@ww_nero/media 1.1.0 → 1.3.0

package/index.js

@@ -114,7 +114,7 @@ const resolveAudioFile = (workingDir, rawPath) => {
 const server = new Server(
   {
     name: 'media',
-    version: '1.1.0',
+    version: '1.3.0',
   },
   {
     capabilities: {
@@ -157,6 +157,15 @@ server.setRequestHandler(ListToolsRequestSchema, async () => ({
             type: 'string',
             description: '需要合成语音的文本内容',
           },
+          gender: {
+            type: 'string',
+            enum: ['female', 'male'],
+            description: '语音性别，female 为女声，male 为男声，默认为 female',
+          },
+          rate: {
+            type: 'number',
+            description: '语速，可调区间为 0.5-2，数字越大语速越快，默认为 1',
+          },
         },
         required: ['working_directory', 'text'],
       },
@@ -180,7 +189,7 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
     }
 
     if (name === 'tts') {
-      const { working_directory, text } = args;
+      const { working_directory, text, gender, rate } = args;
       if (!working_directory || !text) {
         throw new Error('必须同时提供 working_directory 和 text 参数');
       }
@@ -188,6 +197,8 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
       const filename = await tts({
         workingDir,
         text,
+        gender,
+        rate,
         apiKey: DASHSCOPE_API_KEY,
       });
       return { content: [{ type: 'text', text: `语音合成完成，音频文件已保存到工作目录下：${filename}` }] };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ww_nero/media",
-  "version": "1.1.0",
+  "version": "1.3.0",
   "description": "MCP server for media processing, including ASR speech recognition and TTS speech synthesis",
   "main": "index.js",
   "bin": {

package/utils/asr.js

@@ -1,7 +1,8 @@
 const fs = require('fs');
 const path = require('path');
 
-const ASR_UPLOAD_URL = 'http://fsheep.com:10808/upload';
+const FILE_BASE_URL = 'http://fsheep.com:10801';
+const ASR_UPLOAD_URL = `${FILE_BASE_URL}/file/upload`;
 const ASR_SERVICE_URL = 'https://dashscope.aliyuncs.com/api/v1/services/audio/asr/transcription';
 
 const SUPPORTED_AUDIO_TYPES = ['.mp3', '.wav'];
@@ -73,7 +74,7 @@ const uploadAudio = async (uploadUrl, audioPath) => {
 
   const data = await response.json();
 
-  if (!data.success || !data.data || !data.data.path) {
+  if (data.code !== 0 || !data.data || !data.data.path) {
     throw new Error(`上传响应格式错误: ${JSON.stringify(data)}`);
   }
 
@@ -81,11 +82,10 @@ const uploadAudio = async (uploadUrl, audioPath) => {
 };
 
 /**
- * 根据上传接口 URL 和文件路径构建静态资源 URL
+ * 根据 base URL 和文件路径构建静态资源 URL
  */
-const getStaticUrl = (uploadUrl, filePath) => {
-  const url = new URL(uploadUrl);
-  return `${url.protocol}//${url.host}${filePath}`;
+const getStaticUrl = (baseUrl, filePath) => {
+  return `${baseUrl}${filePath}`;
 };
 
 /**
@@ -202,7 +202,7 @@ const asr = async ({ workingDir, audioPath, apiKey }) => {
   const filePath = await uploadAudio(ASR_UPLOAD_URL, audioPath);
 
   // 2. 构建静态资源 URL
-  const audioUrl = getStaticUrl(ASR_UPLOAD_URL, filePath);
+  const audioUrl = getStaticUrl(FILE_BASE_URL, filePath);
 
   // 3. 提交 ASR 任务
   const taskId = await submitAsrTask([audioUrl], apiKey);

package/utils/tts.js

@@ -5,15 +5,24 @@ const { v4: uuidv4 } = require('uuid');
 
 const TTS_WS_URL = 'wss://dashscope.aliyuncs.com/api-ws/v1/inference/';
 
+// 音色配置
+const VOICE_MAP = {
+  female: 'sambert-zhimiao-emo-v1',
+  male: 'sambert-zhihao-v1',
+};
+
 /**
  * TTS 语音合成
  */
-const tts = async ({ workingDir, text, apiKey }) => {
-  // 固定参数
-  const voice = 'sambert-zhimiao-emo-v1';
+const tts = async ({ workingDir, text, gender = 'female', rate = 1, apiKey }) => {
+  // 根据性别选择音色
+  const voice = VOICE_MAP[gender] || VOICE_MAP.female;
   const format = 'mp3';
   const sampleRate = 16000;
 
+  // 验证语速参数
+  const validRate = Math.max(0.5, Math.min(2, Number(rate) || 1));
+
   // 验证 API Key
   if (!apiKey) {
     throw new Error('请配置 DASHSCOPE_API_KEY 环境变量');
@@ -75,7 +84,7 @@ const tts = async ({ workingDir, text, apiKey }) => {
             format: format,
             sample_rate: sampleRate,
             volume: 50,
-            rate: 1,
+            rate: validRate,
             pitch: 1,
             word_timestamp_enabled: false,
             phoneme_timestamp_enabled: false