@ww_nero/audio 1.0.1

package/index.js

@@ -0,0 +1,388 @@
+#!/usr/bin/env node
+
+const https = require('https');
+const fs = require('fs');
+const path = require('path');
+const { Server } = require('@modelcontextprotocol/sdk/server/index.js');
+const { StdioServerTransport } = require('@modelcontextprotocol/sdk/server/stdio.js');
+const {
+  CallToolRequestSchema,
+  ListToolsRequestSchema,
+} = require('@modelcontextprotocol/sdk/types.js');
+
+const BASE_URL = (process.env.AUDIO_BASE_URL || '').replace(/\/+$/, '');
+const API_KEY = process.env.AUDIO_API_KEY || '';
+
+const TTS_MODEL = 'gpt-4o-mini-tts';
+const ASR_MODEL = 'whisper-1';
+
+const VOICE_MAP = {
+  female: 'marin',
+  male: 'cedar',
+};
+
+const MAX_FILE_SIZE = 20 * 1024 * 1024; // 20MB
+
+const convertPath = (filePath) => {
+  const wslMatch = filePath.match(/^\/mnt\/([a-zA-Z])\/(.*)$/);
+  if (wslMatch) {
+    const drive = wslMatch[1].toUpperCase();
+    const rest = wslMatch[2].replace(/\//g, '\\');
+    return `${drive}:\\${rest}`;
+  }
+
+  const winMatch = filePath.match(/^([a-zA-Z]):\\(.*)$/);
+  if (winMatch) {
+    const drive = winMatch[1].toLowerCase();
+    const rest = winMatch[2].replace(/\\/g, '/');
+    return `/mnt/${drive}/${rest}`;
+  }
+
+  return null;
+};
+
+const resolveExistingPath = (inputPath) => {
+  if (fs.existsSync(inputPath)) {
+    return inputPath;
+  }
+
+  const converted = convertPath(inputPath);
+  if (converted && fs.existsSync(converted)) {
+    return converted;
+  }
+
+  return null;
+};
+
+const resolveWorkingDirectory = (rawPath) => {
+  if (!rawPath || typeof rawPath !== 'string') {
+    throw new Error('必须提供 working_directory');
+  }
+
+  if (!path.isAbsolute(rawPath)) {
+    throw new Error('working_directory 必须是绝对路径');
+  }
+
+  const resolved = resolveExistingPath(rawPath);
+  if (!resolved) {
+    throw new Error('工作目录不存在，请确认路径是否正确');
+  }
+
+  const stats = fs.statSync(resolved);
+  if (!stats.isDirectory()) {
+    throw new Error('working_directory 必须是文件夹路径');
+  }
+
+  return resolved;
+};
+
+const resolveAudioFile = (workingDir, rawPath) => {
+  if (!rawPath || typeof rawPath !== 'string') {
+    throw new Error('音频文件路径不能为空');
+  }
+
+  const trimmed = rawPath.trim();
+  if (!trimmed) {
+    throw new Error('音频文件路径不能为空');
+  }
+
+  const absolutePath = path.isAbsolute(trimmed)
+    ? trimmed
+    : path.join(workingDir, trimmed);
+
+  const resolved = resolveExistingPath(absolutePath);
+  if (!resolved) {
+    throw new Error(`音频文件不存在: ${trimmed}`);
+  }
+
+  const stats = fs.statSync(resolved);
+  if (!stats.isFile()) {
+    throw new Error(`音频路径不是文件: ${trimmed}`);
+  }
+
+  if (stats.size > MAX_FILE_SIZE) {
+    throw new Error(`音频文件大小超过限制 (最大 20MB)，当前大小: ${(stats.size / 1024 / 1024).toFixed(2)}MB`);
+  }
+
+  return resolved;
+};
+
+const postTTS = (text, voice, speed, instructions) =>
+  new Promise((resolve, reject) => {
+    if (!BASE_URL) {
+      reject(new Error('未配置 AUDIO_BASE_URL 环境变量'));
+      return;
+    }
+    if (!API_KEY) {
+      reject(new Error('未配置 AUDIO_API_KEY 环境变量'));
+      return;
+    }
+
+    const urlObj = new URL(`${BASE_URL}/v1/audio/speech`);
+    const body = {
+      model: TTS_MODEL,
+      voice,
+      input: text,
+      speed,
+    };
+
+    if (instructions) {
+      body.instructions = instructions;
+    }
+
+    const data = JSON.stringify(body);
+
+    const options = {
+      hostname: urlObj.hostname,
+      port: urlObj.port || 443,
+      path: urlObj.pathname,
+      method: 'POST',
+      headers: {
+        'Authorization': `Bearer ${API_KEY}`,
+        'Content-Type': 'application/json',
