feat: 实现语音转文字并支持流式输出；

fix: 修复了语音解密失败的问题

src/components/AnimatedStreamingText.tsx

@@ -0,0 +1,63 @@
+import React, { memo, useEffect, useState, useRef } from 'react'
+
+interface AnimatedStreamingTextProps {
+    text: string
+    className?: string
+    loading?: boolean
+}
+
+export const AnimatedStreamingText = memo(({ text, className, loading }: AnimatedStreamingTextProps) => {
+    const [displayedSegments, setDisplayedSegments] = useState<string[]>([])
+    const prevTextRef = useRef('')
+
+    useEffect(() => {
+        const currentText = (text || '').trim()
+        const prevText = prevTextRef.current
+
+        if (currentText === prevText) return
+        if (!currentText.startsWith(prevText) && prevText !== '') {
+            // 如果不是追加而是全新的文本（比如重新识别），则重置
+            setDisplayedSegments([currentText])
+            prevTextRef.current = currentText
+            return
+        }
+
+        const newPart = currentText.slice(prevText.length)
+        if (newPart) {
+            // 将新部分作为单独的段加入，以触发动画
+            setDisplayedSegments(prev => [...prev, newPart])
+        }
+        prevTextRef.current = currentText
+    }, [text])
+
+    // 处理 loading 状态的显示
+    if (loading && !text) {
+        return <span className={className}>转写中<span className="dot-flashing">...</span></span>
+    }
+
+    return (
+        <span className={className}>
+            {displayedSegments.map((segment, index) => (
+                <span key={index} className="fade-in-text">
+                    {segment}
+                </span>
+            ))}
+            <style>{`
+        .fade-in-text {
+          animation: fadeIn 0.5s ease-out forwards;
+          opacity: 0;
+        }
+        @keyframes fadeIn {
+          from { opacity: 0; transform: translateY(2px); }
+          to { opacity: 1; transform: translateY(0); }
+        }
+        .dot-flashing {
+           animation: blink 1s infinite;
+        }
+        @keyframes blink { 50% { opacity: 0; } }
+      `}</style>
+        </span>
+    )
+})
+
+AnimatedStreamingText.displayName = 'AnimatedStreamingText'

src/components/VoiceTranscribeDialog.scss

@@ -0,0 +1,255 @@
+.voice-transcribe-dialog-overlay {
+    position: fixed;
+    top: 0;
+    left: 0;
+    right: 0;
+    bottom: 0;
+    background: rgba(0, 0, 0, 0.6);
+    backdrop-filter: blur(4px);
+    display: flex;
+    align-items: center;
+    justify-content: center;
+    z-index: 10000;
+    animation: fadeIn 0.2s ease-out;
+}
+
+.voice-transcribe-dialog {
+    background: var(--color-bg-elevated);
+    border-radius: 16px;
+    box-shadow: 0 8px 32px rgba(0, 0, 0, 0.3);
+    width: 90%;
+    max-width: 480px;
+    animation: slideUp 0.3s ease-out;
+    overflow: hidden;
+}
+
+.dialog-header {
+    display: flex;
+    align-items: center;
+    justify-content: space-between;
+    padding: 20px 24px;
+    border-bottom: 1px solid var(--color-border);
+
+    h3 {
+        margin: 0;
+        font-size: 18px;
+        font-weight: 600;
+        color: var(--color-text-primary);
+    }
+
+    .close-button {
+        background: none;
+        border: none;
+        cursor: pointer;
+        padding: 4px;
+        color: var(--color-text-secondary);
+        border-radius: 6px;
+        transition: all 0.15s ease;
+
+        &:hover {
+            background: var(--color-bg-hover);
+            color: var(--color-text-primary);
+        }
+    }
+}
+
+.dialog-content {
+    padding: 24px;
+}
+
+.info-section {
+    display: flex;
+    flex-direction: column;
+    align-items: center;
+    text-align: center;
+    gap: 16px;
+
+    .info-icon {
+        color: var(--color-primary);
+        opacity: 0.8;
+    }
+
+    .info-text {
+        font-size: 15px;
+        color: var(--color-text-primary);
+        margin: 0;
+    }
+
+    .model-info {
+        width: 100%;
+        background: var(--color-bg);
+        border-radius: 12px;
+        padding: 16px;
+        display: flex;
+        flex-direction: column;
+        gap: 12px;
+
+        .model-item {
+            display: flex;
+            justify-content: space-between;
+            align-items: center;
+            font-size: 14px;
+
+            .label {
+                color: var(--color-text-secondary);
+            }
+
+            .value {
+                color: var(--color-text-primary);
+                font-weight: 500;
+            }
+        }
+    }
+}
+
+.download-section {
+    display: flex;
+    flex-direction: column;
+    align-items: center;
+    gap: 16px;
+    padding: 20px 0;
+
+    .download-icon {
+        .downloading-icon {
+            color: var(--color-primary);
+            animation: bounce 1s ease-in-out infinite;
+        }
+    }
+
+    .download-text {
+        font-size: 15px;
+        color: var(--color-text-primary);
+        margin: 0;
+    }
+
+    .progress-bar {
+        width: 100%;
+        height: 6px;
+        background: var(--color-bg);
+        border-radius: 3px;
+        overflow: hidden;
+
+        .progress-fill {
+            height: 100%;
+            background: linear-gradient(90deg, var(--color-primary), var(--color-accent));
+            border-radius: 3px;
+            transition: width 0.3s ease;
+        }
+    }
+
+    .progress-text {
+        font-size: 14px;
+        color: var(--color-text-secondary);
+        margin: 0;
+        font-variant-numeric: tabular-nums;
+    }
+}
+
+.complete-section {
+    display: flex;
+    flex-direction: column;
+    align-items: center;
+    gap: 16px;
+    padding: 20px 0;
+
+    .complete-icon {
+        color: var(--color-success);
+    }
+
+    .complete-text {
+        font-size: 15px;
+        color: var(--color-text-primary);
+        margin: 0;
+    }
+}
+
+.error-message {
+    display: flex;
+    align-items: center;
+    gap: 8px;
+    padding: 12px 16px;
+    background: rgba(239, 68, 68, 0.1);
+    border: 1px solid rgba(239, 68, 68, 0.3);
+    border-radius: 8px;
+    color: #ef4444;
+    font-size: 14px;
+    margin-top: 16px;
+}
+
+.dialog-actions {
+    display: flex;
+    gap: 12px;
+    margin-top: 24px;
+
+    button {
+        flex: 1;
+        padding: 12px 20px;
+        border-radius: 8px;
+        font-size: 14px;
+        font-weight: 500;
+        cursor: pointer;
+        transition: all 0.15s ease;
+        border: none;
+        display: flex;
+        align-items: center;
+        justify-content: center;
+        gap: 6px;
+
+        &.btn-secondary {
+            background: var(--color-bg);
+            color: var(--color-text-primary);
+
+            &:hover {
+                background: var(--color-bg-hover);
+            }
+        }
+
+        &.btn-primary {
+            background: var(--color-primary);
+            color: white;
+
+            &:hover {
+                opacity: 0.9;
+                transform: translateY(-1px);
+            }
+
+            &:active {
+                transform: translateY(0);
+            }
+        }
+    }
+}
+
+@keyframes fadeIn {
+    from {
+        opacity: 0;
+    }
+
+    to {
+        opacity: 1;
+    }
+}
+
+@keyframes slideUp {
+    from {
+        opacity: 0;
+        transform: translateY(20px);
+    }
+
+    to {
+        opacity: 1;
+        transform: translateY(0);
+    }
+}
+
+@keyframes bounce {
+
+    0%,
+    100% {
+        transform: translateY(0);
+    }
+
+    50% {
+        transform: translateY(-10px);
+    }
+}

