0-beta.50",
-    "@amaster.ai/tts-client": "1.1.0-beta.50",
-    "@amaster.ai/workflow-client": "1.1.0-beta.50"
+    "@amaster.ai/copilot-client": "1.1.0-beta.51",
+    "@amaster.ai/auth-client": "1.1.0-beta.51",
+    "@amaster.ai/bpm-client": "1.1.0-beta.51",
+    "@amaster.ai/asr-client": "1.1.0-beta.51",
+    "@amaster.ai/s3-client": "1.1.0-beta.51",
+    "@amaster.ai/http-client": "1.1.0-beta.51",
+    "@amaster.ai/function-client": "1.1.0-beta.51",
+    "@amaster.ai/tts-client": "1.1.0-beta.51",
+    "@amaster.ai/workflow-client": "1.1.0-beta.51",
+    "@amaster.ai/entity-client": "1.1.0-beta.51"
   },
   "peerDependencies": {
     "axios": "^1.11.0"

package/types/asr.d.ts

@@ -1,88 +1,211 @@
 /**
- *  * Real-time speech recognition using WebSocket connection.
- * 
- * @module asr
- */
-
-/**
- * ASR Client Configuration
+ * ASR Realtime WebSocket Client for Qwen-ASR Realtime API
+ *
+ * WebSocket-based real-time speech recognition for streaming transcription.
+ * Follows the Qwen-ASR Realtime API protocol with proper event handling.
  *
+ * @example
+ * ```typescript
+ * const asrClient = client.asr({
+ *   language: "zh",
+ *   enableVAD: true,
+ *   onReady() {
+ *     console.log("ASR connected");
+ *   },
+ *   onTranscript(text, isFinal) {
+ *     console.log(isFinal ? "[Final]" : "[Interim]", text);
+ *   },
+ *   onError(err) {
+ *     console.error("ASR error:", err);
+ *   },
+ * });
+ *
+ * await asrClient.connect();
+ * await asrClient.startRecording();
+ * // ... stop ...
+ * await asrClient.stopRecording();
+ * await asrClient.close();
+ * ```
  */
+export type ASRLanguage = "zh" | "yue" | "en" | "ja" | "de" | "ko" | "ru" | "fr" | "pt" | "ar" | "it" | "es" | "hi" | "id" | "th" | "tr" | "uk" | "vi" | "cs" | "da" | "fil" | "fi" | "is" | "ms" | "no" | "pl" | "sv";
+export type ClientEventType = "session.update" | "input_audio_buffer.append" | "input_audio_buffer.commit" | "session.finish";
+export type ServerEventType = "session.created" | "session.updated" | "input_audio_buffer.speech_started" | "input_audio_buffer.speech_stopped" | "input_audio_buffer.committed" | "conversation.item.input_audio_transcription.text" | "conversation.item.input_audio_transcription.completed" | "session.finished" | "error";
+export interface BaseEvent {
+    event_id: string;
+    type: ClientEventType | ServerEventType;
+}
+export interface SessionUpdateEvent extends BaseEvent {
+    type: "session.update";
+    session: SessionConfig;
+}
+export interface InputAudioBufferAppendEvent extends BaseEvent {
+    type: "input_audio_buffer.append";
+    audio: string;
+}
+export interface InputAudioBufferCommitEvent extends BaseEvent {
+    type: "input_audio_buffer.commit";
+}
+export interface SessionFinishEvent extends BaseEvent {
+    type: "session.finish";
+}
+type ClientEvent = SessionUpdateEvent | InputAudioBufferAppendEvent | InputAudioBufferCommitEvent | SessionFinishEvent;
+export interface SessionCreatedEvent extends BaseEvent {
+    type: "session.created";
+    session: {
+        id: string;
+    };
+}
+export interface SessionUpdatedEvent extends BaseEvent {
+    type: "session.updated";
+    session: SessionConfig;
+}
+export interface SpeechStartedEvent extends BaseEvent {
+    type: "input_audio_buffer.speech_started";
+}
+export interface SpeechStoppedEvent extends BaseEvent {
+    type: "input_audio_buffer.speech_stopped";
+}
+export interface InputAudioBufferCommittedEvent extends BaseEvent {
+    type: "input_audio_buffer.committed";
+}
+export interface TranscriptionTextEvent extends BaseEvent {
+    type: "conversation.item.input_audio_transcription.text";
+    text?: string;
+    stash?: string;
+    transcript?: string;
+}
+export interface TranscriptionCompletedEvent extends BaseEvent {
+    type: "conversation.item.input_audio_transcription.completed";
+    text?: string;
+    transcript?: string;
+}
+export interface SessionFinishedEvent extends BaseEvent {
+    type: "session.finished";
+}
+export interface ErrorEvent extends BaseEvent {
+    type: "error";
+    error: {
+        message: string;
+        code?: string;
+    };
+}
+export type ServerEvent = SessionCreatedEvent | SessionUpdatedEvent | SpeechStartedEvent | SpeechStoppedEvent | InputAudioBufferCommittedEvent | TranscriptionTextEvent | TranscriptionCompletedEvent | SessionFinishedEvent | ErrorEvent;
+export interface TurnDetectionConfig {
+    type: "server_vad";
+    /** VAD检测阈值，推荐设为 0.0，默认值 0.2，范围 [-1, 1] */
+    threshold?: number;
+    /** VAD断句检测阈值(ms)，推荐设为 400，默认值 800，范围 [200, 6000] */
+    silence_duration_ms?: number;
+}
+export interface InputAudioTranscriptionConfig {
+    language?: ASRLanguage;
+}
+export interface SessionConfig {
+    input_audio_format?: "pcm" | "opus";
