Update main.ts

main.ts

@@ -1,21 +1,220 @@
+// import { serve } from "https://deno.land/std/http/server.ts";
+// import { EdgeSpeechTTS } from "https://esm.sh/@lobehub/tts@1";
+
+// const AUTH_TOKEN = Deno.env.get("AUTH_TOKEN");
+// const VOICES_URL = "https://speech.platform.bing.com/consumer/speech/synthesize/readaloud/voices/list?trustedclienttoken=6A5AA1D4EAFF4E9FB37E23D68491D6F4";
+
+// async function fetchVoiceList() {
+//   const response = await fetch(VOICES_URL);
+//   const voices = await response.json();
+//   return voices.reduce((acc: Record<string, { model: string, name: string, friendlyName: string, locale: string }[]>, voice: any) => {
+//     const { ShortName: model, ShortName: name, FriendlyName: friendlyName, Locale: locale } = voice;
+//     if (!acc[locale]) acc[locale] = [];
+//     acc[locale].push({ model, name, friendlyName, locale });
+//     return acc;
+//   }, {});
+// }
+
+// async function synthesizeSpeech(model: string, voice: string, text: string) {
+//   let voiceName;
+//   let rate = 0;
+//   let pitch = 0;
+
+//   if (model.includes("tts")) {
+//     rate = 0.1;
+//     pitch = 0.2;
+
+//     switch (voice) {
+//       case "alloy":
+//         voiceName = "zh-CN-YunjianNeural";
+//         break;
+//       case "echo":
+//         voiceName = "zh-CN-YunyangNeural";
+//         break;
+//       case "fable":
+//         voiceName = "zh-CN-XiaoxiaoNeural";
+//         break;
+//       case "onyx":
+//         voiceName = "zh-TW-HsiaoChenNeural";
+//         break;
+//       default:
+//         voiceName = "zh-CN-YunxiNeural";
+//         break;
+//     }
+//   } else {
+//     voiceName = model;
+//     const params = Object.fromEntries(
+//       voice.split("|").map((p) => p.split(":") as [string, string])
+//     );
+//     rate = Number(params["rate"] || 0);
+//     pitch = Number(params["pitch"] || 0);
+//   }
+
+//   const tts = new EdgeSpeechTTS();
+  
+//   const payload = {
+//     input: text,
+//     options: { 
+//       rate: rate,
+//       pitch: pitch,
+//       voice: voiceName
+//      },
+//   };
+//   const response = await tts.create(payload);
+//   const mp3Buffer = new Uint8Array(await response.arrayBuffer());
+
+//   console.log(`Successfully synthesized speech, returning audio/mpeg response`);
+//   return new Response(mp3Buffer, {
+//     headers: { "Content-Type": "audio/mpeg" },
+//   });
+// }
+
+// function unauthorized(req: Request) {
+//   const authHeader = req.headers.get("Authorization");
+//   return AUTH_TOKEN && authHeader !== `Bearer ${AUTH_TOKEN}`;
+// }
+
+// function validateContentType(req: Request, expected: string) {
+//   const contentType = req.headers.get("Content-Type");
+//   if (contentType !== expected) {
+//     console.log(`Invalid Content-Type ${contentType}, expected ${expected}`);
+//     return new Response("Bad Request", { status: 400 });
+//   }
+// }
+
+// async function handleDebugRequest(req: Request) {
+//   const url = new URL(req.url);
+//   const voice = url.searchParams.get("voice") || "";
+//   const model = url.searchParams.get("model") || "";
+//   const text = url.searchParams.get("text") || "";
+
+//   console.log(`Debug request with model=${model}, voice=${voice}, text=${text}`);
+
+//   if (!voice || !model || !text) {
+//     console.log("Missing required parameters");
+//     return new Response("Bad Request", { status: 400 });
+//   }
+
+//   return synthesizeSpeech(model, voice, text);
+// }
+
+// async function handleSynthesisRequest(req: Request) {
+//   if (unauthorized(req)) {
+//     console.log("Unauthorized request");
+//     return new Response("Unauthorized", { status: 401 });
+//   }
+
+//   if (req.method !== "POST") {
+//     console.log(`Invalid method ${req.method}, expected POST`);
+//     return new Response("Method Not Allowed", { status: 405 });
+//   }
+
+//   const invalidContentType = validateContentType(req, "application/json");
+//   if (invalidContentType) return invalidContentType;
+
+//   const { model, input, voice } = await req.json();
+//   console.log(`Synthesis request with model=${model}, input=${input}, voice=${voice}`);
+
+//   return synthesizeSpeech(model, voice, input);
+// }
+
+
+// async function handleDemoRequest(req: Request) {
+//     const groupedVoiceList = await fetchVoiceList();
+  
+//     const html = `<!DOCTYPE html><html lang="en"><head><meta charset="UTF-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><title>语音合成演示</title><link href="https://fonts.googleapis.com/css2?family=Noto+Sans+SC:wght@400;700&display=swap" rel="stylesheet"><style>:root{--primary-color:#6c8bd6;--primary-light:#a2b3e3;--primary-dark:#3d5b8f;--secondary-color:#f08080;--text-color:#333;--text-secondary:#777;--bg-color:#fff}body{font-family:'Noto Sans SC','Arial',sans-serif;color:var(--text-color);margin:0;padding:0;display:flex;justify-content:center;background-color:#fafafa;background-image:linear-gradient(135deg,#f5f7fa 0%,#c3cfe2 100%);position:relative;overflow:hidden}body::before{content:"";position:absolute;top:0;left:0;right:0;bottom:0;background:repeating-radial-gradient(circle at 50% 50%,rgba(255,255,255,0.8) 0%,rgba(255,255,255,0.8) 2%,transparent 2%,transparent 4%,rgba(255,255,255,0.8) 4%,rgba(255,255,255,0.8) 6%,transparent 6%,transparent 8%,rgba(255,255,255,0.8) 8%,rgba(255,255,255,0.8) 10%,transparent 10%),repeating-linear-gradient(45deg,#D4F4FF 0%,#D4F4FF 5%,#E6F9FF 5%,#E6F9FF 10%,#F0FAFF 10%,#F0FAFF 15%,#E6F9FF 15%,#E6F9FF 20%,#D4F4FF 20%,#D4F4FF 25%);background-blend-mode:multiply;opacity:0.8;z-index:-1;animation:glitch 15s infinite}.container{display:flex;max-width:1200px;width:100%;margin:40px;background:#fff;border-radius:12px;position:relative;background-color:rgba(255,255,255,0.8);z-index:1}@keyframes glitch{0%{background-position:0 0,0 0;filter:hue-rotate(0deg)}50%{background-position:10px 10px,-10px 10px;filter:hue-rotate(360deg)}100%{background-position:0 0,0 0;filter:hue-rotate(0deg)}}.input-area,.output-area{padding:30px;width:50%}.input-area{border-right:1px solid #E0E0E0}h1{font-size:36px;color:var(--primary-color);margin-bottom:30px}.filter-section{margin-bottom:30px}.filter-section label{display:block;font-size:16px;color:var(--text-secondary);margin-bottom:10px}.filter-section input{font-size:16px;padding:10px 15px;border:2px solid var(--primary-light);border-radius:8px;outline:none;transition:border-color .3s,box-shadow .3s;width:100%;box-sizing:border-box}.filter-section input:focus{border-color:var(--primary-color);box-shadow:0 0 0 2px var(--primary-light)}.slider-container{margin-bottom:30px}.slider-container label{display:block;font-size:16px;color:var(--text-secondary);margin-bottom:10px}.slider{-webkit-appearance:none;width:100%;height:10px;border-radius:5px;background:linear-gradient(to right,var(--secondary-color) 0%,var(--primary-color) 50%,var(--primary-light) 100%);box-shadow:inset 0 1px 2px rgba(0,0,0,0.1),0 1px rgba(255,255,255,0.1);outline:none;opacity:0.7;-webkit-transition:.2s;transition:opacity .2s;margin-bottom:10px}.slider:hover{opacity:1}.slider::-webkit-slider-thumb{-webkit-appearance:none;appearance:none;width:20px;height:20px;border-radius:50%;background:#fff;border:2px solid var(--primary-color);cursor:pointer}.slider::-moz-range-thumb{width:20px;height:20px;border-radius:50%;background:#fff;border:2px solid var(--primary-color);cursor:pointer}.slider-value{font-size:14px;color:var(--text-secondary)}.textarea-container{margin-bottom:30px}.textarea-container label{display:block;font-size:18px;margin-bottom:10px}.textarea-container textarea{width:100%;padding:10px;font-size:16px;border:2px solid var(--primary-light);border-radius:8px;outline:none;resize:vertical;transition:border-color .3s,box-shadow .3s;box-sizing:border-box;height:200px}.textarea-container textarea:focus{border-color:var(--primary-color);box-shadow:0 0 0 2px var(--primary-light)}.voice-group{margin-bottom:20px;border:2px solid var(--primary-light);border-radius:12px;overflow:hidden;cursor:move;background:#fff}.voice-header{padding:15px 20px;font-size:18px;background:var(--primary-light);color:#fff;cursor:pointer;display:flex;justify-content:space-between;align-items:center}.voice-header:hover{background:var(--primary-color)}.voice-buttons{padding:20px;display:none;gap:12px;flex-wrap:wrap}.voice-button{background:var(--secondary-color);color:#fff;border:none;padding:10px 20px;border-radius:50px;cursor:pointer;transition:filter .3s}.voice-button:hover{filter:brightness(0.9)}.chevron{transition:transform .3s}.voice-group.open .voice-buttons{display:flex}.voice-group.open .chevron{transform:rotate(180deg)}.dragging{opacity:0.5}</style></head><body><div class="container"><div class="input-area"><h1>输入文本</h1><div class="filter-section"><label for="keywords">Speaker筛选:</label><input type="text" id="keywords" value="multilingual,-TW,-CN"></div><div class="slider-container"><label for="rate">语速:</label><input type="range" min="-1" max="1" step="0.1" value="-0.1" class="slider" id="rate"><div class="slider-value" id="rateValue">-0.1</div><label for="pitch">音调:</label><input type="range" min="-1" max="1" step="0.1" value="0.1" class="slider" id="pitch"><div class="slider-value" id="pitchValue">0.1</div></div><div class="textarea-container"><label for="inputText">输入文本:</label><textarea id="inputText">Hello world</textarea></div></div><div class="output-area"><h1>选择语音</h1><div id="voices"></div></div></div><script>const voiceList = ${JSON.stringify(groupedVoiceList)};let audio=null;function filterVoices(){const keywords=document.getElementById('keywords').value.split(',').map(k=>k.trim().toLowerCase());const voicesDiv=document.getElementById('voices');voicesDiv.innerHTML='';const filteredVoices={};for(const[locale,voices]of Object.entries(voiceList)){const filtered=voices.filter(({name,friendlyName})=>keywords.some(keyword=>name.toLowerCase().includes(keyword)||friendlyName.toLowerCase().includes(keyword)));if(filtered.length>0){filteredVoices[locale]=filtered}}for(const[locale,voices]of Object.entries(filteredVoices)){const group=document.createElement('div');group.className='voice-group';group.draggable=true;const header=document.createElement('div');header.className='voice-header';header.textContent=locale.toUpperCase();const chevron=document.createElement('span');chevron.className='chevron';chevron.innerHTML='&#9660;';header.appendChild(chevron);const buttonsContainer=document.createElement('div');buttonsContainer.className='voice-buttons';voices.forEach(({model,name})=>{const button=document.createElement('button');button.className='voice-button';button.textContent=name;button.onclick=()=>synthesize(model);buttonsContainer.appendChild(button)});header.onclick=()=>{group.classList.toggle('open')};group.appendChild(header);group.appendChild(buttonsContainer);voicesDiv.appendChild(group)}addDragDropListeners()}function synthesize(model){const text=document.getElementById('inputText').value||'Hello world';const rate=document.getElementById('rate').value||'-0.1';const pitch=document.getElementById('pitch').value||'0.1';const voice=\`rate:\${rate}|pitch:\${pitch}\`;if(audio){audio.pause();audio.currentTime=0}fetch('/v1/audio/speech',{method:'POST',headers:{'Content-Type':'application/json'},body:JSON.stringify({model,input:text,voice})}).then(response=>response.blob()).then(blob=>{const audioUrl=URL.createObjectURL(blob);audio=new Audio(audioUrl);audio.play()})}function addDragDropListeners(){const voicesDiv=document.getElementById('voices');let draggedItem=null;voicesDiv.addEventListener('dragstart',e=>{draggedItem=e.target;e.target.classList.add('dragging')});voicesDiv.addEventListener('dragend',e=>{e.target.classList.remove('dragging');draggedItem=null});voicesDiv.addEventListener('dragover',e=>{e.preventDefault();const afterElement=getDragAfterElement(voicesDiv,e.clientY);if(afterElement==null){voicesDiv.appendChild(draggedItem)}else{voicesDiv.insertBefore(draggedItem,afterElement)}})}function getDragAfterElement(container,y){const draggableElements=[...container.querySelectorAll('.voice-group:not(.dragging)')];return draggableElements.reduce((closest,child)=>{const box=child.getBoundingClientRect();const offset=y-box.top-box.height/2;if(offset<0&&offset>closest.offset){return{offset:offset,element:child}}else{return closest}},{offset:Number.NEGATIVE_INFINITY}).element}filterVoices();document.getElementById('keywords').addEventListener('input',filterVoices);const rateSlider=document.getElementById('rate');const rateValue=document.getElementById('rateValue');rateSlider.oninput=function(){rateValue.innerHTML=this.value};const pitchSlider=document.getElementById('pitch');const pitchValue=document.getElementById('pitchValue');pitchSlider.oninput=function(){pitchValue.innerHTML=this.value}</script></body></html>`;
+  
+//     return new Response(html, {
+//       headers: { "Content-Type": "text/html" },
+//     });
+//   }
+  
+
+// serve(async (req) => {
+//   try {
+//     const url = new URL(req.url);
+
+//     if (url.pathname === "/") {
+//       return handleDemoRequest(req);
+//     }
+
+//     if (url.pathname === "/tts") {
+//       return handleDebugRequest(req);
+//     }
+
+//     if (url.pathname !== "/v1/audio/speech") {
+//       console.log(`Unhandled path ${url.pathname}`);
+//       return new Response("Not Found", { status: 404 });
+//     }
+
+//     return handleSynthesisRequest(req);
+//   } catch (err) {
+//     console.error(`Error processing request: ${err.message}`);
+//     return new Response(`Internal Server Error\n${err.message}`, {
+//       status: 500,
+//     });
+//   }
+// });
 import { serve } from "https://deno.land/std/http/server.ts";
-import { EdgeSpeechTTS } from "https://esm.sh/@lobehub/tts@1";
 
 const AUTH_TOKEN = Deno.env.get("AUTH_TOKEN");
 const VOICES_URL = "https://speech.platform.bing.com/consumer/speech/synthesize/readaloud/voices/list?trustedclienttoken=6A5AA1D4EAFF4E9FB37E23D68491D6F4";
 
