diff --git a/public/禁用更新.png b/public/禁用更新.png
new file mode 100644
index 0000000..d6b4621
Binary files /dev/null and b/public/禁用更新.png differ
diff --git a/src/main/index.ts b/src/main/index.ts
index 433b860..cb0266b 100644
--- a/src/main/index.ts
+++ b/src/main/index.ts
@@ -1800,7 +1800,7 @@ app.whenReady().then(async () => {
   ipcMain.handle('wechat-personal:selectImage', async (event) => {
     const window = BrowserWindow.fromWebContents(event.sender)
     const result = await dialog.showOpenDialog(window!, {
-      title: '选择要通过个人微信测试发送的图片',
+      title: '选择要通过个人微信发送的图片',
       properties: ['openFile'],
       filters: [
         { name: '图片', extensions: ['png', 'jpg', 'jpeg', 'gif', 'webp'] },
@@ -1817,7 +1817,7 @@ app.whenReady().then(async () => {
   ipcMain.handle('wechat-personal:selectVoice', async (event) => {
     const window = BrowserWindow.fromWebContents(event.sender)
     const result = await dialog.showOpenDialog(window!, {
-      title: '选择要通过个人微信测试发送的语音',
+      title: '选择要通过个人微信发送的语音',
       properties: ['openFile'],
       filters: [
         { name: '语音', extensions: ['silk', 'mp3', 'wav', 'm4a', 'aac', 'ogg', 'flac'] },
diff --git a/src/main/services/personal-wechat-send-service.ts b/src/main/services/personal-wechat-send-service.ts
index 1ee2138..fc7244e 100644
--- a/src/main/services/personal-wechat-send-service.ts
+++ b/src/main/services/personal-wechat-send-service.ts
@@ -500,7 +500,7 @@ export class PersonalWechatSendService {
       }
       if (text.length > 2_000) {
         const status = await this.getStatus()
-        return { success: false, status, error: '测试消息不能超过 2000 个字符' }
+        return { success: false, status, error: '消息不能超过 2000 个字符' }
       }
       request = { ...request, to, text }
     } else {
@@ -520,7 +520,7 @@ export class PersonalWechatSendService {
         return {
           success: false,
           status,
-          error: `测试${request.type === 'voice' ? '语音' : '图片'}必须小于 20 MB`
+          error: `${request.type === 'voice' ? '语音' : '图片'}必须小于 20 MB`
         }
       }
       fileBase64 = (
@@ -725,7 +725,7 @@ export class PersonalWechatSendService {
         status: {
           ...base,
           state: 'unsupported_platform',
-          message: '个人微信测试发送当前仅支持 Apple Silicon Mac'
+          message: '个人微信发送当前仅支持 Apple Silicon Mac'
         }
       }
     }
diff --git a/src/main/wcdb4-client.ts b/src/main/wcdb4-client.ts
index 5a3c641..189b3e7 100644
--- a/src/main/wcdb4-client.ts
+++ b/src/main/wcdb4-client.ts
@@ -979,6 +979,11 @@ export class Wcdb4Client {
     this.cachedSessions = null
     this.cachedChatTables = null
     this.sessionsInFlight = null
+    // Session rows can change their folded/muted flags independently of the
+    // contact list. Invalidate the status snapshot together with sessions so
+    // the next hydrated contact load reads the current WeChat state.
+    this.sessionStatusCache.clear()
+    this.sessionStatusesUpdatedAt = 0
     this.sessionDisplayNamesHydrated = false
   }
 
diff --git a/src/renderer/src/components/chat/PersonalWechatChatComposer.tsx b/src/renderer/src/components/chat/PersonalWechatChatComposer.tsx
new file mode 100644
index 0000000..7449460
--- /dev/null
+++ b/src/renderer/src/components/chat/PersonalWechatChatComposer.tsx
@@ -0,0 +1,411 @@
+import { useCallback, useEffect, useRef, useState } from 'react'
+import type {
+  PersonalWechatSendRequest,
+  PersonalWechatSenderStatus
+} from '../../../../shared/personal-wechat'
+import type { TextToSpeechSettings, TextToSpeechVoice } from '../../../../shared/text-to-speech'
+import { Button, SegmentedControl, SegmentedControlItem, Textarea } from '../ui'
+
+export type PersonalWechatComposerMode = 'text' | 'image' | 'voice'
+type VoiceSource = 'generated' | 'file'
+type SelectedLocalFile = { path: string; name: string }
+type GeneratedVoice = { filePath: string; audioDataUrl: string }
+
+export type ChatMessage = {
+  id: string
+  type: PersonalWechatComposerMode | 'system'
+  text?: string
+  fileName?: string
+  outgoing: boolean
+}
+
+interface PersonalWechatChatComposerProps {
+  status: PersonalWechatSenderStatus
+  targetId: string
+  isGroupChat: boolean
+  initialMode?: PersonalWechatComposerMode
+  initialImage?: SelectedLocalFile | null
+  onOpenTextToSpeechSettings?: () => void
+  onCancel: () => void
+  onSend: (
+    request: PersonalWechatSendRequest,
+    displayText: string
+  ) => Promise<{ success: boolean; error?: string }>
+  onMessage: (message: ChatMessage) => void
+  busy: boolean
+}
+
+function formatAudioTime(value: number): string {
+  if (!Number.isFinite(value) || value < 0) return '0:00'
+  const seconds = Math.floor(value)
+  return `${Math.floor(seconds / 60)}:${String(seconds % 60).padStart(2, '0')}`
+}
+
+export function PersonalWechatChatComposer({
+  status,
+  targetId,
+  isGroupChat,
+  initialMode = 'text',
+  initialImage = null,
+  onOpenTextToSpeechSettings,
+  onCancel,
+  onSend,
+  onMessage,
+  busy
+}: PersonalWechatChatComposerProps): React.ReactElement {
+  const [mode, setMode] = useState<PersonalWechatComposerMode>(initialMode)
+  const [text, setText] = useState('')
+  const [image, setImage] = useState<SelectedLocalFile | null>(initialImage)
+  const [voice, setVoice] = useState<SelectedLocalFile | null>(null)
+  const [voiceSource, setVoiceSource] = useState<VoiceSource>('generated')
+  const [voiceText, setVoiceText] = useState('')
+  const [ttsSettings, setTtsSettings] = useState<TextToSpeechSettings | null>(null)
+  const [ttsVoices, setTtsVoices] = useState<TextToSpeechVoice[]>([])
+  const [generatedVoice, setGeneratedVoice] = useState<GeneratedVoice | null>(null)
+  const [isGenerating, setIsGenerating] = useState(false)
+  const [isPreviewPlaying, setIsPreviewPlaying] = useState(false)
+  const [previewCurrentTime, setPreviewCurrentTime] = useState(0)
+  const [previewDuration, setPreviewDuration] = useState(0)
+  const [feedback, setFeedback] = useState<{ success: boolean; message: string } | null>(null)
+  const generatedVoiceRef = useRef<GeneratedVoice | null>(null)
+  const generatedAudioRef = useRef<HTMLAudioElement | null>(null)
+  const selectedTypeReady =
+    mode === 'text'
+      ? status.canSendText
+      : mode === 'image'
+        ? status.canSendImage
+        : status.canSendVoice
+  const isBusy = busy || isGenerating
+  const selectedTtsVoice = ttsVoices.find((item) => item.id === ttsSettings?.selectedVoiceId)
+
+  useEffect(() => {
+    let active = true
+    void window.api
+      .getTextToSpeechSettings()
+      .then(async (response) => {
+        if (!active) return
+        setTtsSettings(response.settings)
+        if (!response.settings.hasApiKey) return
+        const voicesResponse = await window.api.listTextToSpeechVoices({
+          pageNumber: 1,
+          pageSize: 24
+        })
+        if (active && voicesResponse.success) setTtsVoices(voicesResponse.items)
+      })
+      .catch(() => undefined)
+    return () => {
+      active = false
+    }
+  }, [])
+
+  const clearGeneratedVoice = useCallback((): void => {
+    const current = generatedVoiceRef.current
+    generatedVoiceRef.current = null
+    generatedAudioRef.current?.pause()
+    setGeneratedVoice(null)
+    setIsPreviewPlaying(false)
+    setPreviewCurrentTime(0)
+    setPreviewDuration(0)
+    if (current?.filePath)
+      void window.api.removeGeneratedTextToSpeechAudio(current.filePath).catch(() => undefined)
+  }, [])
+
+  useEffect(() => () => clearGeneratedVoice(), [clearGeneratedVoice])
+
+  const selectImage = async (): Promise<void> => {
+    if (isBusy) return
+    const selection = await window.api.selectPersonalWechatImage()
+    if (!selection.canceled && selection.path) {
+      setImage({
+        path: selection.path,
+        name: selection.name || selection.path.split(/[\\/]/).pop() || '图片'
+      })
+      setFeedback(null)
+    }
+  }
+
+  const selectVoice = async (): Promise<void> => {
+    if (isBusy) return
+    const selection = await window.api.selectPersonalWechatVoice()
+    if (!selection.canceled && selection.path) {
+      setVoice({
+        path: selection.path,
+        name: selection.name || selection.path.split(/[\\/]/).pop() || '语音'
+      })
+      setFeedback(null)
+    }
+  }
+
+  const send = async (
+    request: PersonalWechatSendRequest,
+    displayText: string
+  ): Promise<boolean> => {
+    if (isBusy || !selectedTypeReady) return false
+    setFeedback(null)
+    const result = await onSend(request, displayText)
+    setFeedback({
+      success: result.success,
+      message: result.success ? '已发送' : result.error || '发送失败'
+    })
+    if (result.success)
+      onMessage({
+        id: `${Date.now()}-${Math.random()}`,
+        type: request.type,
+        text: request.type === 'text' ? request.text : displayText,
+        fileName: request.type === 'text' ? undefined : displayText,
+        outgoing: true
+      })
+    return result.success
+  }
+
+  const generateVoice = async (): Promise<void> => {
+    if (isBusy || !ttsSettings?.hasApiKey || !ttsSettings.selectedVoiceId || !voiceText.trim())
+      return
+    clearGeneratedVoice()
+    setIsGenerating(true)
+    setFeedback(null)
+    try {
+      const generated = await window.api.synthesizeTextToSpeech({
+        text: voiceText.trim(),
+        referenceId: ttsSettings.selectedVoiceId
+      })
+      if (!generated.success || !generated.filePath || !generated.audioDataUrl) {
+        setFeedback({ success: false, message: generated.error || '语音生成失败' })
+        return
+      }
+      const nextVoice = { filePath: generated.filePath, audioDataUrl: generated.audioDataUrl }
+      generatedVoiceRef.current = nextVoice
+      setGeneratedVoice(nextVoice)
+    } catch (error) {
+      setFeedback({
+        success: false,
+        message: error instanceof Error ? error.message : String(error)
+      })
+    } finally {
+      setIsGenerating(false)
+    }
+  }
+
+  const sendCurrent = async (): Promise<void> => {
+    if (mode === 'text') {
+      const value = text.trim()
+      if (value)
+        await send({ type: 'text', to: targetId, text: value, isGroup: isGroupChat }, value)
+    } else if (mode === 'image' && image) {
+      await send(
+        { type: 'image', to: targetId, filePath: image.path, isGroup: isGroupChat },
+        image.name
+      )
+    } else if (mode === 'voice' && voiceSource === 'file' && voice) {
+      await send(
+        { type: 'voice', to: targetId, filePath: voice.path, isGroup: isGroupChat },
+        voice.name
+      )
+    }
+  }
+
+  const sendGenerated = async (): Promise<void> => {
+    if (!generatedVoiceRef.current) return
+    const result = await send(
+      {
+        type: 'voice',
+        to: targetId,
+        filePath: generatedVoiceRef.current.filePath,
+        isGroup: isGroupChat
+      },
+      voiceText.trim() || '语音消息'
+    )
+    if (result) clearGeneratedVoice()
+  }
+
+  const canSend = Boolean(
+    selectedTypeReady &&
+    !isBusy &&
+    ((mode === 'text' && text.trim()) ||
+      (mode === 'image' && image) ||
+      (mode === 'voice' &&
+        ((voiceSource === 'file' && voice) || (voiceSource === 'generated' && generatedVoice))))
+  )
+  const generatedReady = Boolean(
+    ttsSettings?.hasApiKey && ttsSettings.selectedVoiceId && voiceText.trim()
+  )
+  const previewProgress = previewDuration > 0 ? (previewCurrentTime / previewDuration) * 100 : 0
+
+  return (
+    <section className="personal-wechat-composer" aria-label="发送消息">
+      <SegmentedControl
+        className="personal-wechat-composer-mode"
+        aria-label="消息类型"
+        value={mode}
+        onValueChange={(value) => {
+          clearGeneratedVoice()
+          setMode(value as PersonalWechatComposerMode)
+          setFeedback(null)
+        }}
+        disabled={isBusy}
+      >
+        <SegmentedControlItem value="text">文字</SegmentedControlItem>
+        <SegmentedControlItem value="image">图片</SegmentedControlItem>
+        <SegmentedControlItem value="voice">语音</SegmentedControlItem>
+      </SegmentedControl>
+
+      {mode === 'text' && (
+        <div className="personal-wechat-text-editor">
+          <Textarea
+            aria-label="消息内容"
+            placeholder="输入消息"
+            value={text}
+            maxLength={2000}
+            rows={3}
+            disabled={isBusy}
+            onChange={(event) => {
+              setText(event.target.value)
+              setFeedback(null)
+            }}
+          />
+          <span>{text.length} / 2000</span>
+        </div>
+      )}
+
+      {mode === 'image' && (
+        <div className="personal-wechat-file-editor">
+          <div>
+            <strong>{image?.name || '还没有选择图片'}</strong>
+            <small>{image ? image.path : '支持 PNG、JPG、GIF 和 WebP，最大 20 MB'}</small>
+          </div>
+          <Button variant="outline" size="sm" onClick={() => void selectImage()} disabled={isBusy}>
+            {image ? '重新选择' : '选择图片'}
+          </Button>
+        </div>
+      )}
+
+      {mode === 'voice' && (
+        <div className="personal-wechat-voice-editor">
+          <SegmentedControl
+            aria-label="语音来源"
+            value={voiceSource}
+            onValueChange={(value) => {
+              clearGeneratedVoice()
+              setVoiceSource(value as VoiceSource)
+              setFeedback(null)
+            }}
+            disabled={isBusy}
+          >
+            <SegmentedControlItem value="generated">输入文字生成</SegmentedControlItem>
+            <SegmentedControlItem value="file">选择本地文件</SegmentedControlItem>
+          </SegmentedControl>
+          {voiceSource === 'generated' ? (
+            <>
+              <div className="personal-wechat-voice-heading">
+                <span>{selectedTtsVoice?.name || '尚未选择音色'}</span>
+                {onOpenTextToSpeechSettings && (
+                  <Button variant="link" size="sm" onClick={onOpenTextToSpeechSettings}>
+                    语音设置
+                  </Button>
+                )}
+              </div>
+              <Textarea
+                aria-label="语音文字"
+                placeholder="输入要生成的文字"
+                value={voiceText}
+                maxLength={1000}
+                rows={2}
+                disabled={isBusy}
+                onChange={(event) => {
+                  clearGeneratedVoice()
+                  setVoiceText(event.target.value)
+                  setFeedback(null)
+                }}
+              />
+              {generatedVoice ? (
+                <div className="personal-wechat-generated-result">
+                  <audio
+                    ref={generatedAudioRef}
+                    src={generatedVoice.audioDataUrl}
+                    preload="metadata"
+                    onLoadedMetadata={(event) => setPreviewDuration(event.currentTarget.duration)}
+                    onTimeUpdate={(event) => setPreviewCurrentTime(event.currentTarget.currentTime)}
+                    onPause={() => setIsPreviewPlaying(false)}
+                    onPlay={() => setIsPreviewPlaying(true)}
+                    onEnded={() => {
+                      setIsPreviewPlaying(false)
+                      setPreviewCurrentTime(0)
+                    }}
+                  />
+                  <div>
+                    <strong>语音已生成</strong>
+                    <span>
+                      {formatAudioTime(previewCurrentTime)} / {formatAudioTime(previewDuration)}
+                    </span>
+                    <div className="personal-wechat-preview-track">
+                      <span style={{ width: `${Math.min(100, previewProgress)}%` }} />
+                    </div>
+                  </div>
+                  <div>
+                    <Button
+                      variant="outline"
+                      size="sm"
+                      onClick={async () => {
+                        const audio = generatedAudioRef.current
+                        if (!audio) return
+                        if (audio.paused) {
+                          await audio.play()
+                          setIsPreviewPlaying(true)
+                        } else {
+                          audio.pause()
+                        }
+                      }}
+                      disabled={busy}
+                    >
+                      {isPreviewPlaying ? '暂停' : '试听'}
+                    </Button>
+                    <Button size="sm" onClick={() => void sendGenerated()} disabled={!canSend}>
+                      {busy ? '发送中…' : '发送'}
+                    </Button>
+                  </div>
+                </div>
+              ) : (
+                <Button
+                  size="sm"
+                  onClick={() => void generateVoice()}
+                  disabled={!generatedReady || isBusy}
+                >
+                  {isGenerating ? '正在生成…' : '生成语音'}
+                </Button>
+              )}
+            </>
+          ) : (
+            <div className="personal-wechat-file-editor">
+              <div>
+                <strong>{voice?.name || '还没有选择语音文件'}</strong>
+                <small>{voice ? voice.path : '支持 SILK、MP3、WAV、M4A、AAC、OGG 和 FLAC'}</small>
+              </div>
+              <Button
+                variant="outline"
+                size="sm"
+                onClick={() => void selectVoice()}
+                disabled={isBusy}
+              >
+                {voice ? '重新选择' : '选择语音'}
+              </Button>
+            </div>
+          )}
+        </div>
+      )}
+
+      <div className="personal-wechat-composer-footer">
+        {feedback && (
+          <span className={feedback.success ? 'is-success' : 'is-error'}>{feedback.message}</span>
+        )}
+        <Button variant="outline" size="sm" onClick={onCancel} disabled={isBusy}>
+          取消
+        </Button>
+        {!(mode === 'voice' && voiceSource === 'generated' && generatedVoice) && (
+          <Button size="sm" onClick={() => void sendCurrent()} disabled={!canSend}>
+            {busy ? '正在发送…' : '发送消息'}
+          </Button>
+        )}
+      </div>
+    </section>
+  )
+}
diff --git a/src/renderer/src/components/chat/PersonalWechatSendDialog.tsx b/src/renderer/src/components/chat/PersonalWechatSendDialog.tsx
index 4d4ef3e..a95ccf1 100644
--- a/src/renderer/src/components/chat/PersonalWechatSendDialog.tsx
+++ b/src/renderer/src/components/chat/PersonalWechatSendDialog.tsx
@@ -1,21 +1,21 @@
-import { useCallback, useEffect, useMemo, useRef, useState } from 'react'
+import { useCallback, useEffect, useRef, useState } from 'react'
 import type { Contact } from '../../../../shared/types'
-import type { PersonalWechatSenderStatus } from '../../../../shared/personal-wechat'
-import type { TextToSpeechSettings, TextToSpeechVoice } from '../../../../shared/text-to-speech'
+import type {
+  PersonalWechatSendRequest,
+  PersonalWechatSenderStatus
+} from '../../../../shared/personal-wechat'
+import type {
+  PersonalWechatRuntimeProgressEvent,
+  PersonalWechatRuntimeStatus
+} from '../../../../shared/personal-wechat-runtime'
+import { Button, Dialog, DialogContent, DialogDescription, DialogHeader, DialogTitle } from '../ui'
 import {
-  Button,
-  Dialog,
-  DialogContent,
-  DialogDescription,
-  DialogHeader,
-  DialogTitle,
-  SegmentedControl,
-  SegmentedControlItem,
-  Textarea
-} from '../ui'
+  PersonalWechatChatComposer,
+  type ChatMessage,
+  type PersonalWechatComposerMode
+} from './PersonalWechatChatComposer'
+import { PersonalWechatSetupGuide } from './PersonalWechatSetupGuide'
 