+        'Content-Length': Buffer.byteLength(data),
+      },
+    };
+
+    const req = https.request(options, (res) => {
+      const chunks = [];
+
+      res.on('data', (chunk) => {
+        chunks.push(chunk);
+      });
+
+      res.on('end', () => {
+        if (res.statusCode !== 200) {
+          const respData = Buffer.concat(chunks).toString();
+          reject(new Error(`请求失败，状态码: ${res.statusCode}，响应: ${respData}`));
+          return;
+        }
+
+        resolve(Buffer.concat(chunks));
+      });
+    });
+
+    req.on('error', (err) => reject(new Error(`请求错误: ${err.message}`)));
+
+    req.write(data);
+    req.end();
+  });
+
+const postASR = (filePath) =>
+  new Promise((resolve, reject) => {
+    if (!BASE_URL) {
+      reject(new Error('未配置 AUDIO_BASE_URL 环境变量'));
+      return;
+    }
+    if (!API_KEY) {
+      reject(new Error('未配置 AUDIO_API_KEY 环境变量'));
+      return;
+    }
+
+    const urlObj = new URL(`${BASE_URL}/v1/audio/transcriptions`);
+    const boundary = `----FormBoundary${Date.now()}`;
+    const fileName = path.basename(filePath);
+    const fileContent = fs.readFileSync(filePath);
+
+    const formParts = [];
+
+    // model field
+    formParts.push(
+      `--${boundary}\r\n`,
+      `Content-Disposition: form-data; name="model"\r\n\r\n`,
+      `${ASR_MODEL}\r\n`
+    );
+
+    // response_format field
+    formParts.push(
+      `--${boundary}\r\n`,
+      `Content-Disposition: form-data; name="response_format"\r\n\r\n`,
+      `srt\r\n`
+    );
+
+    // file field
+    formParts.push(
+      `--${boundary}\r\n`,
+      `Content-Disposition: form-data; name="file"; filename="${fileName}"\r\n`,
+      `Content-Type: application/octet-stream\r\n\r\n`
+    );
+
+    const formHeader = Buffer.from(formParts.join(''));
+    const formFooter = Buffer.from(`\r\n--${boundary}--\r\n`);
+    const body = Buffer.concat([formHeader, fileContent, formFooter]);
+
+    const options = {
+      hostname: urlObj.hostname,
+      port: urlObj.port || 443,
+      path: urlObj.pathname,
+      method: 'POST',
+      headers: {
+        'Authorization': `Bearer ${API_KEY}`,
+        'Content-Type': `multipart/form-data; boundary=${boundary}`,
+        'Content-Length': body.length,
+      },
+    };
+
+    const req = https.request(options, (res) => {
+      let respData = '';
+
+      res.on('data', (chunk) => {
+        respData += chunk;
+      });
+
+      res.on('end', () => {
+        if (res.statusCode !== 200) {
+          reject(new Error(`请求失败，状态码: ${res.statusCode}，响应: ${respData}`));
+          return;
+        }
+
+        resolve(respData);
+      });
+    });
+
+    req.on('error', (err) => reject(new Error(`请求错误: ${err.message}`)));
+
+    req.write(body);
+    req.end();
+  });
+
+const tts = async ({ text, gender, speed, instructions, working_directory }) => {
+  const workingDir = resolveWorkingDirectory(working_directory);
+
+  const voice = VOICE_MAP[gender] || VOICE_MAP.female;
+  const audioBuffer = await postTTS(text, voice, speed, instructions);
+
+  const filename = `audio-${Date.now()}.mp3`;
+  const outputPath = path.join(workingDir, filename);
+  fs.writeFileSync(outputPath, audioBuffer);
+
+  return filename;
+};
+
+const asr = async ({ audio_file, working_directory }) => {
+  const workingDir = resolveWorkingDirectory(working_directory);
+  const audioPath = resolveAudioFile(workingDir, audio_file);
+
+  const srtContent = await postASR(audioPath);
+
+  const baseName = path.basename(audio_file, path.extname(audio_file));
+  const filename = `${baseName}.srt`;