+// 简化的Edge TTS实现
+class SimpleEdgeTTS {
+  async create(options) {
+    const { input, options: voiceOptions } = options;
+    const { voice, rate = 0, pitch = 0 } = voiceOptions;
+    
+    const ssml = this.generateSSML(input, voice, rate, pitch);
+    
+    const response = await fetch("https://speech.platform.bing.com/consumer/speech/synthesize/readaloud/edge/v1", {
+      method: "POST",
+      headers: {
+        "Content-Type": "application/ssml+xml",
+        "X-Microsoft-OutputFormat": "audio-24khz-48kbitrate-mono-mp3",
+        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
+      },
+      body: ssml
+    });
+    
+    return response;
+  }
+  
+  generateSSML(text, voice, rate, pitch) {
+    const rateValue = rate >= 0 ? `+${Math.round(rate * 100)}%` : `${Math.round(rate * 100)}%`;
+    const pitchValue = pitch >= 0 ? `+${Math.round(pitch * 50)}Hz` : `${Math.round(pitch * 50)}Hz`;
+    
+    return `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="zh-CN">
+      <voice name="${voice}">
+        <prosody rate="${rateValue}" pitch="${pitchValue}">
+          ${text.replace(/[<>&'"]/g, (char) => {
+            const entities = { '<': '&lt;', '>': '&gt;', '&': '&amp;', '"': '&quot;', "'": '&apos;' };
+            return entities[char];
+          })}
+        </prosody>
+      </voice>
+    </speak>`;
+  }
+}
+
 async function fetchVoiceList() {
-  const response = await fetch(VOICES_URL);
-  const voices = await response.json();
-  return voices.reduce((acc: Record<string, { model: string, name: string, friendlyName: string, locale: string }[]>, voice: any) => {
-    const { ShortName: model, ShortName: name, FriendlyName: friendlyName, Locale: locale } = voice;
-    if (!acc[locale]) acc[locale] = [];
-    acc[locale].push({ model, name, friendlyName, locale });
-    return acc;
-  }, {});
+  try {
+    const response = await fetch(VOICES_URL);
+    const voices = await response.json();
+    return voices.reduce((acc, voice) => {
+      const { ShortName: model, ShortName: name, FriendlyName: friendlyName, Locale: locale } = voice;
+      if (!acc[locale]) acc[locale] = [];
+      acc[locale].push({ model, name, friendlyName, locale });
+      return acc;
+    }, {});
+  } catch (error) {
+    console.error("Failed to fetch voice list:", error);
+    return {};
+  }
 }
 
-async function synthesizeSpeech(model: string, voice: string, text: string) {
+async function synthesizeSpeech(model, voice, text) {
   let voiceName;
   let rate = 0;
   let pitch = 0;
@@ -23,119 +222,962 @@ async function synthesizeSpeech(model: string, voice: string, text: string) {
   if (model.includes("tts")) {
     rate = 0.1;
     pitch = 0.2;
-
-    switch (voice) {
-      case "alloy":
-        voiceName = "zh-CN-YunjianNeural";
-        break;
-      case "echo":
-        voiceName = "zh-CN-YunyangNeural";
-        break;
-      case "fable":
-        voiceName = "zh-CN-XiaoxiaoNeural";
-        break;
-      case "onyx":
-        voiceName = "zh-TW-HsiaoChenNeural";
-        break;
-      default:
-        voiceName = "zh-CN-YunxiNeural";
-        break;
-    }
+    
+    const voiceMap = {
+      "alloy": "zh-CN-YunjianNeural",
+      "echo": "zh-CN-YunyangNeural", 
+      "fable": "zh-CN-XiaoxiaoNeural",
+      "onyx": "zh-TW-HsiaoChenNeural"
+    };
+    voiceName = voiceMap[voice] || "zh-CN-YunxiNeural";
   } else {
     voiceName = model;
     const params = Object.fromEntries(
-      voice.split("|").map((p) => p.split(":") as [string, string])
+      voice.split("|").map((p) => p.split(":"))
     );
     rate = Number(params["rate"] || 0);
     pitch = Number(params["pitch"] || 0);
   }
 
-  const tts = new EdgeSpeechTTS();
+  const tts = new SimpleEdgeTTS();
   
-  const payload = {
-    input: text,
-    options: { 
-      rate: rate,
-      pitch: pitch,
-      voice: voiceName
-     },
-  };
-  const response = await tts.create(payload);
-  const mp3Buffer = new Uint8Array(await response.arrayBuffer());
-
-  console.log(`Successfully synthesized speech, returning audio/mpeg response`);
-  return new Response(mp3Buffer, {
-    headers: { "Content-Type": "audio/mpeg" },
-  });
+  try {
+    const response = await tts.create({
+      input: text,
+      options: { 
+        rate: rate,
+        pitch: pitch,
+        voice: voiceName
+      }
+    });
+    
+    const mp3Buffer = new Uint8Array(await response.arrayBuffer());
+    console.log(`Successfully synthesized speech, returning audio/mpeg response`);
+    return new Response(mp3Buffer, {
+      headers: { "Content-Type": "audio/mpeg" }
+    });
+  } catch (error) {
+    console.error("Synthesis error:", error);
+    return new Response("Synthesis failed", { status: 500 });
+  }
 }
 
-function unauthorized(req: Request) {
+function unauthorized(req) {
   const authHeader = req.headers.get("Authorization");
   return AUTH_TOKEN && authHeader !== `Bearer ${AUTH_TOKEN}`;
 }
 
-function validateContentType(req: Request, expected: string) {
-  const contentType = req.headers.get("Content-Type");
-  if (contentType !== expected) {
-    console.log(`Invalid Content-Type ${contentType}, expected ${expected}`);
-    return new Response("Bad Request", { status: 400 });
-  }
-}
-
-async function handleDebugRequest(req: Request) {
+async function handleDebugRequest(req) {
   const url = new URL(req.url);
   const voice = url.searchParams.get("voice") || "";
   const model = url.searchParams.get("model") || "";
   const text = url.searchParams.get("text") || "";
 
-  console.log(`Debug request with model=${model}, voice=${voice}, text=${text}`);
-
   if (!voice || !model || !text) {
-    console.log("Missing required parameters");
     return new Response("Bad Request", { status: 400 });
   }
 
   return synthesizeSpeech(model, voice, text);
 }
 
-async function handleSynthesisRequest(req: Request) {
+async function handleSynthesisRequest(req) {
   if (unauthorized(req)) {
-    console.log("Unauthorized request");
     return new Response("Unauthorized", { status: 401 });
   }
 
   if (req.method !== "POST") {
-    console.log(`Invalid method ${req.method}, expected POST`);
     return new Response("Method Not Allowed", { status: 405 });
   }
 