src/components/VoiceTranscribeDialog.tsx

@@ -0,0 +1,145 @@
+import React, { useState, useEffect } from 'react'
+import { Download, X, CheckCircle, AlertCircle } from 'lucide-react'
+import './VoiceTranscribeDialog.scss'
+
+interface VoiceTranscribeDialogProps {
+    onClose: () => void
+    onDownloadComplete: () => void
+}
+
+export const VoiceTranscribeDialog: React.FC<VoiceTranscribeDialogProps> = ({
+    onClose,
+    onDownloadComplete
+}) => {
+    const [isDownloading, setIsDownloading] = useState(false)
+    const [downloadProgress, setDownloadProgress] = useState(0)
+    const [downloadError, setDownloadError] = useState<string | null>(null)
+    const [isComplete, setIsComplete] = useState(false)
+
+    useEffect(() => {
+        // 监听下载进度
+        const removeListener = window.electronAPI.whisper?.onDownloadProgress?.((payload) => {
+            if (payload.percent !== undefined) {
+                setDownloadProgress(payload.percent)
+            }
+        })
+
+        return () => {
+            removeListener?.()
+        }
+    }, [])
+
+    const handleDownload = async () => {
+        setIsDownloading(true)
+        setDownloadError(null)
+        setDownloadProgress(0)
+
+        try {
+            const result = await window.electronAPI.whisper?.downloadModel()
+
+            if (result?.success) {
+                setIsComplete(true)
+                setDownloadProgress(100)
+
+                // 延迟关闭弹窗并触发转写
+                setTimeout(() => {
+                    onDownloadComplete()
+                }, 1000)
+            } else {
+                setDownloadError(result?.error || '下载失败')
+                setIsDownloading(false)
+            }
+        } catch (error) {
+            setDownloadError(String(error))
+            setIsDownloading(false)
+        }
+    }
+
+    const handleCancel = () => {
+        if (!isDownloading) {
+            onClose()
+        }
+    }
+
+    return (
+        <div className="voice-transcribe-dialog-overlay" onClick={handleCancel}>
+            <div className="voice-transcribe-dialog" onClick={(e) => e.stopPropagation()}>
+                <div className="dialog-header">
+                    <h3>语音转文字</h3>
+                    {!isDownloading && (
+                        <button className="close-button" onClick={onClose}>
+                            <X size={20} />
+                        </button>
+                    )}
+                </div>
+
+                <div className="dialog-content">
+                    {!isDownloading && !isComplete && (
+                        <>
+                            <div className="info-section">
+                                <AlertCircle size={48} className="info-icon" />
+                                <p className="info-text">
+                                    首次使用语音转文字功能需要下载 AI 模型
+                                </p>
+                                <div className="model-info">
+                                    <div className="model-item">
+                                        <span className="label">模型名称：</span>
+                                        <span className="value">SenseVoiceSmall</span>
+                                    </div>
+                                    <div className="model-item">
+                                        <span className="label">文件大小：</span>
+                                        <span className="value">约 240 MB</span>
+                                    </div>
+                                    <div className="model-item">
+                                        <span className="label">支持语言：</span>
+                                        <span className="value">中文、英文、日文、韩文</span>
+                                    </div>
+                                </div>
+                            </div>
+
+                            {downloadError && (
+                                <div className="error-message">
+                                    <AlertCircle size={16} />
+                                    <span>{downloadError}</span>
+                                </div>
+                            )}
+
+                            <div className="dialog-actions">
+                                <button className="btn-secondary" onClick={onClose}>
+                                    取消
+                                </button>
+                                <button className="btn-primary" onClick={handleDownload}>
+                                    <Download size={16} />
+                                    <span>立即下载</span>
+                                </button>
+                            </div>
+                        </>
+                    )}
+
+                    {isDownloading && !isComplete && (
+                        <div className="download-section">
+                            <div className="download-icon">
+                                <Download size={48} className="downloading-icon" />
+                            </div>
+                            <p className="download-text">正在下载模型...</p>
+                            <div className="progress-bar">
+                                <div
+                                    className="progress-fill"
+                                    style={{ width: `${downloadProgress}%` }}
+                                />
+                            </div>
+                            <p className="progress-text">{downloadProgress.toFixed(1)}%</p>
+                        </div>
+                    )}
+
+                    {isComplete && (
+                        <div className="complete-section">
+                            <CheckCircle size={48} className="complete-icon" />
+                            <p className="complete-text">下载完成！正在转写语音...</p>
+                        </div>
+                    )}
+                </div>
+            </div>
+        </div>
+    )
+}