+    sample_rate?: 16000 | 8000;
+    input_audio_transcription?: InputAudioTranscriptionConfig;
+    turn_detection?: TurnDetectionConfig | null;
+}
 export interface ASRClientConfig {
-  /** Audio format, default 'pcm16' */
-  audioFormat?: "pcm16" | "g711a" | "g711u";
-  
-  /** Sample rate, default 16000 */
-  sampleRate?: number;
-  
-  /** Called when connection is ready */
-  onReady?: () => void;
-  
-  /** Called when speech is detected */
-  onSpeechStart?: () => void;
-  
-  /** Called when speech stops */
-  onSpeechEnd?: () => void;
-  
-  /** 
-   * Called on transcript result
-   * @param text - Recognized text
-   * @param isFinal - Whether this is the final result
-   */
-  onTranscript?: (text: string, isFinal: boolean) => void;
-  
-  /** Called on error */
-  onError?: (error: Error) => void;
-  
-  /** Called on connection close */
-  onClose?: () => void;
+    /**
+     * Audio format
+     * @default "pcm"
+     */
+    audioFormat?: "pcm" | "opus";
+    /**
+     * Sample rate in Hz
+     * @default 16000
+     * @description 支持 16000 和 8000。设置为 8000 时，服务端会先升采样到16000Hz再进行识别，可能引入微小延迟。
+     */
+    sampleRate?: 16000 | 8000;
+    /**
+     * Audio source language
+     * @default "zh"
+     * @description 支持多种语言，包括 zh（中文）、yue（粤语）、en（英文）、ja（日语）等
+     */
+    language?: ASRLanguage;
+    /**
+     * Enable VAD (Voice Activity Detection) mode
+     * @default true
+     * @description true = VAD模式（服务端自动检测语音开始/结束），false = Manual模式（客户端手动控制）
+     */
+    enableVAD?: boolean;
+    /**
+     * VAD detection threshold
+     * @default 0.2
+     * @description 推荐设为 0.0。取值范围 [-1, 1]。较低的阈值会提高 VAD 的灵敏度。
+     */
+    vadThreshold?: number;
+    /**
+     * VAD silence duration threshold in milliseconds
+     * @default 800
+     * @description 推荐设为 400。取值范围 [200, 6000]。静音持续时长超过该阈值将被认为是语句结束。
+     */
+    vadSilenceDurationMs?: number;
+    /**
+     * Get access token for WebSocket authentication
+     */
+    getAccessToken?: () => string | null;
+    /**
+     * Called when connection is ready (session.created received and session.update sent)
+     */
+    onReady?: () => void;
+    /**
+     * Called when speech is detected (VAD mode only)
+     */
+    onSpeechStart?: () => void;
+    /**
+     * Called when speech stops (VAD mode only)
+     */
+    onSpeechEnd?: () => void;
+    /**
+     * Called on transcript result
+     * @param text - Transcribed text
+     * @param isFinal - Whether this is the final result
+     */
+    onTranscript?: (text: string, isFinal: boolean) => void;
+    /**
+     * Called when audio buffer is committed (non-VAD mode only)
+     */
+    onAudioBufferCommitted?: () => void;
+    /**
+     * Called when session is finished
+     */
+    onSessionFinished?: () => void;
+    /**
+     * Called on error
+     */
+    onError?: (error: Error) => void;
+    /**
+     * Called on close
+     */
+    onClose?: () => void;
 }
-
-/**
- * ASR (Automatic Speech Recognition) Client API
- * 
- * Provides real-time speech-to-text conversion via WebSocket.
- * 
- * @since 1.0.0
- */
-export interface ASRClientAPI {
-  /**
-   * Connect to ASR service
-   * 
-   * Establishes WebSocket connection to the speech recognition service.
-   * 
-   * @returns Promise that resolves when connected
-   *
-   */
-  connect(): Promise<void>;
-
-  /**
-   * Start recording from microphone
-   * 
-   * Begins capturing audio from the user's microphone and sends it to the ASR service.
-   * Requires microphone permission from the user.
-   * 
-   * @returns Promise that resolves when recording starts
-   *
-   */
-  startRecording(): Promise<void>;
-
-  /**
-   * Stop recording
-   * 
-   * Stops capturing audio from the microphone but keeps the WebSocket connection open.
-   *
-   */
-  stopRecording(): void;
-
-  /**
-   * Close connection
-   * 
-   * Closes the WebSocket connection and releases resources.
-   *
-   */
-  close(): void;
+export interface ASRClient {
+    /** Connect to ASR service and establish session */
+    connect(): Promise<void>;
+    /** Start recording from microphone */
+    startRecording(): Promise<void>;
+    /**
+     * Stop recording
+     * @description In non-VAD mode, this triggers recognition by sending input_audio_buffer.commit
+     */
+    stopRecording(): Promise<void>;
+    /**
+     * Close connection gracefully
+     * @description Sends session.finish and waits for session.finished before closing
+     */
+    close(): Promise<void>;
+    /**
+     * Check if currently recording
+     */
+    isRecording(): boolean;
+    /**
+     * Check if connected to server
+     */
+    isConnected(): boolean;
 }
+declare const _default$1: (authConfig: Pick<ASRClientConfig, "getAccessToken">) => (config: ASRClientConfig) => ASRClient;
 