-  const invalidContentType = validateContentType(req, "application/json");
-  if (invalidContentType) return invalidContentType;
-
-  const { model, input, voice } = await req.json();
-  console.log(`Synthesis request with model=${model}, input=${input}, voice=${voice}`);
+  const contentType = req.headers.get("Content-Type");
+  if (contentType !== "application/json") {
+    return new Response("Bad Request", { status: 400 });
+  }
 
-  return synthesizeSpeech(model, voice, input);
+  try {
+    const { model, input, voice } = await req.json();
+    return synthesizeSpeech(model, voice, input);
+  } catch (error) {
+    return new Response("Invalid JSON", { status: 400 });
+  }
 }
 
+async function handleDemoRequest() {
+  const groupedVoiceList = await fetchVoiceList();
 
-async function handleDemoRequest(req: Request) {
-    const groupedVoiceList = await fetchVoiceList();
-  
-    const html = `<!DOCTYPE html><html lang="en"><head><meta charset="UTF-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><title>语音合成演示</title><link href="https://fonts.googleapis.com/css2?family=Noto+Sans+SC:wght@400;700&display=swap" rel="stylesheet"><style>:root{--primary-color:#6c8bd6;--primary-light:#a2b3e3;--primary-dark:#3d5b8f;--secondary-color:#f08080;--text-color:#333;--text-secondary:#777;--bg-color:#fff}body{font-family:'Noto Sans SC','Arial',sans-serif;color:var(--text-color);margin:0;padding:0;display:flex;justify-content:center;background-color:#fafafa;background-image:linear-gradient(135deg,#f5f7fa 0%,#c3cfe2 100%);position:relative;overflow:hidden}body::before{content:"";position:absolute;top:0;left:0;right:0;bottom:0;background:repeating-radial-gradient(circle at 50% 50%,rgba(255,255,255,0.8) 0%,rgba(255,255,255,0.8) 2%,transparent 2%,transparent 4%,rgba(255,255,255,0.8) 4%,rgba(255,255,255,0.8) 6%,transparent 6%,transparent 8%,rgba(255,255,255,0.8) 8%,rgba(255,255,255,0.8) 10%,transparent 10%),repeating-linear-gradient(45deg,#D4F4FF 0%,#D4F4FF 5%,#E6F9FF 5%,#E6F9FF 10%,#F0FAFF 10%,#F0FAFF 15%,#E6F9FF 15%,#E6F9FF 20%,#D4F4FF 20%,#D4F4FF 25%);background-blend-mode:multiply;opacity:0.8;z-index:-1;animation:glitch 15s infinite}.container{display:flex;max-width:1200px;width:100%;margin:40px;background:#fff;border-radius:12px;position:relative;background-color:rgba(255,255,255,0.8);z-index:1}@keyframes glitch{0%{background-position:0 0,0 0;filter:hue-rotate(0deg)}50%{background-position:10px 10px,-10px 10px;filter:hue-rotate(360deg)}100%{background-position:0 0,0 0;filter:hue-rotate(0deg)}}.input-area,.output-area{padding:30px;width:50%}.input-area{border-right:1px solid #E0E0E0}h1{font-size:36px;color:var(--primary-color);margin-bottom:30px}.filter-section{margin-bottom:30px}.filter-section label{display:block;font-size:16px;color:var(--text-secondary);margin-bottom:10px}.filter-section input{font-size:16px;padding:10px 15px;border:2px solid var(--primary-light);border-radius:8px;outline:none;transition:border-color .3s,box-shadow .3s;width:100%;box-sizing:border-box}.filter-section input:focus{border-color:var(--primary-color);box-shadow:0 0 0 2px var(--primary-light)}.slider-container{margin-bottom:30px}.slider-container label{display:block;font-size:16px;color:var(--text-secondary);margin-bottom:10px}.slider{-webkit-appearance:none;width:100%;height:10px;border-radius:5px;background:linear-gradient(to right,var(--secondary-color) 0%,var(--primary-color) 50%,var(--primary-light) 100%);box-shadow:inset 0 1px 2px rgba(0,0,0,0.1),0 1px rgba(255,255,255,0.1);outline:none;opacity:0.7;-webkit-transition:.2s;transition:opacity .2s;margin-bottom:10px}.slider:hover{opacity:1}.slider::-webkit-slider-thumb{-webkit-appearance:none;appearance:none;width:20px;height:20px;border-radius:50%;background:#fff;border:2px solid var(--primary-color);cursor:pointer}.slider::-moz-range-thumb{width:20px;height:20px;border-radius:50%;background:#fff;border:2px solid var(--primary-color);cursor:pointer}.slider-value{font-size:14px;color:var(--text-secondary)}.textarea-container{margin-bottom:30px}.textarea-container label{display:block;font-size:18px;margin-bottom:10px}.textarea-container textarea{width:100%;padding:10px;font-size:16px;border:2px solid var(--primary-light);border-radius:8px;outline:none;resize:vertical;transition:border-color .3s,box-shadow .3s;box-sizing:border-box;height:200px}.textarea-container textarea:focus{border-color:var(--primary-color);box-shadow:0 0 0 2px var(--primary-light)}.voice-group{margin-bottom:20px;border:2px solid var(--primary-light);border-radius:12px;overflow:hidden;cursor:move;background:#fff}.voice-header{padding:15px 20px;font-size:18px;background:var(--primary-light);color:#fff;cursor:pointer;display:flex;justify-content:space-between;align-items:center}.voice-header:hover{background:var(--primary-color)}.voice-buttons{padding:20px;display:none;gap:12px;flex-wrap:wrap}.voice-button{background:var(--secondary-color);color:#fff;border:none;padding:10px 20px;border-radius:50px;cursor:pointer;transition:filter .3s}.voice-button:hover{filter:brightness(0.9)}.chevron{transition:transform .3s}.voice-group.open .voice-buttons{display:flex}.voice-group.open .chevron{transform:rotate(180deg)}.dragging{opacity:0.5}</style></head><body><div class="container"><div class="input-area"><h1>输��文本</h1><div class="filter-section"><label for="keywords">Speaker筛选:</label><input type="text" id="keywords" value="multilingual,-TW,-CN"></div><div class="slider-container"><label for="rate">语速:</label><input type="range" min="-1" max="1" step="0.1" value="-0.1" class="slider" id="rate"><div class="slider-value" id="rateValue">-0.1</div><label for="pitch">音调:</label><input type="range" min="-1" max="1" step="0.1" value="0.1" class="slider" id="pitch"><div class="slider-value" id="pitchValue">0.1</div></div><div class="textarea-container"><label for="inputText">输入文本:</label><textarea id="inputText">Hello world</textarea></div></div><div class="output-area"><h1>选择语音</h1><div id="voices"></div></div></div><script>const voiceList = ${JSON.stringify(groupedVoiceList)};let audio=null;function filterVoices(){const keywords=document.getElementById('keywords').value.split(',').map(k=>k.trim().toLowerCase());const voicesDiv=document.getElementById('voices');voicesDiv.innerHTML='';const filteredVoices={};for(const[locale,voices]of Object.entries(voiceList)){const filtered=voices.filter(({name,friendlyName})=>keywords.some(keyword=>name.toLowerCase().includes(keyword)||friendlyName.toLowerCase().includes(keyword)));if(filtered.length>0){filteredVoices[locale]=filtered}}for(const[locale,voices]of Object.entries(filteredVoices)){const group=document.createElement('div');group.className='voice-group';group.draggable=true;const header=document.createElement('div');header.className='voice-header';header.textContent=locale.toUpperCase();const chevron=document.createElement('span');chevron.className='chevron';chevron.innerHTML='&#9660;';header.appendChild(chevron);const buttonsContainer=document.createElement('div');buttonsContainer.className='voice-buttons';voices.forEach(({model,name})=>{const button=document.createElement('button');button.className='voice-button';button.textContent=name;button.onclick=()=>synthesize(model);buttonsContainer.appendChild(button)});header.onclick=()=>{group.classList.toggle('open')};group.appendChild(header);group.appendChild(buttonsContainer);voicesDiv.appendChild(group)}addDragDropListeners()}function synthesize(model){const text=document.getElementById('inputText').value||'Hello world';const rate=document.getElementById('rate').value||'-0.1';const pitch=document.getElementById('pitch').value||'0.1';const