-type SendMode = 'image' | 'voice'
-type VoiceSource = 'generated' | 'file'
 type SelectedLocalFile = { path: string; name: string }
 
 interface PersonalWechatSendDialogProps {
@@ -23,48 +23,34 @@ interface PersonalWechatSendDialogProps {
   isGroupChat: boolean
   onClose: () => void
   onOpenTextToSpeechSettings?: () => void
-  initialMode?: SendMode
+  initialMode?: PersonalWechatComposerMode
   initialImage?: SelectedLocalFile | null
 }
 
-type GeneratedVoice = { filePath: string; audioDataUrl: string }
-const SHOW_SUPPORTED_WECHAT_VERSIONS_KEY = 'wxe:show-supported-wechat-versions'
-
-function formatAudioTime(value: number): string {
-  if (!Number.isFinite(value) || value < 0) return '0:00'
-  const seconds = Math.floor(value)
-  return `${Math.floor(seconds / 60)}:${String(seconds % 60).padStart(2, '0')}`
-}
-
-function statusLabel(status: PersonalWechatSenderStatus | null): string {
-  if (!status) return '正在检测'
-  if (status.state === 'online') return '个人微信已绑定'
-  if (status.state === 'stopped') return '尚未绑定'
-  if (status.state === 'runtime_missing') return '运行时未安装'
-  if (status.state === 'hook_not_ready') return '已绑定，等待消息初始化'
-  if (status.state === 'unsupported_version') return '微信版本不匹配'
-  if (status.state === 'wechat_not_running') return '微信未运行'
-  if (status.state === 'sip_enabled') return 'SIP 未关闭'
-  if (status.state === 'unsupported_platform') return '当前平台不支持'
-  if (status.state === 'starting') return '正在连接'
-  if (status.state === 'rebinding') return '正在重新绑定'
-  if (status.state === 'error') return '绑定异常'
-  return '正在检测'
-}
-
-function readyText(ready: boolean, readyLabel = '正常', waitingLabel = '等待初始化'): string {
-  return ready ? readyLabel : waitingLabel
-}
-
-function statusDescription(status: PersonalWechatSenderStatus | null): string {
-  if (!status) return '正在检查微信版本、进程和 OneBot Hook…'
-  if (status.state === 'unsupported_version') {
-    return `当前微信版本 ${status.wechatVersion || '未知'} 暂不支持。请在设置中查看支持的微信版本，安装完全一致的版本后再试。`
+function fallbackStatus(error: unknown): PersonalWechatSenderStatus {
+  return {
+    state: 'error',
+    platform: 'unknown',
+    arch: 'unknown',
+    sipDisabled: false,
+    wechatRunning: false,
+    runtimeReady: false,
+    endpoint: '127.0.0.1:58080',
+    endpointReady: false,
+    attachReady: false,
+    baseAddressReady: false,
+    textHookInstalled: false,
+    textHookReady: false,
+    imageHookInstalled: false,
+    imageHookReady: false,
+    messageListenerReady: false,
+    canSend: false,
+    canSendText: false,
+    canSendImage: false,
+    canSendVoice: false,
+    message: '无法检测个人微信发送服务',
+    error: error instanceof Error ? error.message : String(error)
   }
-  if (status.state === 'runtime_missing') {
-    return '微信发送组件尚未安装。请前往文字转语音设置下载组件后再试。'
-  }
-  return status.message
 }
 