 export interface ASRHttpClientConfig {
     /** Get access token */

package/types/index.d.ts

@@ -1,40 +1,40 @@
 /**
  *  * A Supabase-inspired unified client for the Amaster platform.
- * 
+ *
  * ## Features
  * - Single client instance for all services (auth, entity, bpm, workflow)
  * - Automatic token management and refresh
  * - Auto-attach authentication to all requests
  * - Centralized error handling
- * 
+ *
  * ## Quick Start
  * ```typescript
  * import { createClient } from '@amaster.ai/client';
- * 
+ *
  * const client = createClient({
  *   baseURL: 'https://api.amaster.ai',
  *   onUnauthorized: () => window.location.href = '/login'
  * });
- * 
+ *
  * // Login
  * await client.auth.login({ email, password });
- * 
+ *
  * // All subsequent requests automatically include auth token
  * await client.entity.list('default', 'users');
  * await client.bpm.startProcess('approval', {});
  * ```
- * 
+ *
  * ## Module Documentation
  * For detailed API documentation, see individual module type definitions:
  * - **Authentication**: {@link ./auth/index.d.ts} (split into user, password-auth, code-auth, oauth, permissions, profile)
  * - **Entity Operations**: {@link ./entity.d.ts}
  * - **BPM (Business Process)**: {@link ./bpm.d.ts}
  * - **Workflow Execution**: {@link ./workflow.d.ts}
- * 
+ *
  * ## AI-Friendly Type Structure
  * This package provides modular type definitions optimized for AI tools and large language models.
  * Types are split into focused modules that can be loaded on-demand:
- * 
+ *
  * **Authentication Types** (70-88% token savings):
  * - `@amaster.ai/client/auth/user` - User types (100 lines)
  * - `@amaster.ai/client/auth/password-auth` - Login/register (200 lines)
@@ -44,15 +44,15 @@
  * - `@amaster.ai/client/auth/profile` - Profile management (100 lines)
  * */
 