src/pages/ChatPage.scss

@@ -1882,3 +1882,31 @@
     transform: translateX(0);
   }
 }
+/* 语音转文字按钮样式 */
+.voice-transcribe-btn {
+  width: 28px;
+  height: 28px;
+  padding: 0;
+  margin-left: 8px;
+  border: none;
+  background: var(--primary-light);
+  border-radius: 50%;
+  color: var(--primary);
+  cursor: pointer;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  transition: all 0.2s;
+  flex-shrink: 0;
+
+  &:hover {
+    background: var(--primary);
+    color: #fff;
+    transform: scale(1.05);
+  }
+
+  svg {
+    width: 14px;
+    height: 14px;
+  }
+}

src/pages/ChatPage.tsx

@@ -5,8 +5,21 @@ import { useChatStore } from '../stores/chatStore'
 import type { ChatSession, Message } from '../types/models'
 import { getEmojiPath } from 'wechat-emojis'
 import { ImagePreview } from '../components/ImagePreview'
+import { VoiceTranscribeDialog } from '../components/VoiceTranscribeDialog'
+import { AnimatedStreamingText } from '../components/AnimatedStreamingText'
 import './ChatPage.scss'
 
+// 系统消息类型常量
+const SYSTEM_MESSAGE_TYPES = [
+  10000,        // 系统消息
+  266287972401, // 拍一拍
+]
+
+// 判断是否为系统消息
+function isSystemMessage(localType: number): boolean {
+  return SYSTEM_MESSAGE_TYPES.includes(localType)
+}
+
 interface ChatPageProps {
   // 保留接口以备将来扩展
 }
@@ -138,6 +151,8 @@ function ChatPage(_props: ChatPageProps) {
   const [highlightedMessageKeys, setHighlightedMessageKeys] = useState<string[]>([])
   const [isRefreshingSessions, setIsRefreshingSessions] = useState(false)
   const [hasInitialMessages, setHasInitialMessages] = useState(false)
+  const [showVoiceTranscribeDialog, setShowVoiceTranscribeDialog] = useState(false)
+  const [pendingVoiceTranscriptRequest, setPendingVoiceTranscriptRequest] = useState<{ sessionId: string; messageId: string } | null>(null)
 
   // 联系人信息加载控制
   const isEnrichingRef = useRef(false)
@@ -1128,10 +1143,10 @@ function ChatPage(_props: ChatPageProps) {
                   const prevMsg = index > 0 ? messages[index - 1] : undefined
                   const showDateDivider = shouldShowDateDivider(msg, prevMsg)
 
-                  // 显示时间：第一条消息，或者与上一条消息间隔超过5分钟
+                  // 显示时间:第一条消息,或者与上一条消息间隔超过5分钟
                   const showTime = !prevMsg || (msg.createTime - prevMsg.createTime > 300)
                   const isSent = msg.isSend === 1
-                  const isSystem = msg.localType === 10000
+                  const isSystem = isSystemMessage(msg.localType)
 