 export function PersonalWechatSendDialog({
@@ -72,338 +58,137 @@ export function PersonalWechatSendDialog({
   isGroupChat,
   onClose,
   onOpenTextToSpeechSettings,
-  initialMode = 'image',
+  initialMode = 'text',
   initialImage = null
 }: PersonalWechatSendDialogProps): React.ReactElement {
-  const [status, setStatus] = useState<PersonalWechatSenderStatus | null>(null)
-  const [mode, setMode] = useState<SendMode>(initialMode)
-  const [image, setImage] = useState<SelectedLocalFile | null>(initialImage)
-  const [voice, setVoice] = useState<SelectedLocalFile | null>(null)
-  const [voiceSource, setVoiceSource] = useState<VoiceSource>('generated')
-  const [voiceText, setVoiceText] = useState('1')
-  const [ttsSettings, setTtsSettings] = useState<TextToSpeechSettings | null>(null)
-  const [ttsVoices, setTtsVoices] = useState<TextToSpeechVoice[]>([])
-  const [generatedVoice, setGeneratedVoice] = useState<GeneratedVoice | null>(null)
-  const [isGenerating, setIsGenerating] = useState(false)
-  const [isPreviewPlaying, setIsPreviewPlaying] = useState(false)
-  const [previewCurrentTime, setPreviewCurrentTime] = useState(0)
-  const [previewDuration, setPreviewDuration] = useState(0)
-  const [isSending, setIsSending] = useState(false)
-  const [isRebinding, setIsRebinding] = useState(false)
-  const [result, setResult] = useState<{ success: boolean; message: string } | null>(null)
-  const generatedVoiceRef = useRef<GeneratedVoice | null>(null)
-  const generatedAudioRef = useRef<HTMLAudioElement | null>(null)
+  const [senderStatus, setSenderStatus] = useState<PersonalWechatSenderStatus | null>(null)
+  const [runtimeStatus, setRuntimeStatus] = useState<PersonalWechatRuntimeStatus | null>(null)
+  const [runtimeProgress, setRuntimeProgress] = useState<PersonalWechatRuntimeProgressEvent | null>(
+    null
+  )
+  const [detecting, setDetecting] = useState(true)
+  const [binding, setBinding] = useState(false)
+  const [runtimeBusy, setRuntimeBusy] = useState(false)
+  const [sendBusy, setSendBusy] = useState(false)
+  const [detectionAttempted, setDetectionAttempted] = useState(false)
+  const [messages, setMessages] = useState<ChatMessage[]>([])
+  const [sendError, setSendError] = useState<string | null>(null)
+  const [composerStarted, setComposerStarted] = useState(Boolean(initialImage))
+  const requestIdRef = useRef(0)
   const restoreFocusRef = useRef<HTMLElement | null>(null)
   const closingRef = useRef(false)
   const displayName = contact.m_nsNickName || contact.m_nsUsrName || '未命名会话'
   const targetId = contact.m_nsUsrName
-  const selectedTypeReady = mode === 'voice' ? status?.canSendVoice : status?.canSendImage
-  const hasContent =
-    mode === 'voice'
-      ? voiceSource === 'file'
-        ? Boolean(voice?.path)
-        : Boolean(voiceText.trim())
-      : Boolean(image?.path)
-  const isBusy = isSending || isGenerating || isRebinding
-  const generatedVoiceReady = Boolean(
-    ttsSettings?.hasApiKey && ttsSettings.selectedVoiceId && voiceText.trim()
+  const isBusy = binding || runtimeBusy || sendBusy
+  const setupReady = Boolean(
+    senderStatus?.canSendText && (senderStatus?.canSendImage || senderStatus?.canSendVoice)
   )
-  const canSubmit = Boolean(
-    selectedTypeReady && hasContent && !isBusy && (mode !== 'voice' || voiceSource !== 'generated')
-  )
-  const canGenerate = Boolean(generatedVoiceReady && !isBusy)
-  const canSendGenerated = Boolean(status?.canSendVoice && generatedVoice && !isBusy)
-  const previewProgress = previewDuration > 0 ? (previewCurrentTime / previewDuration) * 100 : 0
-  const selectedTtsVoice = ttsVoices.find((item) => item.id === ttsSettings?.selectedVoiceId)
-  const statusTone = useMemo(() => {
-    if (status?.attachReady && status.baseAddressReady) return 'ready'
-    if (!status || status.state === 'checking' || status.state === 'starting') return 'checking'
-    return 'blocked'
-  }, [status])
-
-  const clearGeneratedVoice = useCallback((removeFile = true): void => {
-    const current = generatedVoiceRef.current
-    generatedVoiceRef.current = null
-    generatedAudioRef.current?.pause()
-    setGeneratedVoice(null)
-    setIsPreviewPlaying(false)
-    setPreviewCurrentTime(0)
-    setPreviewDuration(0)
-    if (removeFile && current?.filePath) {
-      void window.api.removeGeneratedTextToSpeechAudio(current.filePath).catch(() => undefined)
-    }
-  }, [])
-
-  const handleClose = useCallback((): void => {
-    if (isBusy || closingRef.current) return
-    closingRef.current = true
-    const restoreFocus = restoreFocusRef.current
-    restoreFocusRef.current = null
-    clearGeneratedVoice()
-    onClose()
-    queueMicrotask(() => restoreFocus?.focus())
-  }, [clearGeneratedVoice, isBusy, onClose])
-
-  const handleOpenTextToSpeechSettings = (showSupportedVersions = false): void => {
-    if (!onOpenTextToSpeechSettings || isBusy) return
-    if (showSupportedVersions) {
-      try {
-        sessionStorage.setItem(SHOW_SUPPORTED_WECHAT_VERSIONS_KEY, '1')
-      } catch {
-        // The settings page can still open if session storage is unavailable.
-      }
-    }
-    clearGeneratedVoice()
-    onClose()
-    onOpenTextToSpeechSettings()
-  }
 
   const refreshStatus = useCallback(async (): Promise<void> => {
-    setResult(null)
+    const requestId = ++requestIdRef.current
+    setDetecting(true)
+    setSendError(null)
     try {
-      setStatus(await window.api.getPersonalWechatSenderStatus())
+      const [nextRuntime, nextSender] = await Promise.all([
+        window.api.getPersonalWechatRuntimeStatus?.() || Promise.resolve(null),
+        window.api.getPersonalWechatSenderStatus()
+      ])
+      if (requestId !== requestIdRef.current) return
+      setRuntimeStatus(nextRuntime)
+      setRuntimeProgress(nextRuntime?.state === 'downloading' ? nextRuntime : null)
+      setSenderStatus(nextSender)
     } catch (error) {
-      setStatus({
-        state: 'error',
-        platform: 'unknown',
-        arch: 'unknown',
-        sipDisabled: false,
-        wechatRunning: false,
-        runtimeReady: false,
-        endpoint: '127.0.0.1:58080',
-        endpointReady: false,
-        attachReady: false,
-        baseAddressReady: false,
-        textHookInstalled: false,
-        textHookReady: false,
-        imageHookInstalled: false,
-        imageHookReady: false,
-        messageListenerReady: false,
-        canSend: false,
-        canSendText: false,
-        canSendImage: false,
-        canSendVoice: false,
-        message: '无法检测个人微信发送服务',
-        error: error instanceof Error ? error.message : String(error)
-      })
+      if (requestId === requestIdRef.current) setSenderStatus(fallbackStatus(error))
+    } finally {
+      if (requestId === requestIdRef.current) setDetecting(false)
     }
   }, [])
 
   useEffect(() => {
     void refreshStatus()
+    const unsubscribe = window.api.onPersonalWechatRuntimeProgress?.((status) => {
+      setRuntimeProgress(status)
+      setRuntimeStatus(status)
+      if (status.state === 'ready') void refreshStatus()
+    })
+    return unsubscribe
   }, [refreshStatus])
 
-  useEffect(() => {
-    let active = true
-    void window.api
-      .getTextToSpeechSettings()
-      .then(async (response) => {
-        if (!active) return
-        setTtsSettings(response.settings)
-        if (!response.settings.hasApiKey) return
-        const voicesResponse = await window.api.listTextToSpeechVoices({
-          pageNumber: 1,
-          pageSize: 24
-        })
-        if (!active || !voicesResponse.success) return
-        setTtsVoices(voicesResponse.items)
-      })
-      .catch(() => undefined)
-    return () => {
-      active = false
+  const handleDownloadRuntime = async (): Promise<void> => {
+    if (runtimeBusy) return
+    setRuntimeBusy(true)
+    setSendError(null)
+    try {
+      const result = await window.api.downloadPersonalWechatRuntime()
+      setRuntimeStatus(result.status)
+      if (!result.success && result.error) setSendError(result.error)
+      if (result.success) await refreshStatus()
+    } catch (error) {
+      setSendError(error instanceof Error ? error.message : String(error))
+    } finally {
+      setRuntimeBusy(false)
     }
-  }, [])
+  }
 
-  useEffect(() => {
-    return () => {
-      const current = generatedVoiceRef.current
-      generatedVoiceRef.current = null
-      if (current?.filePath) {
-        void window.api.removeGeneratedTextToSpeechAudio(current.filePath).catch(() => undefined)
-      }
-    }
-  }, [])
-
-  useEffect(() => {
-    const audio = generatedAudioRef.current
-    return () => audio?.pause()
-  }, [generatedVoice])
-
-  const handleRebind = async (): Promise<void> => {
-    if (isBusy) return
-    setIsRebinding(true)
-    setResult(null)
+  const handleBind = async (): Promise<void> => {
+    if (binding) return
+    setBinding(true)
+    setDetectionAttempted(false)
+    setSendError(null)
     try {
       const nextStatus = await window.api.rebindPersonalWechatSender()
-      setStatus(nextStatus)
-      setResult({
-        success: nextStatus.attachReady && nextStatus.baseAddressReady,
-        message: nextStatus.message
-      })
+      setSenderStatus(nextStatus)
+      if (nextStatus.state !== 'online' && nextStatus.message) setSendError(nextStatus.message)
     } catch (error) {
-      setResult({ success: false, message: error instanceof Error ? error.message : String(error) })
+      setSendError(error instanceof Error ? error.message : String(error))
     } finally {
-      setIsRebinding(false)
+      setBinding(false)
     }
   }
 
-  const handleSelectImage = async (): Promise<void> => {
-    if (isBusy) return
-    const selection = await window.api.selectPersonalWechatImage()
-    if (!selection.canceled && selection.path) {
-      setImage({ path: selection.path, name: selection.name || selection.path.split('/').pop()! })
-      setResult(null)
-    }
+  const handleDetect = async (): Promise<void> => {
+    setDetectionAttempted(true)
+    await refreshStatus()
   }
 
-  const handleSelectVoice = async (): Promise<void> => {
-    if (isBusy) return
-    const selection = await window.api.selectPersonalWechatVoice()
-    if (!selection.canceled && selection.path) {
-      setVoice({ path: selection.path, name: selection.name || selection.path.split('/').pop()! })
-      setResult(null)
-    }
+  const handleClose = (): void => {
+    if (isBusy || closingRef.current) return
+    closingRef.current = true
+    const restoreFocus = restoreFocusRef.current
+    restoreFocusRef.current = null
+    onClose()
+    queueMicrotask(() => restoreFocus?.focus())
   }
 
-  const changeMode = (nextMode: SendMode): void => {
-    if (mode === 'voice' && voiceSource === 'generated' && nextMode !== 'voice') {
-      clearGeneratedVoice()
-    }
-    setMode(nextMode)
-    setResult(null)
+  const handleOpenSettings = (): void => {
+    if (!onOpenTextToSpeechSettings || isBusy) return
+    handleClose()
+    onOpenTextToSpeechSettings()
   }
 
-  const changeVoiceSource = (nextSource: VoiceSource): void => {
-    if (voiceSource === 'generated' && nextSource !== 'generated') clearGeneratedVoice()
-    setVoiceSource(nextSource)
-    setResult(null)
-  }
-
-  const changeVoiceText = (nextText: string): void => {
-    if (generatedVoiceRef.current) clearGeneratedVoice()
-    setVoiceText(nextText)
-    setResult(null)
-  }
-
-  const handleGenerateVoice = async (): Promise<void> => {
-    if (!canGenerate) return
-    clearGeneratedVoice()
-    setIsGenerating(true)
-    setResult(null)
+  const handleSend = async (
+    request: PersonalWechatSendRequest
+  ): Promise<{ success: boolean; error?: string }> => {
+    setSendBusy(true)
+    setSendError(null)
     try {
-      const generated = await window.api.synthesizeTextToSpeech({
-        text: voiceText.trim(),
-        referenceId: ttsSettings!.selectedVoiceId
-      })
-      if (!generated.success || !generated.filePath || !generated.audioDataUrl) {
-        setResult({ success: false, message: generated.error || '语音生成失败' })
-        return
-      }
-      const nextVoice = {
-        filePath: generated.filePath,
-        audioDataUrl: generated.audioDataUrl
-      }
-      generatedVoiceRef.current = nextVoice
-      setGeneratedVoice(nextVoice)
-      setResult({ success: true, message: '语音已生成，可以先试听，确认后再发送' })
+      const response = await window.api.sendPersonalWechatMessage({ ...request, to: targetId })
+      setSenderStatus(response.status)
+      if (!response.success) setSendError(response.error || '发送失败')
+      return { success: response.success, error: response.error }
     } catch (error) {
-      setResult({ success: false, message: error instanceof Error ? error.message : String(error) })
+      const message = error instanceof Error ? error.message : String(error)
+      setSendError(message)
+      return { success: false, error: message }
     } finally {
-      setIsGenerating(false)
+      setSendBusy(false)
     }
   }
 