voice=\`rate:\${rate}|pitch:\${pitch}\`;if(audio){audio.pause();audio.currentTime=0}fetch('/v1/audio/speech',{method:'POST',headers:{'Content-Type':'application/json'},body:JSON.stringify({model,input:text,voice})}).then(response=>response.blob()).then(blob=>{const audioUrl=URL.createObjectURL(blob);audio=new Audio(audioUrl);audio.play()})}function addDragDropListeners(){const voicesDiv=document.getElementById('voices');let draggedItem=null;voicesDiv.addEventListener('dragstart',e=>{draggedItem=e.target;e.target.classList.add('dragging')});voicesDiv.addEventListener('dragend',e=>{e.target.classList.remove('dragging');draggedItem=null});voicesDiv.addEventListener('dragover',e=>{e.preventDefault();const afterElement=getDragAfterElement(voicesDiv,e.clientY);if(afterElement==null){voicesDiv.appendChild(draggedItem)}else{voicesDiv.insertBefore(draggedItem,afterElement)}})}function getDragAfterElement(container,y){const draggableElements=[...container.querySelectorAll('.voice-group:not(.dragging)')];return draggableElements.reduce((closest,child)=>{const box=child.getBoundingClientRect();const offset=y-box.top-box.height/2;if(offset<0&&offset>closest.offset){return{offset:offset,element:child}}else{return closest}},{offset:Number.NEGATIVE_INFINITY}).element}filterVoices();document.getElementById('keywords').addEventListener('input',filterVoices);const rateSlider=document.getElementById('rate');const rateValue=document.getElementById('rateValue');rateSlider.oninput=function(){rateValue.innerHTML=this.value};const pitchSlider=document.getElementById('pitch');const pitchValue=document.getElementById('pitchValue');pitchSlider.oninput=function(){pitchValue.innerHTML=this.value}</script></body></html>`;
-  
-    return new Response(html, {
-      headers: { "Content-Type": "text/html" },
-    });
-  }
-  
+  const html = `<!DOCTYPE html>
+<html lang="zh-CN">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=no">
+    <title>Edge TTS 语音合成</title>
+    <link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&family=Noto+Sans+SC:wght@300;400;500;600;700&display=swap" rel="stylesheet">
+    <style>
+        :root {
+            --primary: #2563eb;
+            --primary-hover: #1d4ed8;
+            --secondary: #64748b;
+            --accent: #0ea5e9;
+            --success: #10b981;
+            --warning: #f59e0b;
+            --error: #ef4444;
+            --background: #ffffff;
+            --surface: #f8fafc;
+            --surface-alt: #f1f5f9;
+            --border: #e2e8f0;
+            --border-light: #f1f5f9;
+            --text-primary: #0f172a;
+            --text-secondary: #475569;
+            --text-muted: #94a3b8;
+            --shadow-sm: 0 1px 2px 0 rgb(0 0 0 / 0.05);
+            --shadow: 0 1px 3px 0 rgb(0 0 0 / 0.1), 0 1px 2px -1px rgb(0 0 0 / 0.1);
+            --shadow-md: 0 4px 6px -1px rgb(0 0 0 / 0.1), 0 2px 4px -2px rgb(0 0 0 / 0.1);
+            --shadow-lg: 0 10px 15px -3px rgb(0 0 0 / 0.1), 0 4px 6px -4px rgb(0 0 0 / 0.1);
+            --radius: 8px;
+            --radius-lg: 12px;
+            --spacing: 1rem;
+        }
+
+        @media (prefers-color-scheme: dark) {
+            :root {
+                --primary: #3b82f6;
+                --primary-hover: #2563eb;
+                --background: #0f172a;
+                --surface: #1e293b;
+                --surface-alt: #334155;
+                --border: #334155;
+                --border-light: #475569;
+                --text-primary: #f1f5f9;
+                --text-secondary: #cbd5e1;
+                --text-muted: #64748b;
+            }
+        }
+
+        * {
+            box-sizing: border-box;
+            margin: 0;
+            padding: 0;
+        }
+
+        body {
+            font-family: 'Inter', 'Noto Sans SC', -apple-system, BlinkMacSystemFont, system-ui, sans-serif;
+            background: var(--background);
+            color: var(--text-primary);
+            line-height: 1.6;
+            font-size: 14px;
+            overflow-x: hidden;
+        }
+
+        .container {
+            max-width: 1400px;
+            margin: 0 auto;
+            padding: 2rem;
+            min-height: 100vh;
+        }
+
+        .header {
+            text-align: center;
+            margin-bottom: 3rem;
+        }
+
+        .header h1 {
+            font-size: clamp(2rem, 5vw, 3rem);
+            font-weight: 700;
+            color: var(--text-primary);
+            margin-bottom: 0.5rem;
+            letter-spacing: -0.025em;
+        }
+
+        .header p {
+            font-size: 1.125rem;
+            color: var(--text-secondary);
+            font-weight: 400;
+        }
+
+        .main-grid {
+            display: grid;
+            grid-template-columns: 1fr;
+            gap: 2rem;
+        }
+
+        @media (min-width: 1024px) {
+            .main-grid {
+                grid-template-columns: 400px 1fr;
+            }
+        }
+
+        .panel {
+            background: var(--surface);
+            border: 1px solid var(--border);
+            border-radius: var(--radius-lg);
+            padding: 2rem;
+            box-shadow: var(--shadow);
+        }
+
+        .panel-title {
+            font-size: 1.25rem;
+            font-weight: 600;
+            color: var(--text-primary);
+            margin-bottom: 1.5rem;
+            display: flex;
+            align-items: center;
+            gap: 0.5rem;
+        }
+
+        .form-group {
+            margin-bottom: 1.5rem;
+        }
+
+        .form-label {
+            display: block;
+            font-weight: 500;
+            color: var(--text-primary);
+            margin-bottom: 0.5rem;
+            font-size: 0.875rem;
+        }
+
+        .form-control {
+            width: 100%;
+            padding: 0.75rem 1rem;
+            border: 1px solid var(--border);
+            border-radius: var(--radius);
+            background: var(--background);
+            color: var(--text-primary);
+            font-family: inherit;
+            font-size: 0.875rem;
+            transition: all 0.2s ease;
+        }
+
+        .form-control:focus {
+            outline: none;
+            border-color: var(--primary);
+            box-shadow: 0 0 0 3px rgb(37 99 235 / 0.1);
+        }
+
+        .form-control::placeholder {
+            color: var(--text-muted);
+        }
+
+        .textarea {
+            min-height: 120px;
+            resize: vertical;
+            font-family: inherit;
+        }
+
+        .filter-tabs {
+            display: flex;
+            gap: 0.5rem;
+            margin-bottom: 1rem;
+            flex-wrap: wrap;
+        }
+
+        .filter-tab {
+            padding: 0.5rem 1rem;
+            border: 1px solid var(--border);
+            border-radius: 20px;
+            background: var(--background);
+            color: var(--text-secondary);
+            cursor: pointer;
+            font-size: 0.8rem;
+            font-weight: 500;
+            transition: all 0.2s ease;
+            white-space: nowrap;
+        }
+
+        .filter-tab:hover {
+            border-color: var(--primary);
+            color: var(--primary);
+        }
+
+        .filter-tab.active {
+            background: var(--primary);
+            border-color: var(--primary);
+            color: white;
+        }
+
+        .slider-group {
+            margin-bottom: 1.5rem;
+        }
+
+        .slider-container {
+            display: flex;
+            align-items: center;
+            gap: 1rem;
+            margin-bottom: 0.75rem;
+        }
+
+        .slider {
+            flex: 1;
+            height: 6px;