                   // 系统消息居中显示
                   const wrapperClass = isSystem ? 'system' : (isSent ? 'sent' : 'received')
@@ -1272,6 +1287,35 @@ function ChatPage(_props: ChatPageProps) {
           </div>
         )}
       </div>
+
+      {/* 语音转文字模型下载弹窗 */}
+      {showVoiceTranscribeDialog && (
+        <VoiceTranscribeDialog
+          onClose={() => {
+            setShowVoiceTranscribeDialog(false)
+            setPendingVoiceTranscriptRequest(null)
+          }}
+          onDownloadComplete={async () => {
+            setShowVoiceTranscribeDialog(false)
+            // 下载完成后，继续转写
+            if (pendingVoiceTranscriptRequest) {
+              try {
+                const result = await window.electronAPI.chat.getVoiceTranscript(
+                  pendingVoiceTranscriptRequest.sessionId,
+                  pendingVoiceTranscriptRequest.messageId
+                )
+                if (result.success) {
+                  const cacheKey = `voice-transcript:${pendingVoiceTranscriptRequest.messageId}`
+                  voiceTranscriptCache.set(cacheKey, (result.transcript || '').trim())
+                }
+              } catch (error) {
+                console.error('[ChatPage] 语音转文字失败:', error)
+              }
+            }
+            setPendingVoiceTranscriptRequest(null)
+          }}
+        />
+      )}
     </div>
   )
 }
@@ -1292,7 +1336,7 @@ function MessageBubble({ message, session, showTime, myAvatarUrl, isGroupChat }:
   myAvatarUrl?: string;
   isGroupChat?: boolean;
 }) {
-  const isSystem = message.localType === 10000
+  const isSystem = isSystemMessage(message.localType)
   const isEmoji = message.localType === 47
   const isImage = message.localType === 3
   const isVoice = message.localType === 34
@@ -1570,7 +1614,7 @@ function MessageBubble({ message, session, showTime, myAvatarUrl, isGroupChat }:
     if (!isImage) return
     if (imageLocalPath) return // 已有图片，不需要解密
     if (!message.imageMd5 && !message.imageDatName) return
-    
+
     const container = imageContainerRef.current
     if (!container) return
 
@@ -1612,8 +1656,32 @@ function MessageBubble({ message, session, showTime, myAvatarUrl, isGroupChat }:
     }
   }, [isVoice])
 
+  // 监听流式转写结果
+  useEffect(() => {
+    if (!isVoice) return
+    const removeListener = window.electronAPI.chat.onVoiceTranscriptPartial?.((payload: { msgId: string; text: string }) => {
+      if (payload.msgId === String(message.localId)) {
+        setVoiceTranscript(payload.text)
+        voiceTranscriptCache.set(voiceTranscriptCacheKey, payload.text)
+      }
+    })
+    return () => removeListener?.()
+  }, [isVoice, message.localId, voiceTranscriptCacheKey])
+
   const requestVoiceTranscript = useCallback(async () => {
     if (voiceTranscriptLoading || voiceTranscriptRequestedRef.current) return
+
+    // 检查模型状态
+    const modelStatus = await window.electronAPI.whisper?.getModelStatus()
+    if (!modelStatus?.exists) {
+      // 模型未下载，抛出错误让外层处理
+      const error: any = new Error('MODEL_NOT_DOWNLOADED')
+      error.requiresDownload = true
+      error.sessionId = session.username
+      error.messageId = String(message.localId)
+      throw error
+    }
+
     voiceTranscriptRequestedRef.current = true
     setVoiceTranscriptLoading(true)
     setVoiceTranscriptError(false)
@@ -1627,7 +1695,13 @@ function MessageBubble({ message, session, showTime, myAvatarUrl, isGroupChat }:
         setVoiceTranscriptError(true)
         voiceTranscriptRequestedRef.current = false
       }
-    } catch {
+    } catch (error: any) {
+      // 检查是否是模型未下载错误
+      if (error?.requiresDownload) {
+        // 不显示错误状态，等待用户手动点击转文字按钮时会触发下载弹窗
+        voiceTranscriptRequestedRef.current = false
+        return
+      }
       setVoiceTranscriptError(true)
       voiceTranscriptRequestedRef.current = false
     } finally {
@@ -1635,13 +1709,23 @@ function MessageBubble({ message, session, showTime, myAvatarUrl, isGroupChat }:
     }
   }, [message.localId, session.username, voiceTranscriptCacheKey, voiceTranscriptLoading])
 
