use node-edge-tts

index.js

@@ -10,7 +10,7 @@ const lancedb = require('vectordb')
 const { STORE_PATH, LOG_PATH, AUDIO_PATH } = require('./utils/initFile.js')
 const { getStore, setStore } = require('./modules/store.js')
 const { getSpeechText } = require('./modules/whisper.js')
-const { EdgeTTS } = require('./modules/edge-tts.js')
+const { EdgeTTS } = require('node-edge-tts')
 const { openaiChatStream, openaiEmbedding, azureOpenaiChatStream, azureOpenaiEmbedding } = require('./modules/common.js')
 const { functionAction, functionInfo, functionList } = require('./modules/functions.js')
 const { config: {
@@ -60,7 +60,8 @@ let tts = new EdgeTTS({
   voice: SpeechSynthesisVoiceName,
   lang: 'zh-CN',
   outputFormat: 'audio-24khz-96kbitrate-mono-mp3',
-  proxy: proxyString
+  proxy: proxyString,
+  saveSubtitles: true
 })
 
 let mainWindow

modules/edge-tts.js

@@ -1,23 +1,28 @@
 const { randomBytes } = require('node:crypto')
-const fs = require('node:fs')
+const { writeFileSync, createWriteStream } = require('node:fs')
 const { WebSocket } = require('ws')
 const { HttpsProxyAgent } = require('https-proxy-agent')
 
 
 class EdgeTTS {
+
   voice
   lang
   outputFormat
   proxy
+  saveSubtitles
+
   constructor ({
     voice = 'zh-CN-XiaoyiNeural',
     lang = 'zh-CN',
     outputFormat = 'audio-24khz-48kbitrate-mono-mp3',
+    saveSubtitles = false,
     proxy
   }) {
     this.voice = voice
     this.lang = lang
     this.outputFormat = outputFormat
+    this.saveSubtitles = saveSubtitles
     this.proxy = proxy
   }
 
@@ -37,11 +42,11 @@ class EdgeTTS {
             "context": {
               "synthesis": {
                 "audio": {
-                    "metadataoptions": {
-                      "sentenceBoundaryEnabled": "false",
-                      "wordBoundaryEnabled": "true"
-                    },
-                    "outputFormat": "${this.outputFormat}"
+                  "metadataoptions": {
+                    "sentenceBoundaryEnabled": "false",
+                    "wordBoundaryEnabled": "true"
+                  },
+                  "outputFormat": "${this.outputFormat}"
                 }
               }
             }
@@ -73,13 +78,13 @@ class EdgeTTS {
       }
       cue.part = fullPart
     })
-    fs.writeFileSync(subPath, JSON.stringify(subFile, null, '  '), { encoding: 'utf-8' })
+    writeFileSync(subPath, JSON.stringify(subFile, null, '  '), { encoding: 'utf-8' })
   }
 
   async ttsPromise (text, audioPath) {
     const _wsConnect = await this._connectWebSocket()
     return await new Promise((resolve, reject) => {
-      let audioStream = fs.createWriteStream(audioPath)
+      let audioStream = createWriteStream(audioPath)
       let subFile = []
       _wsConnect.on('message', async (data, isBinary) => {
         if (isBinary) {
@@ -91,7 +96,9 @@ class EdgeTTS {
           let message = data.toString()
           if (message.includes('Path:turn.end')) {
             audioStream.end()
-            this._saveSubFile(subFile, text, audioPath)
+            if (this.saveSubtitles) {
+              this._saveSubFile(subFile, text, audioPath)
+            }
             resolve()
           } else if (message.includes('Path:audio.metadata')) {
             let splitTexts = message.split('\r\n')
@@ -108,7 +115,7 @@ class EdgeTTS {
       _wsConnect.send(`X-RequestId:${requestId}\r\nContent-Type:application/ssml+xml\r\nPath:ssml\r\n\r\n
       ` + `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${this.lang}">
         <voice name="${this.voice}">
-            ${text}
+          ${text}
         </voice>
       </speak>`)
     })

package.json

@@ -57,6 +57,7 @@
     "js-tiktoken": "^1.0.7",
     "lodash": "^4.17.21",
     "nanoid": "^3.3.6",
+    "node-edge-tts": "^1.1.3",
     "openai": "^4.3.0",
     "quickjs-emscripten": "^0.23.0",
     "sound-play": "^1.1.0",