+            border-radius: 3px;
+            background: var(--border);
+            outline: none;
+            -webkit-appearance: none;
+            appearance: none;
+            cursor: pointer;
+        }
+
+        .slider::-webkit-slider-thumb {
+            -webkit-appearance: none;
+            appearance: none;
+            width: 20px;
+            height: 20px;
+            border-radius: 50%;
+            background: var(--primary);
+            cursor: pointer;
+            border: 2px solid white;
+            box-shadow: var(--shadow);
+            transition: all 0.2s ease;
+        }
+
+        .slider::-webkit-slider-thumb:hover {
+            transform: scale(1.1);
+            box-shadow: var(--shadow-md);
+        }
+
+        .slider::-moz-range-thumb {
+            width: 20px;
+            height: 20px;
+            border-radius: 50%;
+            background: var(--primary);
+            cursor: pointer;
+            border: 2px solid white;
+            box-shadow: var(--shadow);
+        }
+
+        .slider-value {
+            font-size: 0.8rem;
+            color: var(--text-secondary);
+            font-weight: 500;
+            min-width: 3rem;
+            text-align: center;
+            background: var(--surface-alt);
+            padding: 0.25rem 0.5rem;
+            border-radius: 4px;
+        }
+
+        .voice-container {
+            max-height: 70vh;
+            overflow-y: auto;
+            border: 1px solid var(--border);
+            border-radius: var(--radius);
+            background: var(--background);
+        }
+
+        .voice-container::-webkit-scrollbar {
+            width: 6px;
+        }
+
+        .voice-container::-webkit-scrollbar-track {
+            background: var(--surface);
+        }
+
+        .voice-container::-webkit-scrollbar-thumb {
+            background: var(--border);
+            border-radius: 3px;
+        }
+
+        .voice-container::-webkit-scrollbar-thumb:hover {
+            background: var(--text-muted);
+        }
+
+        .voice-group {
+            border-bottom: 1px solid var(--border-light);
+        }
+
+        .voice-group:last-child {
+            border-bottom: none;
+        }
+
+        .voice-header {
+            padding: 1rem 1.25rem;
+            cursor: pointer;
+            display: flex;
+            justify-content: space-between;
+            align-items: center;
+            background: var(--surface);
+            transition: all 0.2s ease;
+            border-left: 3px solid transparent;
+        }
+
+        .voice-header:hover {
+            background: var(--surface-alt);
+            border-left-color: var(--primary);
+        }
+
+        .voice-header.active {
+            background: var(--primary);
+            color: white;
+            border-left-color: var(--primary-hover);
+        }
+
+        .voice-header-info {
+            display: flex;
+            align-items: center;
+            gap: 0.75rem;
+        }
+
+        .voice-header-title {
+            font-weight: 500;
+            font-size: 0.9rem;
+        }
+
+        .voice-header-count {
+            font-size: 0.75rem;
+            opacity: 0.8;
+            background: rgba(255, 255, 255, 0.2);
+            padding: 0.125rem 0.5rem;
+            border-radius: 10px;
+        }
+
+        .chevron {
+            transition: transform 0.2s ease;
+            font-size: 0.75rem;
+        }
+
+        .voice-group.open .chevron {
+            transform: rotate(180deg);
+        }
+
+        .voice-buttons {
+            padding: 1rem 1.25rem;
+            display: none;
+            gap: 0.5rem;
+            flex-wrap: wrap;
+            background: var(--surface-alt);
+        }
+
+        .voice-group.open .voice-buttons {
+            display: flex;
+        }
+
+        .voice-button {
+            background: white;
+            color: var(--text-primary);
+            border: 1px solid var(--border);
+            padding: 0.5rem 1rem;
+            border-radius: 20px;
+            cursor: pointer;
+            transition: all 0.2s ease;
+            font-size: 0.8rem;
+            font-weight: 500;
+            white-space: nowrap;
+        }
+
+        .voice-button:hover {
+            border-color: var(--primary);
+            color: var(--primary);
+            transform: translateY(-1px);
+            box-shadow: var(--shadow);
+        }
+
+        .voice-button.playing {
+            background: var(--success);
+            color: white;
+            border-color: var(--success);
+            animation: pulse 1.5s infinite;
+        }
+
+        @keyframes pulse {
+            0%, 100% { transform: scale(1); }
+            50% { transform: scale(1.02); }
+        }
+
+        .chinese-voices {
+            background: linear-gradient(135deg, #fef3c7 0%, #fde68a 100%);
+        }
+
+        @media (prefers-color-scheme: dark) {
+            .chinese-voices {
+                background: linear-gradient(135deg, #451a03 0%, #78350f 100%);
+            }
+        }
+
+        .chinese-voices .voice-header {
+            background: rgba(255, 255, 255, 0.9);
+        }
+
+        @media (prefers-color-scheme: dark) {
+            .chinese-voices .voice-header {
+                background: rgba(0, 0, 0, 0.3);
+            }
+        }
+
+        .status-message {
+            margin-top: 1rem;
+            padding: 1rem;
+            border-radius: var(--radius);
+            font-size: 0.875rem;
+            text-align: center;
+            display: none;
+            font-weight: 500;
+        }
+
+        .status-message.show {
+            display: block;
+        }
+
+        .status-message.success {
+            background: #f0fdf4;
+            color: var(--success);
+            border: 1px solid #bbf7d0;
+        }
+
+        .status-message.error {
+            background: #fef2f2;
+            color: var(--error);
+            border: 1px solid #fecaca;
+        }
+
+        @media (prefers-color-scheme: dark) {
+            .status-message.success {
+                background: #064e3b;
+                border-color: #047857;
+            }
+            
+            .status-message.error {
+                background: #7f1d1d;
+                border-color: #dc2626;
+            }
+        }
+
+        .mobile-controls {
+            position: fixed;
+            bottom: 0;
+            left: 0;
+            right: 0;
+            background: var(--surface);
+            padding: 1rem;
+            box-shadow: var(--shadow-lg);
+            border-top: 1px solid var(--border);
+            display: none;
+            z-index: 1000;
+            backdrop-filter: blur(10px);
+        }
+
+        @media (max-width: 1023px) {
+            .mobile-controls {
+                display: block;
+            }
+            
+            .voice-container {
+                margin-bottom: 100px;
+            }
+        }
+
+        .current-audio-info {
+            font-size: 0.8rem;
+            color: var(--text-secondary);
+            margin-bottom: 0.75rem;
+            text-align: center;
+        }
+
+        .audio-controls {
+            display: flex;
+            gap: 0.75rem;
+            align-items: center;
+        }
+
+        .control-button {
+            background: var(--primary);
+            color: white;
+            border: none;
+            padding: 0.75rem 1.5rem;
+            border-radius: var(--radius);
+            cursor: pointer;
+            font-size: 0.875rem;
+            font-weight: 500;
+            transition: all 0.2s ease;
+            flex: 1;
+        }