+  // 根据设置决定是否自动转写
+  const [autoTranscribeEnabled, setAutoTranscribeEnabled] = useState(false)
+
   useEffect(() => {
+    window.electronAPI.config.get('autoTranscribeVoice').then((value) => {
+      setAutoTranscribeEnabled(value === true)
+    })
+  }, [])
+
+  useEffect(() => {
+    if (!autoTranscribeEnabled) return
     if (!isVoice) return
     if (!voiceDataUrl) return
     if (voiceTranscriptError) return
     if (voiceTranscriptLoading || voiceTranscript !== undefined || voiceTranscriptRequestedRef.current) return
     void requestVoiceTranscript()
-  }, [isVoice, voiceDataUrl, voiceTranscript, voiceTranscriptError, voiceTranscriptLoading, requestVoiceTranscript])
+  }, [autoTranscribeEnabled, isVoice, voiceDataUrl, voiceTranscript, voiceTranscriptError, voiceTranscriptLoading, requestVoiceTranscript])
 
   if (isSystem) {
     return (
@@ -1771,7 +1855,12 @@ function MessageBubble({ message, session, showTime, myAvatarUrl, isGroupChat }:
           setVoiceLoading(true)
           setVoiceError(false)
           try {
-            const result = await window.electronAPI.chat.getVoiceData(session.username, String(message.localId))
+            const result = await window.electronAPI.chat.getVoiceData(
+              session.username,
+              String(message.localId),
+              message.createTime,
+              message.serverId
+            )
             if (result.success && result.data) {
               const url = `data:audio/wav;base64,${result.data}`
               voiceDataUrlCache.set(voiceCacheKey, url)
@@ -1842,6 +1931,22 @@ function MessageBubble({ message, session, showTime, myAvatarUrl, isGroupChat }:
               {showDecryptHint && <span className="voice-hint">点击解密</span>}
               {voiceError && <span className="voice-error">播放失败</span>}
             </div>
+            {/* 转文字按钮 */}
+            {voiceDataUrl && !voiceTranscript && !voiceTranscriptLoading && (
+              <button
+                className="voice-transcribe-btn"
+                onClick={(e) => {
+                  e.stopPropagation()
+                  void requestVoiceTranscript()
+                }}
+                title="转文字"
+                type="button"
+              >
+                <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+                  <path d="M21 15a2 2 0 0 1-2 2H7l-4 4V5a2 2 0 0 1 2-2h14a2 2 0 0 1 2 2z" />
+                </svg>
+              </button>
+            )}
           </div>
           {showTranscript && (
             <div
@@ -1849,7 +1954,16 @@ function MessageBubble({ message, session, showTime, myAvatarUrl, isGroupChat }:
               onClick={handleTranscriptRetry}
               title={voiceTranscriptError ? '点击重试语音转写' : undefined}
             >
-              {transcriptDisplay}
+              {voiceTranscriptError ? (
+                '转写失败，点击重试'
+              ) : !voiceTranscript ? (
+                voiceTranscriptLoading ? '转写中...' : '未识别到文字'
+              ) : (
+                <AnimatedStreamingText
+                  text={transcriptText}
+                  loading={voiceTranscriptLoading}
+                />
+              )}
             </div>
           )}
         </div>

src/pages/ExportPage.tsx

@@ -20,6 +20,7 @@ interface ExportOptions {
   exportImages: boolean
   exportVoices: boolean
   exportEmojis: boolean
+  exportVoiceAsText: boolean
 }
 
 interface ExportResult {
@@ -54,7 +55,8 @@ function ExportPage() {
     exportMedia: false,
     exportImages: true,
     exportVoices: true,
-    exportEmojis: true
+    exportEmojis: true,
+    exportVoiceAsText: false
   })
 
   const loadSessions = useCallback(async () => {
@@ -158,6 +160,7 @@ function ExportPage() {
         exportImages: options.exportMedia && options.exportImages,
         exportVoices: options.exportMedia && options.exportVoices,
         exportEmojis: options.exportMedia && options.exportEmojis,
+        exportVoiceAsText: options.exportMedia && options.exportVoiceAsText,
         dateRange: options.useAllTime ? null : options.dateRange ? {
           start: Math.floor(options.dateRange.start.getTime() / 1000),
           // 将结束日期设置为当天的 23:59:59,以包含当天的所有消息
@@ -372,9 +375,9 @@ function ExportPage() {
                   <span className="slider"></span>
                 </label>
               </div>
-              
+
               <div className="media-option-divider"></div>
-              
+
               <label className={`media-checkbox-row ${!options.exportMedia ? 'disabled' : ''}`}>
                 <div className="media-checkbox-info">
                   <span className="media-checkbox-title">图片</span>
@@ -387,9 +390,9 @@ function ExportPage() {
                   onChange={e => setOptions({ ...options, exportImages: e.target.checked })}
                 />
               </label>
-              
+
               <div className="media-option-divider"></div>
-              
+
               <label className={`media-checkbox-row ${!options.exportMedia ? 'disabled' : ''}`}>
                 <div className="media-checkbox-info">
                   <span className="media-checkbox-title">语音</span>
@@ -402,9 +405,24 @@ function ExportPage() {
                   onChange={e => setOptions({ ...options, exportVoices: e.target.checked })}
                 />
               </label>
-              
+
               <div className="media-option-divider"></div>
-              
+
+              <label className={`media-checkbox-row ${!options.exportMedia ? 'disabled' : ''}`}>
+                <div className="media-checkbox-info">
+                  <span className="media-checkbox-title">语音转文字</span>
+                  <span className="media-checkbox-desc">将语音消息转换为文字导出</span>
+                </div>
+                <input
+                  type="checkbox"
+                  checked={options.exportVoiceAsText}
+                  disabled={!options.exportMedia}
+                  onChange={e => setOptions({ ...options, exportVoiceAsText: e.target.checked })}
+                />
+              </label>
+
+              <div className="media-option-divider"></div>
+
               <label className={`media-checkbox-row ${!options.exportMedia ? 'disabled' : ''}`}>
                 <div className="media-checkbox-info">
                   <span className="media-checkbox-title">表情</span>

src/pages/SettingsPage.tsx

@@ -13,19 +13,6 @@ import './SettingsPage.scss'
 
 type SettingsTab = 'appearance' | 'database' | 'whisper' | 'cache' | 'about'
 