-  const handleToggleGeneratedPreview = async (): Promise<void> => {
-    const audio = generatedAudioRef.current
-    if (!audio) return
-    if (!audio.paused) {
-      audio.pause()
-      setIsPreviewPlaying(false)
-      return
-    }
-    try {
-      await audio.play()
-      setIsPreviewPlaying(true)
-    } catch {
-      setResult({ success: false, message: '语音试听播放失败' })
-    }
-  }
-
-  const handleSendGeneratedVoice = async (): Promise<void> => {
-    const current = generatedVoiceRef.current
-    if (!canSendGenerated || !current) return
-    setIsSending(true)
-    setResult(null)
-    generatedAudioRef.current?.pause()
-    setIsPreviewPlaying(false)
-    try {
-      const response = await window.api.sendPersonalWechatMessage({
-        type: 'voice',
-        to: targetId,
-        filePath: current.filePath,
-        isGroup: isGroupChat
-      })
-      setStatus(response.status)
-      setResult({
-        success: response.success,
-        message: response.success
-          ? '语音已提交给微信；请从另一设备或群成员处确认实际送达'
-          : response.error || '发送失败'
-      })
-      if (response.success) clearGeneratedVoice()
-    } catch (error) {
-      setResult({ success: false, message: error instanceof Error ? error.message : String(error) })
-    } finally {
-      setIsSending(false)
-    }
-  }
-
-  const handleSend = async (): Promise<void> => {
-    if (!canSubmit || (mode === 'voice' && voiceSource === 'generated')) return
-    setIsSending(true)
-    setResult(null)
-    try {
-      const response = await window.api.sendPersonalWechatMessage(
-        mode === 'voice'
-          ? { type: 'voice', to: targetId, filePath: voice!.path, isGroup: isGroupChat }
-          : { type: 'image', to: targetId, filePath: image!.path, isGroup: isGroupChat }
-      )
-      setStatus(response.status)
-      setResult({
-        success: response.success,
-        message: response.success
-          ? `${mode === 'voice' ? '语音' : '图片'}已提交给微信；请从另一设备或群成员处确认实际送达`
-          : response.error || '发送失败'
-      })
-    } catch (error) {
-      setResult({ success: false, message: error instanceof Error ? error.message : String(error) })
-    } finally {
-      setIsSending(false)
-    }
-  }
-
-  const statusItems = [
-    ['微信进程', status?.wechatPid ? `PID ${status.wechatPid}` : '未检测到'],
-    [
-      'OneBot',
-      status?.oneBotPid
-        ? `PID ${status.oneBotPid}${status.boundWechatPid ? ` · 绑定 ${status.boundWechatPid}` : ''}`
-        : '未启动'
-    ],
-    [
-      '接口',
-      `${status?.endpoint || '127.0.0.1:58080'} · ${readyText(Boolean(status?.endpointReady), '监听中', '未监听')}`
-    ],
-    ['基址扫描', status?.baseAddress || readyText(Boolean(status?.baseAddressReady))],
-    ['图片 Hook', readyText(Boolean(status?.imageHookReady), '已捕获，可发送', '等待手动发图片')],
-    ['语音能力', readyText(Boolean(status?.canSendVoice), '可发送', '等待媒体 Hook 初始化')],
-    ['消息监听', readyText(Boolean(status?.messageListenerReady), '正常', '等待收到微信消息')]
-  ]
-
   return (
     <Dialog open onOpenChange={(open) => !open && handleClose()}>
       <DialogContent
-        className="personal-wechat-send-dialog max-h-[calc(100vh-3rem)] max-w-[620px] gap-4 overflow-y-auto p-[22px]"
+        className="personal-wechat-send-dialog max-h-[calc(100vh-2rem)] max-w-[720px] gap-0 overflow-y-auto p-0"
         onOpenAutoFocus={() => {
           restoreFocusRef.current =
             document.activeElement instanceof HTMLElement ? document.activeElement : null
@@ -411,291 +196,91 @@ export function PersonalWechatSendDialog({
         onEscapeKeyDown={(event) => isBusy && event.preventDefault()}
         onPointerDownOutside={(event) => isBusy && event.preventDefault()}
       >
-        <DialogHeader className="flex-row items-center justify-between space-y-0 pr-10">
-          <div>
-            <span className="text-[11px] font-bold tracking-normal text-primary">实验性功能</span>
-            <DialogTitle className="mt-0.5 text-[19px] leading-[26px] tracking-normal">
-              个人微信测试发送
-            </DialogTitle>
+        <DialogHeader className="personal-wechat-chat-header">
+          <div className="personal-wechat-chat-avatar" aria-hidden>
+            {displayName.slice(0, 1)}
           </div>
-          <DialogDescription className="sr-only">
-            向当前微信联系人或群聊测试发送图片和语音。
-          </DialogDescription>
+          <div className="personal-wechat-chat-heading">
+            <DialogTitle>{displayName}</DialogTitle>
+            <DialogDescription>
+              {isGroupChat ? '群聊' : '联系人'} · {setupReady ? '微信已连接' : '配置微信消息发送'}
+            </DialogDescription>
+          </div>
+          <span
+            className={`personal-wechat-connection-dot ${setupReady ? 'is-online' : ''}`}
+            aria-label={setupReady ? '微信已连接' : '微信尚未配置'}
+          />
         </DialogHeader>
 