+
+        .control-button:hover {
+            background: var(--primary-hover);
+            transform: translateY(-1px);
+        }
+
+        .control-button:disabled {
+            background: var(--text-muted);
+            color: var(--background);
+            cursor: not-allowed;
+            transform: none;
+        }
+
+        .loading {
+            display: inline-block;
+            width: 14px;
+            height: 14px;
+            border: 2px solid rgba(255, 255, 255, 0.3);
+            border-radius: 50%;
+            border-top-color: white;
+            animation: spin 1s linear infinite;
+            margin-right: 0.5rem;
+        }
+
+        @keyframes spin {
+            to { transform: rotate(360deg); }
+        }
+
+        @media (max-width: 640px) {
+            .container {
+                padding: 1rem;
+            }
+            
+            .panel {
+                padding: 1.5rem;
+            }
+            
+            .header {
+                margin-bottom: 2rem;
+            }
+            
+            .filter-tabs {
+                gap: 0.25rem;
+            }
+            
+            .filter-tab {
+                font-size: 0.75rem;
+                padding: 0.375rem 0.75rem;
+            }
+        }
+    </style>
+</head>
+
+<body>
+    <div class="container">
+        <div class="header">
+            <h1>Edge TTS 语音合成</h1>
+            <p>高质量的多语言语音合成服务</p>
+        </div>
+
+        <div class="main-grid">
+            <div class="panel">
+                <h2 class="panel-title">
+                    ⚙️ 设置
+                </h2>
+
+                <div class="form-group">
+                    <label class="form-label">语音筛选</label>
+                    <div class="filter-tabs">
+                        <div class="filter-tab active" data-filter="chinese">中文</div>
+                        <div class="filter-tab" data-filter="english">英文</div>
+                        <div class="filter-tab" data-filter="multilingual">多语言</div>
+                        <div class="filter-tab" data-filter="all">全部</div>
+                    </div>
+                    <input type="text" id="customFilter" class="form-control" placeholder="自定义筛选...">
+                </div>
+
+                <div class="slider-group">
+                    <label class="form-label">语速</label>
+                    <div class="slider-container">
+                        <input type="range" min="-1" max="1" step="0.1" value="-0.1" class="slider" id="rate">
+                        <div class="slider-value" id="rateValue">-0.1</div>
+                    </div>
+                </div>
+
+                <div class="slider-group">
+                    <label class="form-label">音调</label>
+                    <div class="slider-container">
+                        <input type="range" min="-1" max="1" step="0.1" value="0.1" class="slider" id="pitch">
+                        <div class="slider-value" id="pitchValue">0.1</div>
+                    </div>
+                </div>
+
+                <div class="form-group">
+                    <label class="form-label" for="inputText">输入文本</label>
+                    <textarea id="inputText" class="form-control textarea" placeholder="请输入要转换的文本...">你好，欢迎使用Edge TTS语音合成服务！这里支持多种中文语音选择。</textarea>
+                </div>
+
+                <div class="status-message" id="statusMessage"></div>
+            </div>
+
+            <div class="panel">
+                <h2 class="panel-title">
+                    🎤 语音选择
+                </h2>
+                
+                <div class="voice-container" id="voices"></div>
+            </div>
+        </div>
+
+        <div class="mobile-controls">
+            <div class="current-audio-info" id="currentAudioInfo">选择语音后开始合成</div>
+            <div class="audio-controls">
+                <button class="control-button" id="pauseBtn" disabled>暂停</button>
+                <button class="control-button" id="stopBtn" disabled>停止</button>
+            </div>
+        </div>
+    </div>
+
+    <script>
+        const voiceList = ${JSON.stringify(groupedVoiceList)};
+        let audio = null;
+        let currentVoiceButton = null;
+        let currentFilter = 'chinese';
+
+        const chineseVoiceMapping = {
+            'zh-CN': '中文 (普通话)',
+            'zh-HK': '中文 (粤语)',
+            'zh-TW': '中文 (台湾话)',
+            'zh-CN-liaoning': '中文 (辽宁话)',
+            'zh-CN-shaanxi': '中文 (陕西话)'
+        };
+
+        const filterPresets = {
+            chinese: ['zh-CN', 'zh-HK', 'zh-TW', 'zh-CN-liaoning', 'zh-CN-shaanxi'],
+            english: ['en-US', 'en-GB', 'en-AU', 'en-CA', 'en-IN'],
+            multilingual: Object.keys(voiceList).filter(locale => 
+                !locale.startsWith('zh-') && !locale.startsWith('en-')
+            ).slice(0, 10)
+        };
+
+        function showStatusMessage(message, type = 'success') {
+            const statusEl = document.getElementById('statusMessage');
+            statusEl.textContent = message;
+            statusEl.className = \`status-message show \${type}\`;
+            setTimeout(() => {
+                statusEl.classList.remove('show');
+            }, 3000);
+        }
+
+        function updateMobileControls(voiceName = '') {
+            const infoEl = document.getElementById('currentAudioInfo');
+            const pauseBtn = document.getElementById('pauseBtn');
+            const stopBtn = document.getElementById('stopBtn');
+            
+            if (voiceName) {
+                infoEl.textContent = \`当前语音: \${voiceName}\`;
+                pauseBtn.disabled = false;
+                stopBtn.disabled = false;
+            } else {
+                infoEl.textContent = '选择语音后开始合成';
+                pauseBtn.disabled = true;
+                stopBtn.disabled = true;
+            }
+        }
+
+        function filterVoices(filterType = 'chinese', customKeyword = '') {
+            const voicesDiv = document.getElementById('voices');
+            voicesDiv.innerHTML = '';
+            
+            let filteredVoices = {};
+            
+            if (filterType === 'all') {
+                filteredVoices = voiceList;
+            } else if (filterPresets[filterType]) {
+                for (const locale of filterPresets[filterType]) {
+                    if (voiceList[locale]) {
+                        filteredVoices[locale] = voiceList[locale];
+                    }
+                }
+            }
+            
+            if (customKeyword.trim()) {
+                const keyword = customKeyword.trim().toLowerCase();
+                const tempFiltered = {};
+                
+                for (const [locale, voices] of Object.entries(filteredVoices)) {
+                    const matchingVoices = voices.filter(voice => 
+                        voice.name.toLowerCase().includes(keyword) || 
+                        voice.friendlyName.toLowerCase().includes(keyword) ||
+                        locale.toLowerCase().includes(keyword)
+                    );
+                    
+                    if (matchingVoices.length > 0) {
+                        tempFiltered[locale] = matchingVoices;
+                    }
+                }
+                
+                filteredVoices = tempFiltered;
+            }
+
+            for (const [locale, voices] of Object.entries(filteredVoices)) {
+                const group = document.createElement('div');
+                group.className = \`voice-group \${filterType === 'chinese' ? 'chinese-voices' : ''}\`;
+                
+                const header = document.createElement('div');