-import type { AuthClientAPI } from './auth/index';
-import type { EntityClientAPI } from './entity';
-import type { BpmClientAPI } from './bpm';
-import type { WorkflowClientAPI } from './workflow';
-import type { ASRClientAPI } from './asr';
-import type { CopilotClientAPI } from './copilot';
-import type { FunctionClientAPI } from './function';
-import type { TTSClientAPI } from './tts';
-import type { S3ClientAPI } from './s3';
+import type { AuthClientAPI } from "./auth/index";
+import type { EntityClientAPI } from "./entity";
+import type { BpmClientAPI } from "./bpm";
+import type { WorkflowClientAPI } from "./workflow";
+import type { ASRClientConfig, ASRClient, ASRHttpClientConfig, ASRHttpClient } from "./asr";
+import type { CopilotClientAPI } from "./copilot";
+import type { FunctionClientAPI } from "./function";
+import type { TTSClientAPI } from "./tts";
+import type { S3ClientAPI } from "./s3";
 
 /**
  * Configuration options for creating an Amaster client
@@ -67,7 +67,7 @@ export interface AmasterClientOptions {
 
   /**
    * Optional custom headers to include in ALL requests
-   * 
+   *
    * These headers will be merged with authentication headers.
    * Useful for tenant IDs, API keys, or application metadata.
    *
@@ -76,7 +76,7 @@ export interface AmasterClientOptions {
 
   /**
    * Callback triggered when a 401 Unauthorized response is received
-   * 
+   *
    * Use this to redirect to login page or show authentication modal.
    * This callback is invoked BEFORE the request promise rejects.
    *
@@ -85,7 +85,7 @@ export interface AmasterClientOptions {
 
   /**
    * Callback triggered when the access token expires
-   * 
+   *
    * Note: Token refresh is handled automatically by the client.
    * This callback is for additional actions like logging or analytics.
    *
@@ -94,31 +94,31 @@ export interface AmasterClientOptions {
 
   /**
    * Enable automatic token refresh (default: true)
-   * 
+   *
    * When enabled, tokens are automatically refreshed before expiry.
-   * 
+   *
    * @default true
    */
   autoRefresh?: boolean;
 
   /**
    * Token refresh threshold in seconds (default: 300)
-   * 
+   *
    * Tokens will be refreshed this many seconds before expiry.
    * For example, if set to 300 (5 minutes), a token expiring at
    * 10:00:00 will be refreshed at 9:55:00.
-   * 
+   *
    * @default 300
    */
   refreshThreshold?: number;
 
   /**
    * Auto-handle OAuth callback on page load (default: true)
-   * 
+   *
    * Automatically detects and processes `#access_token` in URL hash,
    * then clears the hash for security. Set to `false` to manually
    * call `client.auth.handleOAuthCallback()`.
-   * 
+   *
    * @default true
    */
   autoHandleOAuthCallback?: boolean;
@@ -126,16 +126,16 @@ export interface AmasterClientOptions {
 
 /**
  * Unified Amaster Client interface
- * 
+ *
  * This is the main client object returned by `createClient()`.
  * It provides access to all Amaster services through a unified interface.
- * 
+ *
  * ## Core Modules
  * - {@link auth} - Authentication and user management
  * - {@link entity} - CRUD operations for entities
  * - {@link bpm} - Business process management
  * - {@link workflow} - Workflow execution
- * 
+ *
  * ## Token Management
  * - {@link isAuthenticated} - Check if user is authenticated
  * - {@link getAccessToken} - Get current access token
@@ -146,9 +146,9 @@ export interface AmasterClientOptions {
 export interface AmasterClient {
   /**
    * Authentication module
-   * 
+   *
    * Provides methods for user authentication, registration, and management.
-   * 
+   *
    * For detailed documentation, see {@link ./auth.d.ts}
    *
    */
@@ -156,9 +156,9 @@ export interface AmasterClient {
 
   /**
    * Entity CRUD operations module
-   * 
+   *
    * Provides methods for creating, reading, updating, and deleting entities.
-   * 
+   *
    * For detailed documentation, see {@link ./entity.d.ts}
    *
    */
@@ -166,9 +166,9 @@ export interface AmasterClient {
 
   /**
    * Business Process Management (BPM) module
-   * 
+   *
    * Provides methods for managing BPMN processes and tasks.
-   * 
+   *
    * For detailed documentation, see {@link ./bpm.d.ts}
    *
    */
@@ -176,9 +176,9 @@ export interface AmasterClient {
 
   /**
    * Workflow execution module
-   * 
+   *
    * Provides methods for executing Dify-style workflows.
-   * 
+   *
    * For detailed documentation, see {@link ./workflow.d.ts}
    *
    */
@@ -186,18 +186,27 @@ export interface AmasterClient {
 