-        <div className="personal-wechat-send-device-note" role="note">
-          <span aria-hidden>i</span>
-          <p>
-            <strong>请在其他设备确认发送结果</strong>
-            通过测试功能发送的消息不会显示在本机微信会话中，请切换到手机、平板等其他设备确认是否送达。
-          </p>
-        </div>
-
-        <div className="personal-wechat-send-target">
-          <span>{isGroupChat ? '发送到群聊' : '发送给联系人'}</span>
-          <strong>{displayName}</strong>
-          <code>{targetId}</code>
-        </div>
-
-        <div className={`personal-wechat-send-status ${statusTone}`}>
-          <span className="personal-wechat-send-status-dot" />
-          <div>
-            <strong>{statusLabel(status)}</strong>
-            <p>{statusDescription(status)}</p>
-            {status?.wechatVersion && <small>微信版本：{status.wechatVersion}</small>}
-            {status?.error &&
-              status.state !== 'unsupported_version' &&
-              status.state !== 'runtime_missing' && <small className="error">{status.error}</small>}
-          </div>
-          <div className="personal-wechat-send-status-actions">
-            {(status?.state === 'unsupported_version' || status?.state === 'runtime_missing') &&
-            onOpenTextToSpeechSettings ? (
-              <Button
-                variant="link"
-                size="sm"
-                className="h-auto p-0"
-                onClick={() =>
-                  handleOpenTextToSpeechSettings(status.state === 'unsupported_version')
-                }
-                disabled={isBusy}
-              >
-                {status.state === 'runtime_missing' ? '前往下载组件' : '查看支持版本'}
-              </Button>
-            ) : null}
-            <Button
-              variant="link"
-              size="sm"
-              className="h-auto p-0"
-              onClick={() => void refreshStatus()}
-              disabled={isBusy}
-            >
-              重新检测
-            </Button>
-            <Button
-              variant="link"
-              size="sm"
-              className="h-auto p-0"
-              onClick={() => void handleRebind()}
-              disabled={isBusy}
-            >
-              {isRebinding ? '绑定中…' : '尝试重新绑定'}
-            </Button>
-          </div>
-        </div>
-
-        <dl className="personal-wechat-send-diagnostics">
-          {statusItems.map(([label, value]) => (
-            <div key={label}>
-              <dt>{label}</dt>
-              <dd>{value}</dd>
+        <div className="personal-wechat-chat-body">
+          {setupReady && composerStarted && (
+            <div className="personal-wechat-message-list" aria-label="消息列表">
+              {messages.length === 0 ? (
+                <div className="personal-wechat-empty-message">还没有发送消息。</div>
+              ) : (
+                messages.map((message) => (
+                  <div
+                    key={message.id}
+                    className={`personal-wechat-message-bubble ${message.outgoing ? 'is-outgoing' : ''}`}
+                  >
+                    <span className="personal-wechat-message-kind">
+                      {message.type === 'text'
+                        ? '文字'
+                        : message.type === 'image'
+                          ? '图片'
+                          : '语音'}
+                    </span>
+                    <span>{message.text || message.fileName}</span>
+                  </div>
+                ))
+              )}
             </div>
-          ))}
-        </dl>
-
-        <SegmentedControl
-          className="personal-wechat-send-mode grid w-full grid-cols-2"
-          aria-label="测试消息类型"
-          value={mode}
-          onValueChange={(value) => changeMode(value as SendMode)}
-          disabled={isBusy}
-        >
-          <SegmentedControlItem value="image" className="w-full">
-            图片
-          </SegmentedControlItem>
-          <SegmentedControlItem value="voice" className="w-full">
-            语音
-          </SegmentedControlItem>
-        </SegmentedControl>
-
-        {mode === 'image' ? (
-          <div className="personal-wechat-send-image-picker">
-            <span>图片</span>
-            <Button
-              variant="outline"
-              size="sm"
-              onClick={() => void handleSelectImage()}
-              disabled={isBusy}
-            >
-              {image ? '重新选择图片' : '选择图片'}
-            </Button>
-            {image ? (
-              <div>
-                <strong>{image.name}</strong>
-                <small>{image.path}</small>
-              </div>
-            ) : (
-              <small>支持 PNG、JPG、GIF 和 WebP，最大 20 MB</small>
-            )}
-          </div>
-        ) : (
-          <div className="personal-wechat-voice-composer">
-            <SegmentedControl
-              className="personal-wechat-voice-source grid w-full grid-cols-2"
-              aria-label="语音来源"
-              value={voiceSource}
-              onValueChange={(value) => changeVoiceSource(value as VoiceSource)}
-              disabled={isBusy}
-            >
-              <SegmentedControlItem value="generated" className="w-full">
-                输入文字生成
-              </SegmentedControlItem>
-              <SegmentedControlItem value="file" className="w-full">
-                选择本地文件
-              </SegmentedControlItem>
-            </SegmentedControl>
-
-            {voiceSource === 'generated' ? (
-              <div className="personal-wechat-generated-voice">
-                <div className="personal-wechat-generated-voice-heading">
-                  <div>
-                    <span>文字生成语音</span>
-                    <strong>{selectedTtsVoice?.name || '尚未选择音色'}</strong>
-                  </div>
-                  {onOpenTextToSpeechSettings ? (
-                    <Button
-                      variant="outline"
-                      size="sm"
-                      onClick={() => handleOpenTextToSpeechSettings()}
-                    >
-                      前往文字转语音设置
-                    </Button>
-                  ) : null}
-                </div>
-                <label className="personal-wechat-send-editor">
-                  <span>要生成的文字</span>
-                  <Textarea
-                    aria-label="要生成的文字"
-                    value={voiceText}
-                    maxLength={1000}
-                    rows={3}
-                    disabled={isBusy}
-                    onChange={(event) => changeVoiceText(event.target.value)}
-                  />
-                  <small>{voiceText.length} / 1000</small>
-                </label>
-                <div
-                  className={`personal-wechat-tts-readiness ${ttsSettings?.hasApiKey ? 'ready' : ''}`}
-                >
-                  <strong>
-                    {generatedVoiceReady
-                      ? 'API Key 与音色已准备'
-                      : ttsSettings?.hasApiKey
-                        ? '还需要在设置中选择音色'
-                        : '还需要配置语音服务 API Key'}
-                  </strong>
-                  <span>点击右下角“生成语音”，生成完成后可以试听，再决定是否发送。</span>
-                </div>
-                {isGenerating ? (
-                  <div className="personal-wechat-generation-progress" aria-live="polite">
-                    <div>
-                      <strong>正在生成语音…</strong>
-                      <span>通常需要几秒，请稍候</span>
-                    </div>
-                    <div className="personal-wechat-generation-track">
-                      <span />
-                    </div>
-                  </div>
-                ) : generatedVoice ? (
-                  <div className="personal-wechat-generated-result">
-                    <audio
-                      ref={generatedAudioRef}
-                      src={generatedVoice.audioDataUrl}
-                      preload="metadata"
-                      onLoadedMetadata={(event) => setPreviewDuration(event.currentTarget.duration)}
-                      onTimeUpdate={(event) =>
-                        setPreviewCurrentTime(event.currentTarget.currentTime)
-                      }
-                      onPause={() => setIsPreviewPlaying(false)}
-                      onPlay={() => setIsPreviewPlaying(true)}
-                      onEnded={() => {
-                        setIsPreviewPlaying(false)
-                        setPreviewCurrentTime(0)
-                      }}
-                    />
-                    <div className="personal-wechat-generated-result-copy">
-                      <strong>语音生成完成</strong>
-                      <span>
-                        {formatAudioTime(previewCurrentTime)} / {formatAudioTime(previewDuration)}
-                      </span>
-                      <div className="personal-wechat-preview-track">
-                        <span style={{ width: `${Math.min(100, previewProgress)}%` }} />
-                      </div>
-                    </div>
-                    <div className="personal-wechat-generated-result-actions">
-                      <Button
-                        variant="outline"
-                        size="sm"
-                        onClick={() => void handleToggleGeneratedPreview()}
-                        disabled={isSending}
-                      >
-                        {isPreviewPlaying ? '暂停' : '播放'}
-                      </Button>
-                      <Button
-                        size="sm"
-                        onClick={() => void handleSendGeneratedVoice()}
-                        disabled={!canSendGenerated}
-                      >
-                        {isSending ? '正在发送…' : `发送到${isGroupChat ? '群聊' : '联系人'}`}
-                      </Button>
-                    </div>
-                  </div>
-                ) : null}
-              </div>
-            ) : (
-              <div className="personal-wechat-send-image-picker">
-                <span>本地语音文件</span>
-                <Button
-                  variant="outline"
-                  size="sm"
-                  onClick={() => void handleSelectVoice()}
-                  disabled={isBusy}
-                >
-                  {voice ? '重新选择语音' : '选择语音'}
-                </Button>
-                {voice ? (
-                  <div>
-                    <strong>{voice.name}</strong>
-                    <small>{voice.path}</small>
-                  </div>
-                ) : (
-                  <small>支持 SILK、MP3、WAV、M4A、AAC、OGG 和 FLAC，最大 20 MB</small>
-                )}
-              </div>
-            )}
-          </div>
-        )}
-
-        <p className="personal-wechat-send-note">
-          {mode === 'voice' && voiceSource === 'generated'
-            ? '文字生成语音会复用现有媒体上传能力；请先在微信中给任意好友手动发送一张普通图片，再点击重新检测。'
-            : mode === 'voice'
-              ? '本地语音复用媒体上传 Hook，请先在微信中给任意好友手动发送一张普通图片，再点击重新检测。'
-              : '如果想测试图片，请先在微信中给任意好友手动发送一张普通图片，再点击重新检测。'}
-          微信重新登录导致 PID 改变时，请点击“尝试重新绑定”。
-        </p>
-
-        {result && (
-          <div className={`personal-wechat-send-result ${result.success ? 'success' : 'error'}`}>
-            {result.message}
-          </div>
-        )}
-
-        <footer>
-          <Button variant="outline" onClick={handleClose} disabled={isBusy}>
-            取消
-          </Button>
-          {mode === 'voice' && voiceSource === 'generated' ? (
-            <Button disabled={!canGenerate} onClick={() => void handleGenerateVoice()}>
-              {isGenerating ? '正在生成语音…' : generatedVoice ? '重新生成语音' : '生成语音'}
-            </Button>
-          ) : (
-            <Button disabled={!canSubmit} onClick={handleSend}>
-              {isSending
-                ? '正在发送…'
-                : `测试发送${mode === 'voice' ? '语音' : '图片'}到${isGroupChat ? '群聊' : '联系人'}`}
-            </Button>
           )}
-        </footer>
+
+          {(!setupReady || !composerStarted) && senderStatus && (
+            <PersonalWechatSetupGuide
+              runtimeStatus={runtimeStatus}
+              senderStatus={senderStatus}
+              runtimeProgress={runtimeProgress}
+              runtimeBusy={runtimeBusy}
+              binding={binding}
+              detecting={detecting}
+              onDownloadRuntime={() => void handleDownloadRuntime()}
+              onBind={() => void handleBind()}
+              detectionAttempted={detectionAttempted}
+              onDetect={() => void handleDetect()}
+              onStartSending={() => setComposerStarted(true)}
+              onOpenTextToSpeechSettings={handleOpenSettings}
+            />
+          )}
+
+          {setupReady && composerStarted && (
+            <PersonalWechatChatComposer
+              status={senderStatus!}
+              targetId={targetId}
+              isGroupChat={isGroupChat}
+              initialMode={initialMode}
+              initialImage={initialImage}
+              onOpenTextToSpeechSettings={handleOpenSettings}
+              onCancel={handleClose}
+              onSend={handleSend}
+              onMessage={(message) => setMessages((current) => [...current, message])}
+              busy={sendBusy}
+            />
+          )}
+          {sendError && (
+            <div className="personal-wechat-global-error" role="alert">
+              {sendError}
+            </div>
+          )}
+        </div>
+        {(!setupReady || !composerStarted) && (
+          <div className="personal-wechat-chat-footer">
+            <Button variant="outline" onClick={handleClose} disabled={isBusy}>
+              关闭
+            </Button>
+          </div>
+        )}
       </DialogContent>
     </Dialog>
   )
diff --git a/src/renderer/src/components/chat/PersonalWechatSetupGuide.tsx b/src/renderer/src/components/chat/PersonalWechatSetupGuide.tsx
new file mode 100644
index 0000000..9f4fc46
--- /dev/null
+++ b/src/renderer/src/components/chat/PersonalWechatSetupGuide.tsx
@@ -0,0 +1,306 @@
+import { useState } from 'react'
+import type { PersonalWechatSenderStatus } from '../../../../shared/personal-wechat'
+import type {
+  PersonalWechatRuntimeProgressEvent,
+  PersonalWechatRuntimeStatus
+} from '../../../../shared/personal-wechat-runtime'
+import { Button, Dialog, DialogContent, DialogDescription, DialogHeader, DialogTitle } from '../ui'
+import { PersonalWechatSupportedVersionsContent } from './PersonalWechatSupportedVersionsContent'
+
+interface PersonalWechatSetupGuideProps {
+  runtimeStatus: PersonalWechatRuntimeStatus | null
+  senderStatus: PersonalWechatSenderStatus | null
+  runtimeProgress: PersonalWechatRuntimeProgressEvent | null
+  runtimeBusy: boolean
+  binding: boolean
+  detecting: boolean
+  detectionAttempted: boolean
+  onDownloadRuntime: () => void
+  onBind: () => void
+  onDetect: () => void
+  onStartSending: () => void
+  onOpenTextToSpeechSettings?: () => void
+}
+
+function capabilityLabel(ready: boolean, detectionAttempted: boolean): string {
+  if (ready) return '已就绪'
+  return detectionAttempted ? '未检测' : '待检测'
+}
+
+function diagnosticValue(value: unknown): string {
+  if (value === undefined || value === null || value === '') return '未检测到'
+  return String(value)
+}
+
+function bindingHint(status: PersonalWechatSenderStatus): string {
+  if (status.state === 'wechat_not_running') return '请先启动并登录微信。'
+  if (status.state === 'unsupported_platform') return '当前系统暂不支持个人微信发送。'
+  if (status.state === 'unsupported_version')
+    return '当前微信版本暂不支持，请查看语音设置中的支持版本。'
+  if (status.state === 'runtime_missing') return '请先完成语音模型准备。'
+  if (status.state === 'error') return '连接微信时遇到问题，请稍后重试。'
+  return '请启动并登录当前微信，TraceMemo 会自动绑定正在使用的账号。'
+}
+
+function isWechatBound(status: PersonalWechatSenderStatus): boolean {
+  if (status.state === 'online') return true
+  return Boolean(
+    status.state === 'hook_not_ready' &&
+    status.wechatPid &&
+    status.boundWechatPid === status.wechatPid &&
+    status.attachReady &&
+    status.baseAddressReady
+  )
+}
+
+export function PersonalWechatSetupGuide({
+  runtimeStatus,
+  senderStatus,
+  runtimeProgress,
+  runtimeBusy,
+  binding,
+  detecting,
+  detectionAttempted,
+  onDownloadRuntime,
+  onBind,
+  onDetect,
+  onStartSending,
+  onOpenTextToSpeechSettings
+}: PersonalWechatSetupGuideProps): React.ReactElement {
+  const [showSupportedVersions, setShowSupportedVersions] = useState(false)
+  const runtimeReady = runtimeStatus?.state === 'ready' || senderStatus?.runtimeReady === true
+  const runtimeDownloading = runtimeBusy || runtimeStatus?.state === 'downloading'
+  const connected = senderStatus ? isWechatBound(senderStatus) : false
+  const canSendText = Boolean(senderStatus?.canSendText)
+  const canSendImage = Boolean(senderStatus?.canSendImage)
+  const canSendVoice = Boolean(senderStatus?.canSendVoice)
+  const mediaReady = canSendImage || canSendVoice
+  const verificationComplete = canSendText && mediaReady
+  const allReady = verificationComplete
+  const progress = runtimeProgress || runtimeStatus
+  const progressPercent = Math.max(0, Math.min(100, Math.round((progress?.progress || 0) * 100)))
+
+  return (
+    <section className="personal-wechat-setup" aria-label="微信消息功能配置">
+      <div className="personal-wechat-setup-heading">
+        <div>
+          <span className="personal-wechat-eyebrow">首次使用</span>
+          <h2>先准备好微信消息功能</h2>
+          <p>只需几步，TraceMemo 就能像普通聊天窗口一样发送文字、图片和语音。</p>
+        </div>
+        {allReady && <span className="personal-wechat-setup-complete">已完成</span>}
+      </div>
+
+      <ol className="personal-wechat-steps">
+        <li
+          className={runtimeReady ? 'is-complete' : runtimeDownloading ? 'is-active' : 'is-current'}
+        >
+          <span className="personal-wechat-step-number">{runtimeReady ? '✓' : '1'}</span>
+          <div className="personal-wechat-step-content">
+            <strong>准备语音模型</strong>
+            <p>发送语音消息需要本地模型，首次使用时下载一次即可。</p>
+            {runtimeDownloading ? (
+              <div className="personal-wechat-download-progress" aria-live="polite">
+                <div>
+                  <span>正在准备模型</span>
+                  <span>{progressPercent}%</span>
+                </div>
+                <div className="personal-wechat-progress-track">
+                  <span style={{ width: `${progressPercent}%` }} />
+                </div>
+              </div>
+            ) : runtimeReady ? (
+              <span className="personal-wechat-step-status">✓ 语音模型已准备</span>
+            ) : (
+              <Button
+                size="sm"
+                onClick={onDownloadRuntime}
+                disabled={runtimeBusy || runtimeStatus?.state === 'unsupported'}
+              >
+                下载模型
+              </Button>
+            )}
+            {runtimeStatus?.error && !runtimeDownloading && !runtimeReady && (
+              <p className="personal-wechat-step-error">模型准备失败，请重试或查看语音设置。</p>
+            )}
+          </div>
+        </li>
+
+        <li
+          className={
+            connected
+              ? 'is-complete'
+              : binding
+                ? 'is-active'
+                : runtimeReady
+                  ? 'is-current'
+                  : 'is-pending'
+          }
+        >
+          <span className="personal-wechat-step-number">{connected ? '✓' : '2'}</span>
+          <div className="personal-wechat-step-content">
+            <strong>绑定个人微信</strong>
+            <p>请启动并登录当前微信，TraceMemo 会自动绑定正在使用的账号。</p>
+            {connected ? (
+              <span className="personal-wechat-step-status">✓ 微信已绑定</span>
+            ) : (
+              <Button
+                size="sm"
+                variant="outline"
+                onClick={onBind}
+                disabled={!runtimeReady || binding || runtimeDownloading}
+              >
+                {binding ? '正在绑定…' : '绑定微信'}
+              </Button>
+            )}
+            {!connected && senderStatus && (
+              <p className="personal-wechat-step-hint">{bindingHint(senderStatus)}</p>
+            )}
+            {!connected && (
+              <p className="personal-wechat-step-warning" role="note">
+                绑定微信可能导致当前微信异常闪退，这是正常现象。若微信退出，请重新启动微信后，再回到这里重新检测/绑定。
+                <br />
+                微信总是自动更新？请在微信左下角打开“设置 → 通用”，取消勾选“有更新时自动升级微信”，否则版本变化后可能无法绑定。
+              </p>
+            )}
+            <Button
+              className="w-fit"
+              variant="link"
+              size="sm"
+              onClick={() => setShowSupportedVersions(true)}
+            >
+              查看支持的微信版本
+            </Button>
+          </div>
+        </li>
+
+        <li
+          className={verificationComplete ? 'is-complete' : connected ? 'is-current' : 'is-pending'}
+        >
+          <span className="personal-wechat-step-number">{verificationComplete ? '✓' : '3'}</span>
+          <div className="personal-wechat-step-content">
+            <strong>验证消息能力</strong>
+            <p>
+              请打开手机微信，给任意好友手动发送：
+              <br />① 一条文字消息
+              <br />② 一张图片
+            </p>
+            <p className="personal-wechat-step-hint">
+              TraceMemo 需要通过你主动发送的消息初始化微信消息能力。完成后点击“重新检测”。
+            </p>
+            {!verificationComplete && (
+              <Button
+                size="sm"
+                variant="outline"
+                onClick={onDetect}
+                disabled={detecting || binding}
+              >
+                {detecting ? '正在检测…' : '重新检测'}
+              </Button>
+            )}
+            {detectionAttempted && !detecting && !verificationComplete && (
+              <p className="personal-wechat-step-hint" role="status">
+                暂未检测到新的消息，请确认已在手机微信中发送文字和图片，然后再次检测。
+              </p>
+            )}
+          </div>
+        </li>
+
+        <li
+          className={allReady ? 'is-complete' : verificationComplete ? 'is-current' : 'is-pending'}
+        >
+          <span className="personal-wechat-step-number">{allReady ? '✓' : '4'}</span>
+          <div className="personal-wechat-step-content">
+            <strong>能力检测</strong>
+            <div className="personal-wechat-capabilities" aria-label="微信消息能力">
+              {[
+                ['文字消息', canSendText],
+                ['图片和语音消息', mediaReady]
+              ].map(([label, ready]) => (
+                <span key={String(label)} className={ready ? 'is-ready' : ''}>
+                  <b aria-hidden>{ready ? '✓' : '−'}</b>
+                  {label}
+                  <small>{capabilityLabel(Boolean(ready), detectionAttempted)}</small>
+                </span>
+              ))}
+            </div>
+            {allReady ? (
+              <>
+                <span className="personal-wechat-step-status">微信消息发送已配置完成</span>
+                <Button size="sm" onClick={onStartSending}>
+                  开始发送
+                </Button>
+              </>
+            ) : verificationComplete ? (
+              <>
+                <p className="personal-wechat-step-hint">
+                  语音能力还未就绪，请在微信中完成一次媒体消息初始化后重新检测。
+                </p>
+                <Button size="sm" variant="outline" onClick={onDetect} disabled={detecting}>
+                  {detecting ? '正在检测…' : '重新检测能力'}
+                </Button>
+              </>
+            ) : null}
+          </div>
+        </li>
+      </ol>
+
+      {!runtimeReady && runtimeStatus?.state === 'unsupported' && onOpenTextToSpeechSettings && (
+        <Button
+          variant="link"
+          size="sm"
+          className="personal-wechat-setup-link"
+          onClick={onOpenTextToSpeechSettings}
+        >
+          查看语音设置
+        </Button>
+      )}
+
+      <Dialog open={showSupportedVersions} onOpenChange={setShowSupportedVersions}>
+        <DialogContent className="max-h-[calc(100vh-3rem)] max-w-[620px] overflow-y-auto">
+          <DialogHeader className="pr-8">
+            <DialogTitle className="text-lg">支持的微信版本</DialogTitle>
+            <DialogDescription>请安装下列完整版本之一。</DialogDescription>
+          </DialogHeader>
+          <PersonalWechatSupportedVersionsContent />
+        </DialogContent>
+      </Dialog>
+
+      <details className="personal-wechat-diagnostics">
+        <summary>高级诊断</summary>
+        <p>仅用于排查连接问题，普通使用无需关注这些信息。</p>
+        <dl>
+          {[
+            ['微信进程', senderStatus?.wechatPid ? `PID ${senderStatus.wechatPid}` : '未检测到'],
+            ['OneBot', senderStatus?.oneBotPid ? `PID ${senderStatus.oneBotPid}` : '未启动'],
+            ['绑定进程', diagnosticValue(senderStatus?.boundWechatPid)],
+            [
+              '接口监听',
+              `${diagnosticValue(senderStatus?.endpoint)} · ${senderStatus?.endpointReady ? '监听中' : '未监听'}`
+            ],
+            [
+              '基址扫描',
+              senderStatus?.baseAddress || (senderStatus?.baseAddressReady ? '已完成' : '未完成')
+            ],
+            ['文字 Hook', senderStatus?.textHookReady ? '已就绪' : '未就绪'],
+            ['图片 Hook', senderStatus?.imageHookReady ? '已就绪' : '未就绪'],
+            ['语音能力', senderStatus?.canSendVoice ? '可发送' : '未就绪'],
+            ['消息监听', senderStatus?.messageListenerReady ? '正常' : '未就绪'],
+            ['微信版本', diagnosticValue(senderStatus?.wechatVersion)],
+            [
+              '运行时',
+              runtimeStatus?.version
+                ? `${runtimeStatus.version} · ${runtimeStatus.state}`
+                : diagnosticValue(senderStatus?.runtimeReady)
+            ]
+          ].map(([label, value]) => (
+            <div key={label}>
+              <dt>{label}</dt>
+              <dd>{value}</dd>
+            </div>
+          ))}
+        </dl>
+      </details>
+    </section>
+  )
+}
diff --git a/src/renderer/src/components/chat/PersonalWechatSupportedVersionsContent.tsx b/src/renderer/src/components/chat/PersonalWechatSupportedVersionsContent.tsx
new file mode 100644
index 0000000..345599c
--- /dev/null
+++ b/src/renderer/src/components/chat/PersonalWechatSupportedVersionsContent.tsx
@@ -0,0 +1,46 @@
+import type React from 'react'
+
+export const WECHAT_VERSION_DOWNLOAD_URL = 'https://github.com/zsbai/wechat-versions/releases'
+
+export const BUNDLED_WECHAT_VERSIONS = [
+  '4.1.6.12',
+  '4.1.6.46',
+  '4.1.6.47',
+  '4.1.7.31',
+  '4.1.7.55',
+  '4.1.7.57',
+  '4.1.8.28',
+  '4.1.8.29',
+  '4.1.8.104',
+  '4.1.8.107',
+  '4.1.9.52',
+  '4.1.9.55',
+  '4.1.9.58',
+  '4.1.10.53',
+  '4.1.11.53'
+] as const
+
+export function PersonalWechatSupportedVersionsContent(): React.ReactElement {
+  return (
+    <>
+      <a
+        className="w-fit rounded-md border border-primary/30 bg-primary/10 px-3 py-2 text-xs font-semibold text-primary hover:border-primary"
+        href={WECHAT_VERSION_DOWNLOAD_URL}
+        target="_blank"
+        rel="noreferrer"
+      >
+        下载微信历史版本 ↗
+      </a>
+      <div className="grid grid-cols-2 gap-2 sm:grid-cols-3">
+        {BUNDLED_WECHAT_VERSIONS.map((version) => (
+          <span
+            className="rounded-md border border-border bg-background px-2 py-2 text-center text-xs text-muted-foreground"
+            key={version}
+          >
+            {version}
+          </span>
+        ))}
+      </div>
+    </>
+  )
+}
diff --git a/src/renderer/src/features/settings/pages/TextToSpeechPage.tsx b/src/renderer/src/features/settings/pages/TextToSpeechPage.tsx
index fdb8519..ddee9da 100644
--- a/src/renderer/src/features/settings/pages/TextToSpeechPage.tsx
+++ b/src/renderer/src/features/settings/pages/TextToSpeechPage.tsx
@@ -19,28 +19,11 @@ import {
   SelectTrigger,
   SelectValue
 } from '../../../components/ui'
+import { PersonalWechatSupportedVersionsContent } from '../../../components/chat/PersonalWechatSupportedVersionsContent'
 import { isMac, isWindows } from '../../../utils/runtime-environment'
 