-const whisperModels = [
-  { value: 'tiny', label: 'tiny (75 MB)' },
-  { value: 'base', label: 'base (142 MB)' },
-  { value: 'small', label: 'small (466 MB)' },
-  { value: 'medium', label: 'medium (1.5 GB)' },
-  { value: 'large-v3', label: 'large-v3 (2.9 GB)' }
-]
-
-const whisperSources = [
-  { value: 'official', label: 'HuggingFace 官方' },
-  { value: 'tsinghua', label: '清华镜像 (hf-mirror)' }
-]
-
 const tabs: { id: SettingsTab; label: string; icon: React.ElementType }[] = [
   { id: 'appearance', label: '外观', icon: Palette },
   { id: 'database', label: '数据库连接', icon: Database },
@@ -57,10 +44,10 @@ function SettingsPage() {
   const [logEnabled, setLogEnabled] = useState(false)
   const [whisperModelName, setWhisperModelName] = useState('base')
   const [whisperModelDir, setWhisperModelDir] = useState('')
-  const [whisperDownloadSource, setWhisperDownloadSource] = useState('tsinghua')
   const [isWhisperDownloading, setIsWhisperDownloading] = useState(false)
   const [whisperDownloadProgress, setWhisperDownloadProgress] = useState(0)
-  const [whisperModelStatus, setWhisperModelStatus] = useState<{ exists: boolean; path?: string } | null>(null)
+  const [whisperModelStatus, setWhisperModelStatus] = useState<{ exists: boolean; modelPath?: string; tokensPath?: string } | null>(null)
+  const [autoTranscribeVoice, setAutoTranscribeVoice] = useState(false)
 
   const [isLoading, setIsLoadingState] = useState(false)
   const [isTesting, setIsTesting] = useState(false)
@@ -124,7 +111,7 @@ function SettingsPage() {
       const savedImageAesKey = await configService.getImageAesKey()
       const savedWhisperModelName = await configService.getWhisperModelName()
       const savedWhisperModelDir = await configService.getWhisperModelDir()
-      const savedWhisperSource = await configService.getWhisperDownloadSource()
+      const savedAutoTranscribe = await configService.getAutoTranscribeVoice()
 
       if (savedKey) setDecryptKey(savedKey)
       if (savedPath) setDbPath(savedPath)
@@ -135,9 +122,8 @@ function SettingsPage() {
       }
       if (savedImageAesKey) setImageAesKey(savedImageAesKey)
       setLogEnabled(savedLogEnabled)
-      if (savedWhisperModelName) setWhisperModelName(savedWhisperModelName)
+      setAutoTranscribeVoice(savedAutoTranscribe)
       if (savedWhisperModelDir) setWhisperModelDir(savedWhisperModelDir)
-      if (savedWhisperSource) setWhisperDownloadSource(savedWhisperSource)
     } catch (e) {
       console.error('加载配置失败:', e)
     }
@@ -145,14 +131,15 @@ function SettingsPage() {
 
 
 
-  const refreshWhisperStatus = async (modelNameValue = whisperModelName, modelDirValue = whisperModelDir) => {
+  const refreshWhisperStatus = async (modelDirValue = whisperModelDir) => {
     try {
-      const result = await window.electronAPI.whisper?.getModelStatus({
-        modelName: modelNameValue,
-        downloadDir: modelDirValue || undefined
-      })
+      const result = await window.electronAPI.whisper?.getModelStatus()
       if (result?.success) {
-        setWhisperModelStatus({ exists: Boolean(result.exists), path: result.path })
+        setWhisperModelStatus({
+          exists: Boolean(result.exists),
+          modelPath: result.modelPath,
+          tokensPath: result.tokensPath
+        })
       }
     } catch {
       setWhisperModelStatus(null)
@@ -178,17 +165,16 @@ function SettingsPage() {
 
   useEffect(() => {
     const removeListener = window.electronAPI.whisper?.onDownloadProgress?.((payload) => {
-      if (payload.modelName !== whisperModelName) return
       if (typeof payload.percent === 'number') {
         setWhisperDownloadProgress(payload.percent)
       }
     })
     return () => removeListener?.()
-  }, [whisperModelName])
+  }, [])
 
   useEffect(() => {
-    void refreshWhisperStatus(whisperModelName, whisperModelDir)
-  }, [whisperModelName, whisperModelDir])
+    void refreshWhisperStatus(whisperModelDir)
+  }, [whisperModelDir])
 
   const handleCheckUpdate = async () => {
     setIsCheckingUpdate(true)
@@ -331,30 +317,21 @@ function SettingsPage() {
     await configService.setWhisperModelName(value)
   }
 