+  const outputPath = path.join(workingDir, filename);
+  fs.writeFileSync(outputPath, srtContent);
+
+  return filename;
+};
+
+const server = new Server(
+  {
+    name: 'audio',
+    version: '1.0.1',
+  },
+  {
+    capabilities: {
+      tools: {},
+    },
+  }
+);
+
+server.setRequestHandler(ListToolsRequestSchema, async () => ({
+  tools: [
+    {
+      name: 'tts',
+      description: '语音合成工具，将文本转换为音频文件。',
+      inputSchema: {
+        type: 'object',
+        properties: {
+          working_directory: {
+            type: 'string',
+            description: '工作目录的绝对路径，合成的音频文件将保存到此目录',
+          },
+          text: {
+            type: 'string',
+            description: '需要合成语音的文本内容',
+          },
+          gender: {
+            type: 'string',
+            description: '语音性别，female 为女声，male 为男声，默认为 female',
+            enum: ['female', 'male'],
+          },
+          speed: {
+            type: 'number',
+            description: '语速，可调区间为 0.25-4，数字越大语速越快，默认为 1',
+          },
+          instructions: {
+            type: 'string',
+            description: '对语音语调的要求说明，如 "Speak in a cheerful and positive tone."',
+          },
+        },
+        required: ['working_directory', 'text'],
+      },
+    },
+    {
+      name: 'asr',
+      description: '语音识别工具，将音频文件转换为带时间戳的 SRT 字幕文件。',
+      inputSchema: {
+        type: 'object',
+        properties: {
+          working_directory: {
+            type: 'string',
+            description: '工作目录的绝对路径，识别结果将保存到此目录',
+          },
+          audio_file: {
+            type: 'string',
+            description: '音频文件的相对路径（相对于工作目录），支持 mp3/wav 格式，文件大小不超过 20MB',
+          },
+        },
+        required: ['working_directory', 'audio_file'],
+      },
+    },
+  ],
+}));
+
+server.setRequestHandler(CallToolRequestSchema, async (request) => {
+  const { name, arguments: args } = request.params;
+
+  try {
+    if (name === 'tts') {
+      const { working_directory, text, gender, speed, instructions } = args;
+      if (!working_directory || !text) {
+        throw new Error('必须同时提供 working_directory 和 text 参数');
+      }
+
+      const validSpeed = typeof speed === 'number' ? Math.max(0.25, Math.min(4, speed)) : 1;
+
+      const filename = await tts({
+        text,
+        gender: gender || 'female',
+        speed: validSpeed,
+        instructions,
+        working_directory,
+      });
+      return { content: [{ type: 'text', text: `已保存到工作目录下，文件名为：${filename}` }] };
+    }
+
+    if (name === 'asr') {
+      const { working_directory, audio_file } = args;
+      if (!working_directory || !audio_file) {
+        throw new Error('必须同时提供 working_directory 和 audio_file 参数');
+      }
+      const filename = await asr({ audio_file, working_directory });
+      return { content: [{ type: 'text', text: `已保存到工作目录下，文件名为：${filename}` }] };
+    }
+
+    return {
+      content: [{ type: 'text', text: `未知工具: ${name}` }],
+      isError: true,
+    };
+  } catch (error) {
+    const message = error?.message || '未知错误';
+    return { content: [{ type: 'text', text: message }], isError: true };
+  }
+});
+
+const main = async () => {
+  const transport = new StdioServerTransport();
+  await server.connect(transport);
+};
+
+main().catch(console.error);

package/package.json

@@ -0,0 +1,15 @@
+{
+  "name": "@ww_nero/audio",
+  "version": "1.0.1",
+  "description": "MCP server for TTS and ASR using OpenAI compatible API",
+  "main": "index.js",
+  "bin": {
+    "audio": "index.js"
+  },
+  "files": [
+    "index.js"
+  ],
+  "dependencies": {
+    "@modelcontextprotocol/sdk": "^1.22.0"
+  }
+}