 const VOICE_PAGE_SIZE = 24
-const WECHAT_VERSION_DOWNLOAD_URL = 'https://github.com/zsbai/wechat-versions/releases'
 const SHOW_SUPPORTED_WECHAT_VERSIONS_KEY = 'wxe:show-supported-wechat-versions'
-const BUNDLED_WECHAT_VERSIONS = [
-  '4.1.6.12',
-  '4.1.6.46',
-  '4.1.6.47',
-  '4.1.7.31',
-  '4.1.7.55',
-  '4.1.7.57',
-  '4.1.8.28',
-  '4.1.8.29',
-  '4.1.8.104',
-  '4.1.8.107',
-  '4.1.9.52',
-  '4.1.9.55',
-  '4.1.9.58',
-  '4.1.10.53',
-  '4.1.11.53'
-] as const
 
 const RUNTIME_STATUS_LABELS: Record<PersonalWechatRuntimeStatus['state'], string> = {
   missing: '未下载',
@@ -539,7 +522,7 @@ export function TextToSpeechPage({
                   <span>2</span>
                   <div>
                     <strong>在档案中生成语音</strong>
-                    <p>打开联系人或群聊的测试发送，选择“语音 → 输入文字生成”。</p>
+                    <p>打开联系人或群聊的发送，选择“语音 → 输入文字生成”。</p>
                   </div>
                 </li>
                 <li>
@@ -566,7 +549,11 @@ export function TextToSpeechPage({
                       发送功能需要关闭 SIP 并连接微信进程。关闭 SIP
                       会降低系统安全性，请确认风险后再使用。
                     </li>
-                    <li>微信重新登录或 PID 改变后，需要在测试发送弹窗中点击“尝试重新绑定”。</li>
+                    <li>微信重新登录或 PID 改变后，需要在发送窗口中重新检测或绑定。</li>
+                    <li>
+                      绑定前建议关闭微信自动升级 在微信左下角打开“设置 →
+                      通用”，取消勾选“有更新时自动升级微信”。微信自动更新后，版本可能不再兼容发送组件。
+                    </li>
                   </ul>
                   {personalWechatRuntimeSupported ? (
                     <Button
@@ -941,24 +928,7 @@ export function TextToSpeechPage({
           <DialogTitle className="text-lg">支持的微信版本</DialogTitle>
           <DialogDescription>请安装下列完整版本之一。</DialogDescription>
         </DialogHeader>
-        <a
-          className="w-fit rounded-md border border-primary/30 bg-primary/10 px-3 py-2 text-xs font-semibold text-primary hover:border-primary"
-          href={WECHAT_VERSION_DOWNLOAD_URL}
-          target="_blank"
-          rel="noreferrer"
-        >
-          下载微信历史版本 ↗
-        </a>
-        <div className="grid grid-cols-2 gap-2 sm:grid-cols-3">
-          {BUNDLED_WECHAT_VERSIONS.map((version) => (
-            <span
-              className="rounded-md border border-border bg-background px-2 py-2 text-center text-xs text-muted-foreground"
-              key={version}
-            >
-              {version}
-            </span>
-          ))}
-        </div>
+        <PersonalWechatSupportedVersionsContent />
       </DialogContent>
     </Dialog>
   )
diff --git a/src/renderer/src/styles/chat.scss b/src/renderer/src/styles/chat.scss
index 6911f81..1896fc3 100644
--- a/src/renderer/src/styles/chat.scss
+++ b/src/renderer/src/styles/chat.scss
@@ -564,6 +564,515 @@
   }
 }
 