-  const handleWhisperSourceChange = async (value: string) => {
-    setWhisperDownloadSource(value)
-    await configService.setWhisperDownloadSource(value)
-  }
-
   const handleDownloadWhisperModel = async () => {
     if (isWhisperDownloading) return
     setIsWhisperDownloading(true)
     setWhisperDownloadProgress(0)
     try {
-      const result = await window.electronAPI.whisper.downloadModel({
-        modelName: whisperModelName,
-        downloadDir: whisperModelDir || undefined,
-        source: whisperDownloadSource
-      })
+      const result = await window.electronAPI.whisper.downloadModel()
       if (result.success) {
         setWhisperDownloadProgress(100)
-        showMessage('Whisper 模型下载完成', true)
-        await refreshWhisperStatus(whisperModelName, whisperModelDir)
+        showMessage('SenseVoiceSmall 模型下载完成', true)
+        await refreshWhisperStatus(whisperModelDir)
       } else {
-        showMessage(result.error || 'Whisper 模型下载失败', false)
+        showMessage(result.error || '模型下载失败', false)
       }
     } catch (e) {
-      showMessage(`Whisper 模型下载失败: ${e}`, false)
+      showMessage(`模型下载失败: ${e}`, false)
     } finally {
       setIsWhisperDownloading(false)
     }
@@ -475,9 +452,8 @@ function SettingsPage() {
       } else {
         await configService.setImageAesKey('')
       }
-      await configService.setWhisperModelName(whisperModelName)
       await configService.setWhisperModelDir(whisperModelDir)
-      await configService.setWhisperDownloadSource(whisperDownloadSource)
+      await configService.setAutoTranscribeVoice(autoTranscribeVoice)
       await configService.setOnboardingDone(true)
 
       showMessage('配置保存成功，正在测试连接...', true)
@@ -513,9 +489,8 @@ function SettingsPage() {
       setWxid('')
       setCachePath('')
       setLogEnabled(false)
-      setWhisperModelName('base')
+      setAutoTranscribeVoice(false)
       setWhisperModelDir('')
-      setWhisperDownloadSource('tsinghua')
       setWhisperModelStatus(null)
       setWhisperDownloadProgress(0)
       setIsWhisperDownloading(false)
@@ -674,14 +649,14 @@ function SettingsPage() {
         <label>账号 wxid</label>
         <span className="form-hint">微信账号标识</span>
         <div className="wxid-input-wrapper" ref={wxidDropdownRef}>
-          <input 
-            type="text" 
-            placeholder="例如: wxid_xxxxxx" 
-            value={wxid} 
-            onChange={(e) => setWxid(e.target.value)} 
+          <input
+            type="text"
+            placeholder="例如: wxid_xxxxxx"
+            value={wxid}
+            onChange={(e) => setWxid(e.target.value)}
           />
-          <button 
-            type="button" 
+          <button
+            type="button"
             className={`wxid-dropdown-btn ${showWxidSelect ? 'open' : ''}`}
             onClick={() => wxidOptions.length > 0 ? setShowWxidSelect(!showWxidSelect) : handleScanWxid()}
             title={wxidOptions.length > 0 ? "选择已检测到的账号" : "扫描账号"}
@@ -691,8 +666,8 @@ function SettingsPage() {
           {showWxidSelect && wxidOptions.length > 0 && (
             <div className="wxid-dropdown">
               {wxidOptions.map((opt) => (
-                <div 
-                  key={opt.wxid} 
+                <div
+                  key={opt.wxid}
                   className={`wxid-option ${opt.wxid === wxid ? 'active' : ''}`}
                   onClick={() => handleSelectWxid(opt.wxid)}
                 >
@@ -759,34 +734,31 @@ function SettingsPage() {
   )
   const renderWhisperTab = () => (
     <div className="tab-content">
-      <p className="section-desc">语音解密后自动转写为文字</p>
-      <div className="form-group whisper-section">
-        <label>语音识别模型 (Whisper)</label>
-        <span className="form-hint">语音解密后自动转文字，模型越大越准确但下载更慢</span>
-        <div className="whisper-grid">
-          <div className="whisper-field">
-            <span className="field-label">模型</span>
-            <select
-              value={whisperModelName}
-              onChange={(e) => handleWhisperModelChange(e.target.value)}
-            >
-              {whisperModels.map((model) => (
-                <option key={model.value} value={model.value}>{model.label}</option>
-              ))}
-            </select>
-          </div>
-          <div className="whisper-field">
-            <span className="field-label">下载源</span>
-            <select
-              value={whisperDownloadSource}
-              onChange={(e) => handleWhisperSourceChange(e.target.value)}
-            >
-              {whisperSources.map((source) => (
-                <option key={source.value} value={source.value}>{source.label}</option>
-              ))}
-            </select>
-          </div>
+      <div className="form-group">
+        <label>自动语音转文字</label>
+        <span className="form-hint">语音解密后自动转写为文字（需下载模型）</span>
+        <div className="log-toggle-line">
+          <span className="log-status">{autoTranscribeVoice ? '已开启' : '已关闭'}</span>
+          <label className="switch" htmlFor="auto-transcribe-toggle">
+            <input
+              id="auto-transcribe-toggle"
+              className="switch-input"
+              type="checkbox"
+              checked={autoTranscribeVoice}
+              onChange={async (e) => {
+                const enabled = e.target.checked
+                setAutoTranscribeVoice(enabled)
+                await configService.setAutoTranscribeVoice(enabled)
+                showMessage(enabled ? '已开启自动转文字' : '已关闭自动转文字', true)
+              }}
+            />
+            <span className="switch-slider" />
+          </label>
         </div>
+      </div>
+      <div className="form-group whisper-section">
+        <label>语音识别模型 (SenseVoiceSmall)</label>
+        <span className="form-hint">基于 Sherpa-onnx，支持中文、英文、日文、韩文</span>
         <span className="form-hint">模型下载目录</span>
         <input
           type="text"
@@ -801,9 +773,9 @@ function SettingsPage() {
         </div>
         <div className="whisper-status-line">
           <span className={`status ${whisperModelStatus?.exists ? 'ok' : 'warn'}`}>
-            {whisperModelStatus?.exists ? '已下载' : '未下载'}
+            {whisperModelStatus?.exists ? '已下载 (240 MB)' : '未下载 (240 MB)'}
           </span>
-          {whisperModelStatus?.path && <span className="path">{whisperModelStatus.path}</span>}
+          {whisperModelStatus?.modelPath && <span className="path">{whisperModelStatus.modelPath}</span>}
         </div>
         {isWhisperDownloading ? (
           <div className="whisper-progress">
@@ -917,8 +889,8 @@ function SettingsPage() {
             </div>
             <div className="wxid-dialog-list">
               {wxidOptions.map((opt) => (
-                <div 
-                  key={opt.wxid} 
+                <div
+                  key={opt.wxid}
                   className={`wxid-dialog-item ${opt.wxid === wxid ? 'active' : ''}`}
                   onClick={() => handleSelectWxid(opt.wxid)}
                 >