   /**
    * ASR (Automatic Speech Recognition) module
-   * 
+   *
    * Provides methods for real-time speech-to-text conversion via WebSocket.
-   * 
+   *
    * For detailed documentation, see {@link ./asr.d.ts}
    */
-  asr: ASRClientAPI;
+  asr: (config: ASRClientConfig) => ASRClient;
+
+  /**
+   * ASR HTTP module
+   *
+   * Provides methods for speech-to-text conversion via HTTP API.
+   *
+   * For detailed documentation, see
+   */
+  asrHttp: (config: ASRHttpClientConfig) => ASRHttpClient;
 
   /**
    * Copilot AI Assistant module
-   * 
+   *
    * Provides methods for interactive AI conversations with a2a streaming.
-   * 
+   *
    * For detailed documentation, see {@link ./copilot.d.ts}
    *
    */
@@ -205,34 +214,34 @@ export interface AmasterClient {
 
   /**
    * Serverless Function module
-   * 
+   *
    * Provides methods for invoking serverless functions.
-   * 
+   *
    * For detailed documentation, see {@link ./function.d.ts}
    */
   function: FunctionClientAPI;
 
   /**
    * TTS (Text-to-Speech) module
-   * 
+   *
    * Provides methods for real-time text-to-speech conversion via WebSocket.
-   * 
+   *
    * For detailed documentation, see {@link ./tts.d.ts}
    */
   tts: TTSClientAPI;
 
   /**
    * S3 Storage module
-   * 
+   *
    * Provides methods for file upload, download, and management.
-   * 
+   *
    * For detailed documentation, see {@link ./s3.d.ts}
    */
   s3: S3ClientAPI;
 
   /**
    * Check if the user is currently authenticated
-   * 
+   *
    * @returns `true` if a valid access token exists, `false` otherwise
    *
    */
@@ -240,7 +249,7 @@ export interface AmasterClient {
 
   /**
    * Get the current access token
-   * 
+   *
    * @returns The access token string, or `null` if not authenticated
    *
    */
@@ -248,10 +257,10 @@ export interface AmasterClient {
 
   /**
    * Manually set the access token
-   * 
+   *
    * Useful when you have a token from another source (e.g., SSO, OAuth).
    * After setting the token, all requests will use it automatically.
-   * 
+   *
    * @param token - The JWT access token to set
    *
    */
@@ -259,12 +268,12 @@ export interface AmasterClient {
 
   /**
    * Clear all authentication state
-   * 
+   *
    * This removes:
    * - Access token
    * - User information
    * - Auto-refresh timers
-   * 
+   *
    * After calling this, `isAuthenticated()` will return `false`
    * and all requests will be unauthenticated.
    *
@@ -274,10 +283,10 @@ export interface AmasterClient {
 
 /**
  * Factory function to create a unified Amaster client
- * 
+ *
  * This is the main entry point for using the Amaster client library.
  * It returns a client instance that provides access to all Amaster services.
- * 
+ *
  * @param options - Configuration options for the client
  * @returns A configured Amaster client instance
  *
@@ -285,18 +294,18 @@ export interface AmasterClient {
 export declare function createClient(options: AmasterClientOptions): AmasterClient;
 
 // Re-export shared common types
-export type { ClientResult } from './common';
+export type { ClientResult } from "./common";
 
 // Re-export main client interfaces
-export type { AuthClientAPI } from './auth';
-export type { EntityClientAPI } from './entity';
-export type { BpmClientAPI } from './bpm';
-export type { WorkflowClientAPI } from './workflow';
-export type { ASRClientAPI } from './asr';
-export type { CopilotClientAPI } from './copilot';
-export type { FunctionClientAPI } from './function';
-export type { TTSClientAPI } from './tts';
-export type { S3ClientAPI } from './s3';
+export type { AuthClientAPI } from "./auth";
+export type { EntityClientAPI } from "./entity";
+export type { BpmClientAPI } from "./bpm";
+export type { WorkflowClientAPI } from "./workflow";
+export type { ASRClient, ASRClientConfig, ASRHttpClient, ASRHttpClientConfig } from "./asr";
+export type { CopilotClientAPI } from "./copilot";
+export type { FunctionClientAPI } from "./function";
+export type { TTSClientAPI } from "./tts";
+export type { S3ClientAPI } from "./s3";
 
 // For detailed types, import directly from submodules:
 // import type { LoginParams, User } from '@amaster.ai/client/auth'