+.personal-wechat-chat-header {
+  display: flex;
+  align-items: center;
+  gap: 11px;
+  min-height: 68px;
+  padding: 14px 20px;
+  border-bottom: 1px solid var(--wxex-border);
+  background: var(--wxex-bg-elevated);
+}
+
+.personal-wechat-chat-avatar {
+  display: grid;
+  width: 34px;
+  height: 34px;
+  flex: 0 0 auto;
+  place-items: center;
+  border-radius: 9px;
+  background: var(--wxex-brand-soft);
+  color: var(--wxex-brand);
+  font-size: 14px;
+  font-weight: 700;
+}
+
+.personal-wechat-chat-heading {
+  min-width: 0;
+  flex: 1;
+
+  h2,
+  p {
+    margin: 0;
+  }
+
+  h2 {
+    overflow: hidden;
+    color: var(--wxex-text-primary);
+    font-size: 14px;
+    font-weight: 650;
+    text-overflow: ellipsis;
+    white-space: nowrap;
+  }
+
+  p {
+    margin-top: 2px;
+    color: var(--wxex-text-muted);
+    font-size: 11px;
+  }
+}
+
+.personal-wechat-connection-dot {
+  width: 8px;
+  height: 8px;
+  border-radius: 50%;
+  background: var(--wxex-text-muted);
+
+  &.is-online {
+    background: var(--wxex-success);
+    box-shadow: 0 0 0 3px color-mix(in srgb, var(--wxex-success) 14%, transparent);
+  }
+}
+
+.personal-wechat-chat-body {
+  display: grid;
+  min-height: 0;
+  background: var(--wxex-bg-main);
+}
+
+.personal-wechat-message-list {
+  display: flex;
+  min-height: 96px;
+  flex-direction: column;
+  gap: 8px;
+  padding: 12px 20px;
+  background: color-mix(in srgb, var(--wxex-bg-main) 92%, var(--wxex-bg-sidebar));
+}
+
+.personal-wechat-system-message,
+.personal-wechat-empty-message {
+  align-self: center;
+  color: var(--wxex-text-muted);
+  font-size: 11px;
+}
+
+.personal-wechat-empty-message {
+  margin-top: 15px;
+}
+
+.personal-wechat-message-bubble {
+  display: grid;
+  max-width: min(76%, 440px);
+  gap: 3px;
+  align-self: flex-start;
+  border: 1px solid var(--wxex-border);
+  border-radius: 10px 10px 10px 3px;
+  background: var(--wxex-bg-elevated);
+  padding: 8px 10px;
+  color: var(--wxex-text-primary);
+  font-size: 12px;
+
+  &.is-outgoing {
+    align-self: flex-end;
+    border-color: color-mix(in srgb, var(--wxex-brand) 25%, transparent);
+    border-radius: 10px 10px 3px 10px;
+    background: color-mix(in srgb, var(--wxex-brand) 10%, var(--wxex-bg-elevated));
+  }
+}
+
+.personal-wechat-message-kind {
+  color: var(--wxex-text-muted);
+  font-size: 10px;
+}
+
+.personal-wechat-setup {
+  display: grid;
+  gap: 16px;
+  border-top: 1px solid var(--wxex-border);
+  background: var(--wxex-bg-elevated);
+  padding: 18px 20px 20px;
+}
+
+.personal-wechat-setup-heading {
+  display: flex;
+  align-items: flex-start;
+  justify-content: space-between;
+  gap: 12px;
+
+  h2,
+  p {
+    margin: 0;
+  }
+
+  h2 {
+    margin-top: 3px;
+    color: var(--wxex-text-primary);
+    font-size: 16px;
+    font-weight: 650;
+  }
+
+  p {
+    margin-top: 4px;
+    color: var(--wxex-text-secondary);
+    font-size: 12px;
+    line-height: 1.55;
+  }
+}
+
+.personal-wechat-eyebrow {
+  color: var(--wxex-brand);
+  font-size: 10px;
+  font-weight: 700;
+}
+
+.personal-wechat-setup-complete,
+.personal-wechat-step-status {
+  color: var(--wxex-success);
+  font-size: 11px;
+  font-weight: 650;
+}
+
+.personal-wechat-steps {
+  display: grid;
+  gap: 0;
+  margin: 0;
+  padding: 0;
+  list-style: none;
+
+  > li {
+    position: relative;
+    display: grid;
+    grid-template-columns: 24px minmax(0, 1fr);
+    gap: 10px;
+    padding-bottom: 11px;
+
+    &:not(:last-child)::after {
+      position: absolute;
+      top: 23px;
+      bottom: 0;
+      left: 11px;
+      width: 1px;
+      background: var(--wxex-border);
+      content: '';
+    }
+  }
+}
+
+.personal-wechat-step-number {
+  z-index: 1;
+  display: grid;
+  width: 23px;
+  height: 23px;
+  place-items: center;
+  border: 1px solid var(--wxex-border);
+  border-radius: 50%;
+  background: var(--wxex-bg-elevated);
+  color: var(--wxex-text-muted);
+  font-size: 11px;
+  font-weight: 700;
+}
+
+.personal-wechat-steps > li.is-current .personal-wechat-step-number,
+.personal-wechat-steps > li.is-active .personal-wechat-step-number {
+  border-color: var(--wxex-brand);
+  background: var(--wxex-brand-soft);
+  color: var(--wxex-brand);
+}
+
+.personal-wechat-steps > li.is-complete .personal-wechat-step-number {
+  border-color: color-mix(in srgb, var(--wxex-success) 35%, transparent);
+  background: color-mix(in srgb, var(--wxex-success) 12%, var(--wxex-bg-elevated));
+  color: var(--wxex-success);
+}
+
+.personal-wechat-step-content {
+  display: grid;
+  gap: 7px;
+  min-width: 0;
+
+  > strong {
+    color: var(--wxex-text-primary);
+    font-size: 13px;
+  }
+
+  > p {
+    margin: 0;
+    color: var(--wxex-text-secondary);
+    font-size: 11px;
+    line-height: 1.55;
+  }
+}
+
+.personal-wechat-step-hint {
+  color: var(--wxex-text-muted) !important;
+}
+
+.personal-wechat-step-warning {
+  margin: 0;
+  border-radius: var(--wxex-radius-md);
+  background: color-mix(in srgb, var(--wxex-warning) 10%, var(--wxex-bg-elevated));
+  padding: 9px 11px;
+  color: var(--wxex-text-secondary) !important;
+  font-size: 11px;
+  line-height: 1.55;
+}
+
+.personal-wechat-step-error,
+.personal-wechat-global-error {
+  color: var(--wxex-danger) !important;
+}
+
+.personal-wechat-download-progress {
+  display: grid;
+  gap: 6px;
+  max-width: 360px;
+
+  > div:first-child {
+    display: flex;
+    justify-content: space-between;
+    color: var(--wxex-text-muted);
+    font-size: 10px;
+  }
+}
+
+.personal-wechat-progress-track,
+.personal-wechat-preview-track {
+  position: relative;
+  overflow: hidden;
+  height: 4px;
+  border-radius: 999px;
+  background: color-mix(in srgb, var(--wxex-brand) 14%, transparent);
+
+  > span {
+    display: block;
+    height: 100%;
+    border-radius: inherit;
+    background: var(--wxex-brand);
+  }
+}
+
+.personal-wechat-capabilities {
+  display: grid;
+  grid-template-columns: repeat(2, minmax(0, 1fr));
+  gap: 7px;
+
+  > span {
+    display: grid;
+    grid-template-columns: auto 1fr;
+    align-items: center;
+    gap: 5px;
+    border: 1px solid var(--wxex-border);
+    border-radius: var(--wxex-radius-md);
+    padding: 8px;
+    color: var(--wxex-text-muted);
+    font-size: 11px;
+
+    &.is-ready {
+      border-color: color-mix(in srgb, var(--wxex-success) 30%, transparent);
+      color: var(--wxex-text-primary);
+    }
+
+    &:not(.is-ready) {
+      background: color-mix(in srgb, var(--wxex-bg-main) 70%, var(--wxex-bg-elevated));
+    }
+  }
+
+  b {
+    color: var(--wxex-text-muted);
+
+    .is-ready & {
+      color: var(--wxex-success);
+    }
+  }
+
+  small {
+    grid-column: 2;
+    color: var(--wxex-text-muted);
+    font-size: 10px;
+  }
+}
+
+.personal-wechat-diagnostics {
+  border-top: 1px solid var(--wxex-border);
+  padding-top: 12px;
+
+  summary {
+    cursor: pointer;
+    color: var(--wxex-text-secondary);
+    font-size: 11px;
+    font-weight: 600;
+  }
+
+  > p {
+    margin: 8px 0;
+    color: var(--wxex-text-muted);
+    font-size: 10px;
+  }
+
+  dl {
+    display: grid;
+    grid-template-columns: repeat(2, minmax(0, 1fr));
+    gap: 1px;
+    overflow: hidden;
+    margin: 0;
+    border: 1px solid var(--wxex-border);
+    border-radius: var(--wxex-radius-md);
+    background: var(--wxex-border);
+  }
+
+  dl > div {
+    min-width: 0;
+    background: var(--wxex-bg-main);
+    padding: 7px 9px;
+  }
+
+  dt {
+    color: var(--wxex-text-muted);
+    font-size: 10px;
+  }
+
+  dd {
+    overflow: hidden;
+    margin: 2px 0 0;
+    color: var(--wxex-text-secondary);
+    font-size: 10px;
+    text-overflow: ellipsis;
+    white-space: nowrap;
+  }
+}
+
+.personal-wechat-composer {
+  display: grid;
+  gap: 12px;
+  border-top: 1px solid var(--wxex-border);
+  background: var(--wxex-bg-elevated);
+  padding: 14px 20px 18px;
+}
+
+.personal-wechat-composer-mode {
+  display: grid;
+  grid-template-columns: repeat(3, minmax(0, 1fr));
+}
+
+.personal-wechat-text-editor,
+.personal-wechat-voice-editor,
+.personal-wechat-file-editor {
+  display: grid;
+  gap: 8px;
+}
+
+.personal-wechat-text-editor textarea {
+  resize: vertical;
+}
+
+.personal-wechat-text-editor > span {
+  justify-self: end;
+  color: var(--wxex-text-muted);
+  font-size: 10px;
+}
+
+.personal-wechat-file-editor {
+  grid-template-columns: minmax(0, 1fr) auto;
+  align-items: center;
+  gap: 10px;
+}
+
+.personal-wechat-file-editor > div {
+  min-width: 0;
+}
+
+.personal-wechat-file-editor strong,
+.personal-wechat-file-editor small {
+  display: block;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+
+.personal-wechat-file-editor strong {
+  color: var(--wxex-text-primary);
+  font-size: 12px;
+}
+
+.personal-wechat-file-editor small {
+  margin-top: 3px;
+  color: var(--wxex-text-muted);
+  font-size: 10px;
+}
+
+.personal-wechat-voice-heading,
+.personal-wechat-composer-footer {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  gap: 10px;
+}
+
+.personal-wechat-voice-heading > span {
+  color: var(--wxex-text-secondary);
+  font-size: 11px;
+}
+
+.personal-wechat-generated-result {
+  display: grid;
+  grid-template-columns: minmax(0, 1fr) auto;
+  align-items: center;
+  gap: 10px;
+  border: 1px solid color-mix(in srgb, var(--wxex-brand) 28%, transparent);
+  border-radius: var(--wxex-radius-md);
+  background: color-mix(in srgb, var(--wxex-brand) 7%, var(--wxex-bg-elevated));
+  padding: 9px;
+
+  audio {
+    display: none;
+  }
+
+  strong,
+  span {
+    display: block;
+  }
+
+  strong {
+    color: var(--wxex-text-primary);
+    font-size: 11px;
+  }
+
+  span {
+    color: var(--wxex-text-muted);
+    font-size: 10px;
+  }
+}
+
+.personal-wechat-composer-footer > span {
+  flex: 1;
+  font-size: 11px;
+
+  &.is-success {
+    color: var(--wxex-success);
+  }
+  &.is-error {
+    color: var(--wxex-danger);
+  }
+}
+
+.personal-wechat-global-error {
+  border-top: 1px solid color-mix(in srgb, var(--wxex-danger) 18%, transparent);
+  padding: 10px 20px 0;
+  font-size: 11px;
+}
+
+.personal-wechat-chat-footer {
+  display: flex;
+  justify-content: flex-end;
+  border-top: 1px solid var(--wxex-border);
+  padding: 12px 20px;
+}
+
+@media (max-width: 560px) {
+  .personal-wechat-capabilities {
+    grid-template-columns: 1fr;
+  }
+  .personal-wechat-chat-header,
+  .personal-wechat-message-list,
+  .personal-wechat-setup,
+  .personal-wechat-composer {
+    padding-inline: 14px;
+  }
+  .personal-wechat-generated-result {
+    grid-template-columns: 1fr;
+  }
+}
+
 .chat-header-actions svg,
 .chat-content-search svg,
 .data-trust-bar svg {
diff --git a/src/renderer/src/styles/settings-text-to-speech.scss b/src/renderer/src/styles/settings-text-to-speech.scss
index 5e52ca4..317d349 100644
--- a/src/renderer/src/styles/settings-text-to-speech.scss
+++ b/src/renderer/src/styles/settings-text-to-speech.scss
@@ -96,6 +96,35 @@
   }
 }
 
+.tts-auto-update-guide {
+  display: grid;
+  gap: 6px;
+  border-radius: var(--wxex-radius-md);
+  background: color-mix(in srgb, var(--wxex-warning) 9%, var(--wxex-bg-elevated));
+  padding: 12px 13px;
+
+  strong {
+    color: var(--wxex-text-primary);
+    font-size: 12px;
+  }
+
+  p {
+    margin: 0;
+    color: var(--wxex-text-secondary);
+    font-size: 10px;
+    line-height: 16px;
+  }
+
+  img {
+    width: min(100%, 640px);
+    max-height: 280px;
+    border: 1px solid var(--wxex-border);
+    border-radius: var(--wxex-radius-md);
+    object-fit: contain;
+    object-position: left top;
+  }
+}
+
 .tts-hook-warning {
   display: grid;
   grid-template-columns: auto minmax(0, 1fr);
diff --git a/tests/component/personal-wechat-send-dialog.test.tsx b/tests/component/personal-wechat-send-dialog.test.tsx
index c2e08f5..0401c1e 100644
--- a/tests/component/personal-wechat-send-dialog.test.tsx
+++ b/tests/component/personal-wechat-send-dialog.test.tsx
@@ -1,10 +1,13 @@
-import { useState } from 'react'
 import { fireEvent, render, screen, waitFor } from '@testing-library/react'
 import userEvent from '@testing-library/user-event'
+import type React from 'react'
 import { beforeEach, describe, expect, it, vi } from 'vitest'
 import { PersonalWechatSendDialog } from '../../src/renderer/src/components/chat/PersonalWechatSendDialog'
 
 const getStatus = vi.fn()
+const getRuntimeStatus = vi.fn()
+const downloadRuntime = vi.fn()
+const onRuntimeProgress = vi.fn(() => vi.fn())
 const rebind = vi.fn()
 const selectImage = vi.fn()
 const selectVoice = vi.fn()
@@ -20,7 +23,6 @@ const contact = {
   md5: 'fixture-md5',
   type: 'group' as const
 }
-
 const readyStatus = {
   state: 'online' as const,
   platform: 'darwin',
@@ -48,42 +50,44 @@ const readyStatus = {
   canSendVoice: true,
   message: '个人微信已绑定'
 }
+const readyRuntime = {
+  version: 'v0.0.18',
+  state: 'ready' as const,
+  downloadedBytes: 1,
+  totalBytes: 1,
+  progress: 1,
+  platform: 'darwin' as NodeJS.Platform,
+  architecture: 'arm64',
+  supported: true,
+  removable: true
+}
 