src/services/config.ts

@@ -20,7 +20,8 @@ export const CONFIG_KEYS = {
   IMAGE_AES_KEY: 'imageAesKey',
   WHISPER_MODEL_NAME: 'whisperModelName',
   WHISPER_MODEL_DIR: 'whisperModelDir',
-  WHISPER_DOWNLOAD_SOURCE: 'whisperDownloadSource'
+  WHISPER_DOWNLOAD_SOURCE: 'whisperDownloadSource',
+  AUTO_TRANSCRIBE_VOICE: 'autoTranscribeVoice'
 } as const
 
 // 获取解密密钥
@@ -218,3 +219,14 @@ export async function getOnboardingDone(): Promise<boolean> {
 export async function setOnboardingDone(done: boolean): Promise<void> {
   await config.set(CONFIG_KEYS.ONBOARDING_DONE, done)
 }
+
+// 获取自动语音转文字开关
+export async function getAutoTranscribeVoice(): Promise<boolean> {
+  const value = await config.get(CONFIG_KEYS.AUTO_TRANSCRIBE_VOICE)
+  return value === true
+}
+
+// 设置自动语音转文字开关
+export async function setAutoTranscribeVoice(enabled: boolean): Promise<void> {
+  await config.set(CONFIG_KEYS.AUTO_TRANSCRIBE_VOICE, enabled)
+}

src/types/electron.d.ts

@@ -94,8 +94,9 @@ export interface ElectronAPI {
       error?: string
     }>
     getImageData: (sessionId: string, msgId: string) => Promise<{ success: boolean; data?: string; error?: string }>
-    getVoiceData: (sessionId: string, msgId: string) => Promise<{ success: boolean; data?: string; error?: string }>
+    getVoiceData: (sessionId: string, msgId: string, createTime?: number, serverId?: string | number) => Promise<{ success: boolean; data?: string; error?: string }>
     getVoiceTranscript: (sessionId: string, msgId: string) => Promise<{ success: boolean; transcript?: string; error?: string }>
+    onVoiceTranscriptPartial: (callback: (payload: { msgId: string; text: string }) => void) => () => void
   }
 
   image: {
@@ -297,8 +298,8 @@ export interface ElectronAPI {
     }>
   }
   whisper: {
-    downloadModel: (payload: { modelName: string; downloadDir?: string; source?: string }) => Promise<{ success: boolean; path?: string; error?: string }>
-    getModelStatus: (payload: { modelName: string; downloadDir?: string }) => Promise<{ success: boolean; exists?: boolean; path?: string; sizeBytes?: number; error?: string }>
+    downloadModel: () => Promise<{ success: boolean; modelPath?: string; tokensPath?: string; error?: string }>
+    getModelStatus: () => Promise<{ success: boolean; exists?: boolean; modelPath?: string; tokensPath?: string; sizeBytes?: number; error?: string }>
     onDownloadProgress: (callback: (payload: { modelName: string; downloadedBytes: number; totalBytes?: number; percent?: number }) => void) => () => void
   }
 }