+                header.className = 'voice-header';
+                
+                const displayName = chineseVoiceMapping[locale] || locale.toUpperCase();
+                const headerInfo = document.createElement('div');
+                headerInfo.className = 'voice-header-info';
+                headerInfo.innerHTML = \`
+                    <span class="voice-header-title">\${displayName}</span>
+                    <span class="voice-header-count">\${voices.length}</span>
+                \`;
+                
+                const chevron = document.createElement('span');
+                chevron.className = 'chevron';
+                chevron.innerHTML = '▼';
+                
+                header.appendChild(headerInfo);
+                header.appendChild(chevron);
+                
+                const buttonsContainer = document.createElement('div');
+                buttonsContainer.className = 'voice-buttons';
+                
+                voices.forEach(({model, name, friendlyName}) => {
+                    const button = document.createElement('button');
+                    button.className = 'voice-button';
+                    
+                    const displayName = name.replace(/Neural$/, '').split('-').pop() || name;
+                    button.textContent = displayName;
+                    button.title = friendlyName;
+                    
+                    button.onclick = () => synthesize(model, button, displayName);
+                    buttonsContainer.appendChild(button);
+                });
+                
+                header.onclick = () => {
+                    group.classList.toggle('open');
+                    header.classList.toggle('active');
+                };
+                
+                group.appendChild(header);
+                group.appendChild(buttonsContainer);
+                voicesDiv.appendChild(group);
+                
+                if (filterType === 'chinese') {
+                    group.classList.add('open');
+                    header.classList.add('active');
+                }
+            }
+        }
+
+        function synthesize(model, buttonElement, voiceName) {
+            const text = document.getElementById('inputText').value || '你好，欢迎使用Edge TTS语音合成服务！';
+            const rate = document.getElementById('rate').value || '-0.1';
+            const pitch = document.getElementById('pitch').value || '0.1';
+            const voice = \`rate:\${rate}|pitch:\${pitch}\`;
+            
+            if (currentVoiceButton) {
+                currentVoiceButton.classList.remove('playing');
+                currentVoiceButton.innerHTML = currentVoiceButton.textContent;
+            }
+            
+            currentVoiceButton = buttonElement;
+            buttonElement.classList.add('playing');
+            buttonElement.innerHTML = '<span class="loading"></span>' + buttonElement.textContent;
+            
+            if (audio) {
+                audio.pause();
+                audio.currentTime = 0;
+            }
+            
+            updateMobileControls(voiceName);
+            showStatusMessage('正在合成语音...', 'success');
+            
+            fetch('/v1/audio/speech', {
+                method: 'POST',
+                headers: {'Content-Type': 'application/json'},
+                body: JSON.stringify({model, input: text, voice})
+            })
+            .then(response => {
+                if (!response.ok) {
+                    throw new Error('合成失败');
+                }
+                return response.blob();
+            })
+            .then(blob => {
+                const audioUrl = URL.createObjectURL(blob);
+                audio = new Audio(audioUrl);
+                
+                audio.onplay = () => {
+                    showStatusMessage(\`正在播放: \${voiceName}\`, 'success');
+                };
+                
+                audio.onended = () => {
+                    buttonElement.classList.remove('playing');
+                    buttonElement.innerHTML = buttonElement.textContent;
+                    updateMobileControls();
+                    showStatusMessage('播放完成', 'success');
+                };
+                
+                audio.onerror = () => {
+                    buttonElement.classList.remove('playing');
+                    buttonElement.innerHTML = buttonElement.textContent;
+                    updateMobileControls();
+                    showStatusMessage('播放失败', 'error');
+                };
+                
+                audio.play();
+            })
+            .catch(error => {
+                buttonElement.classList.remove('playing');
+                buttonElement.innerHTML = buttonElement.textContent;
+                updateMobileControls();
+                showStatusMessage('合成失败: ' + error.message, 'error');
+            });
+        }
+
+        document.addEventListener('DOMContentLoaded', function() {
+            document.querySelectorAll('.filter-tab').forEach(tab => {
+                tab.addEventListener('click', function() {
+                    document.querySelectorAll('.filter-tab').forEach(t => t.classList.remove('active'));
+                    this.classList.add('active');
+                    currentFilter = this.dataset.filter;
+                    filterVoices(currentFilter, document.getElementById('customFilter').value);
+                });
+            });
+            
+            document.getElementById('customFilter').addEventListener('input', function() {
+                filterVoices(currentFilter, this.value);
+            });
+            
+            const rateSlider = document.getElementById('rate');
+            const rateValue = document.getElementById('rateValue');
+            rateSlider.oninput = function() {
+                rateValue.textContent = this.value;
+            };
+            
+            const pitchSlider = document.getElementById('pitch');
+            const pitchValue = document.getElementById('pitchValue');
+            pitchSlider.oninput = function() {
+                pitchValue.textContent = this.value;
+            };
+            
+            document.getElementById('pauseBtn').addEventListener('click', function() {
+                if (audio) {
+                    if (audio.paused) {
+                        audio.play();
+                        this.textContent = '暂停';
+                    } else {
+                        audio.pause();
+                        this.textContent = '继续';
+                    }
+                }
+            });
+            
+            document.getElementById('stopBtn').addEventListener('click', function() {
+                if (audio) {
+                    audio.pause();
+                    audio.currentTime = 0;
+                    if (currentVoiceButton) {
+                        currentVoiceButton.classList.remove('playing');
+                        currentVoiceButton.innerHTML = currentVoiceButton.textContent;
+                    }
+                    updateMobileControls();
+                    document.getElementById('pauseBtn').textContent = '暂停';
+                }
+            });
+            
+            filterVoices('chinese');
+        });
+    </script>
+</body>
+</html>`;
+
+  return new Response(html, {
+    headers: { "Content-Type": "text/html" }
+  });
+}
 
 serve(async (req) => {
   try {
     const url = new URL(req.url);
 
     if (url.pathname === "/") {
-      return handleDemoRequest(req);
+      return handleDemoRequest();
     }
 
     if (url.pathname === "/tts") {