@ww_nero/media 1.0.10 → 1.2.0

package/README.md

@@ -27,12 +27,9 @@
 **参数：**
 - `working_directory`: 工作目录的绝对路径，合成的音频文件将保存到此目录
 - `text`: 需要合成语音的文本内容
-- `voice`: （可选）音色模型，默认 `sambert-zhimiao-emo-v1`
-- `format`: （可选）输出音频格式，可选 mp3（默认）、wav、pcm
-- `sample_rate`: （可选）采样率，默认 16000
 
 **输出：**
-- 合成结果保存到工作目录下的 `tts_<timestamp>.<format>` 文件
+- 合成结果保存到工作目录下的 `tts_<timestamp>.mp3` 文件
 
 ## 环境变量
 

package/index.js

@@ -114,7 +114,7 @@ const resolveAudioFile = (workingDir, rawPath) => {
 const server = new Server(
   {
     name: 'media',
-    version: '1.0.10',
+    version: '1.2.0',
   },
   {
     capabilities: {
@@ -157,17 +157,10 @@ server.setRequestHandler(ListToolsRequestSchema, async () => ({
             type: 'string',
             description: '需要合成语音的文本内容',
           },
-          voice: {
+          gender: {
             type: 'string',
-            description: '音色模型，可选值如 sambert-zhimiao-emo-v1（默认）、sambert-zhichu-v1 等',
-          },
-          format: {
-            type: 'string',
-            description: '输出音频格式，可选 mp3（默认）、wav、pcm',
-          },
-          sample_rate: {
-            type: 'number',
-            description: '采样率，默认 16000',
+            enum: ['female', 'male'],
+            description: '语音性别，female 为女声，male 为男声，默认为 female',
           },
         },
         required: ['working_directory', 'text'],
@@ -192,7 +185,7 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
     }
 
     if (name === 'tts') {
-      const { working_directory, text, voice, format, sample_rate } = args;
+      const { working_directory, text, gender } = args;
       if (!working_directory || !text) {
         throw new Error('必须同时提供 working_directory 和 text 参数');
       }
@@ -200,10 +193,8 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
       const filename = await tts({
         workingDir,
         text,
+        gender,
         apiKey: DASHSCOPE_API_KEY,
-        voice: voice || 'sambert-zhimiao-emo-v1',
-        format: format || 'mp3',
-        sampleRate: sample_rate || 16000,
       });
       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": "1.0.10",
+  "version": "1.2.0",
   "description": "MCP server for media processing, including ASR speech recognition and TTS speech synthesis",
   "main": "index.js",
   "bin": {

package/utils/tts.js

@@ -5,10 +5,21 @@ 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, voice = 'sambert-zhimiao-emo-v1', format = 'mp3', sampleRate = 16000 }) => {
+const tts = async ({ workingDir, text, gender = 'female', 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 API Key
   if (!apiKey) {
     throw new Error('请配置 DASHSCOPE_API_KEY 环境变量');