-function Harness({ onClose = vi.fn() }: { onClose?: () => void }): React.ReactElement {
-  const [open, setOpen] = useState(false)
-  return (
-    <>
-      <button type="button" onClick={() => setOpen(true)}>
-        打开测试发送
-      </button>
-      {open && (
-        <PersonalWechatSendDialog
-          contact={contact}
-          isGroupChat
-          onClose={() => {
-            onClose()
-            setOpen(false)
-          }}
-        />
-      )}
-    </>
+function renderDialog(
+  props: Partial<React.ComponentProps<typeof PersonalWechatSendDialog>> = {}
+): React.ReactElement {
+  return render(
+    <PersonalWechatSendDialog contact={contact} isGroupChat onClose={vi.fn()} {...props} />
   )
 }
 
+async function startComposer(): Promise<void> {
+  const start = await screen.findByRole('button', { name: '开始发送' })
+  fireEvent.click(start)
+}
+
 describe('PersonalWechatSendDialog', () => {
   beforeEach(() => {
     getStatus.mockReset().mockResolvedValue(readyStatus)
+    getRuntimeStatus.mockReset().mockResolvedValue(readyRuntime)
+    downloadRuntime.mockReset().mockResolvedValue({ success: true, status: readyRuntime })
+    onRuntimeProgress.mockReset().mockReturnValue(vi.fn())
     rebind.mockReset().mockResolvedValue(readyStatus)
-    selectImage.mockReset().mockResolvedValue({
-      canceled: false,
-      path: '/Users/fixture/test.png',
-      name: 'test.png'
-    })
-    selectVoice.mockReset().mockResolvedValue({
-      canceled: false,
-      path: '/Users/fixture/test.silk',
-      name: 'test.silk'
-    })
+    selectImage
+      .mockReset()
+      .mockResolvedValue({ canceled: false, path: '/Users/fixture/test.png', name: 'test.png' })
+    selectVoice
+      .mockReset()
+      .mockResolvedValue({ canceled: false, path: '/Users/fixture/test.silk', name: 'test.silk' })
     sendMessage.mockReset().mockResolvedValue({ success: true, status: readyStatus })
     getTextToSpeechSettings.mockReset().mockResolvedValue({
       success: true,
@@ -125,6 +129,9 @@ describe('PersonalWechatSendDialog', () => {
       configurable: true,
       value: {
         getPersonalWechatSenderStatus: getStatus,
+        getPersonalWechatRuntimeStatus: getRuntimeStatus,
+        downloadPersonalWechatRuntime: downloadRuntime,
+        onPersonalWechatRuntimeProgress: onRuntimeProgress,
         rebindPersonalWechatSender: rebind,
         selectPersonalWechatImage: selectImage,
         selectPersonalWechatVoice: selectVoice,
@@ -137,14 +144,84 @@ describe('PersonalWechatSendDialog', () => {
     })
   })
 
-  it('switches to voice mode and sends only the selected audio file', async () => {
-    render(<PersonalWechatSendDialog contact={contact} isGroupChat onClose={vi.fn()} />)
-    await screen.findByText('技术交流群')
+  it('shows a user-facing four-step setup and keeps diagnostics collapsed', async () => {
+    getRuntimeStatus
+      .mockResolvedValueOnce({ ...readyRuntime, state: 'missing', progress: 0 })
+      .mockResolvedValue(readyRuntime)
+    getStatus.mockResolvedValue({
+      ...readyStatus,
+      state: 'stopped',
+      runtimeReady: false,
+      canSend: false,
+      canSendText: false,
+      canSendImage: false,
+      canSendVoice: false,
+      message: '尚未绑定当前微信'
+    })
+    renderDialog()
+    expect(await screen.findByText('准备语音模型')).toBeInTheDocument()
+    expect(screen.getByRole('button', { name: '下载模型' })).toBeEnabled()
+    expect(screen.getByText('绑定个人微信')).toBeInTheDocument()
+    expect(screen.getByText('验证消息能力')).toBeInTheDocument()
+    expect(screen.getByText('能力检测')).toBeInTheDocument()
+    expect(screen.getByText('图片和语音消息')).toBeInTheDocument()
+    fireEvent.click(screen.getByRole('button', { name: '查看支持的微信版本' }))
+    const versionsDialog = screen.getByRole('dialog', { name: '支持的微信版本' })
+    expect(versionsDialog).toBeInTheDocument()
+    expect(versionsDialog).toHaveTextContent('4.1.11.53')
+    expect(
+      screen.getByText(
+        '绑定微信可能导致当前微信异常闪退，这是正常现象。若微信退出，请重新启动微信后，再回到这里重新检测/绑定。'
+      )
+    ).toBeInTheDocument()
+    expect(screen.queryByLabelText('消息列表')).not.toBeInTheDocument()
+    expect(screen.queryByText('TraceMemo 消息发送')).not.toBeInTheDocument()
+    expect(screen.queryByText('PID 4668')).not.toBeVisible()
+    fireEvent.click(screen.getByText('高级诊断'))
+    expect(screen.getByText('PID 4668')).toBeInTheDocument()
+  })
+
+  it('sends text through the existing message API and echoes it in the chat', async () => {
+    renderDialog()
+    await startComposer()
+    await userEvent
+      .setup()
+      .type(screen.getByRole('textbox', { name: '消息内容' }), '你好 TraceMemo')
+    fireEvent.click(screen.getByRole('button', { name: '发送消息' }))
+    await waitFor(() =>
+      expect(sendMessage).toHaveBeenCalledWith({
+        type: 'text',
+        to: 'fixture-room@chatroom',
+        text: '你好 TraceMemo',
+        isGroup: true
+      })
+    )
+    expect(
+      screen.getByLabelText('消息列表').querySelector('.personal-wechat-message-bubble')
+    ).toHaveTextContent('你好 TraceMemo')
+  })
+
+  it('supports local image and voice selection', async () => {
+    renderDialog()
+    await startComposer()
+    fireEvent.click(screen.getByRole('radio', { name: '图片' }))
+    fireEvent.click(screen.getByRole('button', { name: '选择图片' }))
+    expect(await screen.findByText('test.png')).toBeInTheDocument()
+    fireEvent.click(screen.getByRole('button', { name: '发送消息' }))
+    await waitFor(() =>
+      expect(sendMessage).toHaveBeenCalledWith({
+        type: 'image',
+        to: 'fixture-room@chatroom',
+        filePath: '/Users/fixture/test.png',
+        isGroup: true
+      })
+    )
+
     fireEvent.click(screen.getByRole('radio', { name: '语音' }))
     fireEvent.click(screen.getByRole('radio', { name: '选择本地文件' }))
     fireEvent.click(screen.getByRole('button', { name: '选择语音' }))
     expect(await screen.findByText('test.silk')).toBeInTheDocument()
-    fireEvent.click(screen.getByRole('button', { name: '测试发送语音到群聊' }))
+    fireEvent.click(screen.getByRole('button', { name: '发送消息' }))
     await waitFor(() =>
       expect(sendMessage).toHaveBeenCalledWith({
         type: 'voice',
@@ -155,136 +232,101 @@ describe('PersonalWechatSendDialog', () => {
     )
   })
 
-  it('shows the selected TTS voice and keeps generation separate from sending', async () => {
-    const openSettings = vi.fn()
-    render(
-      <PersonalWechatSendDialog
-        contact={contact}
-        isGroupChat
-        onClose={vi.fn()}
-        onOpenTextToSpeechSettings={openSettings}
-      />
-    )
-    await screen.findByText('技术交流群')
-    fireEvent.click(screen.getByRole('radio', { name: '语音' }))
-    expect(await screen.findByText('暖阳女声')).toBeInTheDocument()
-    expect(screen.getByRole('textbox', { name: '要生成的文字' })).toHaveValue('1')
-    expect(screen.getByRole('button', { name: '生成语音' })).toBeEnabled()
-    fireEvent.click(screen.getByRole('button', { name: '前往文字转语音设置' }))
-    expect(openSettings).toHaveBeenCalledTimes(1)
-    expect(sendMessage).not.toHaveBeenCalled()
-  })
-
-  it('shows binding diagnostics and only offers image and voice modes', async () => {
-    render(<PersonalWechatSendDialog contact={contact} isGroupChat onClose={vi.fn()} />)
-
-    expect(await screen.findByText('PID 4668')).toBeInTheDocument()
-    expect(screen.getByText('PID 5401 · 绑定 4668')).toBeInTheDocument()
-    expect(screen.getByText('0x114ef8000')).toBeInTheDocument()
-    expect(screen.getByText('已捕获，可发送')).toBeInTheDocument()
-    expect(screen.getByRole('radio', { name: '图片' })).toHaveAttribute('aria-checked', 'true')
-    expect(screen.getByRole('radio', { name: '语音' })).toBeVisible()
-    expect(screen.queryByRole('radio', { name: '文字' })).not.toBeInTheDocument()
-  })
-
-  it('supports segmented keyboard navigation for the message type', async () => {
-    const user = userEvent.setup()
-    render(<PersonalWechatSendDialog contact={contact} isGroupChat onClose={vi.fn()} />)
-
-    await screen.findByText('技术交流群')
-    const imageMode = screen.getByRole('radio', { name: '图片' })
-    const voiceMode = screen.getByRole('radio', { name: '语音' })
-    imageMode.focus()
-    await user.keyboard('{ArrowRight}')
-    expect(voiceMode).toHaveFocus()
-    await user.keyboard(' ')
-    expect(voiceMode).toBeChecked()
-    expect(screen.getByRole('textbox', { name: '要生成的文字' })).toBeVisible()
-  })
-
-  it('switches to image mode and sends only the selected image', async () => {
-    render(<PersonalWechatSendDialog contact={contact} isGroupChat onClose={vi.fn()} />)
-    await screen.findByText('技术交流群')
-    fireEvent.click(screen.getByRole('radio', { name: '图片' }))
-    expect(screen.queryByRole('textbox')).not.toBeInTheDocument()
-    expect(
-      screen.getByText(
-        '如果想测试图片，请先在微信中给任意好友手动发送一张普通图片，再点击重新检测。',
-        { exact: false }
-      )
-    ).toBeInTheDocument()
-    fireEvent.click(screen.getByRole('button', { name: '选择图片' }))
-    expect(await screen.findByText('test.png')).toBeInTheDocument()
-    fireEvent.click(screen.getByRole('button', { name: '测试发送图片到群聊' }))
-
-    await waitFor(() =>
-      expect(sendMessage).toHaveBeenCalledWith({
-        type: 'image',
-        to: 'fixture-room@chatroom',
-        filePath: '/Users/fixture/test.png',
-        isGroup: true
+  it('uses the existing runtime, binding and detection IPC actions', async () => {
+    getRuntimeStatus
+      .mockResolvedValueOnce({ ...readyRuntime, state: 'missing', progress: 0 })
+      .mockResolvedValue(readyRuntime)
+    getStatus
+      .mockResolvedValueOnce({
+        ...readyStatus,
+        state: 'stopped',
+        runtimeReady: false,
+        canSend: false,
+        canSendText: false,
+        canSendImage: false,
+        canSendVoice: false
       })
-    )
+      .mockResolvedValue({
+        ...readyStatus,
+        state: 'stopped',
+        runtimeReady: true,
+        canSend: false,
+        canSendText: false,
+        canSendImage: false,
+        canSendVoice: false
+      })
+    renderDialog()
+    await screen.findByRole('button', { name: '下载模型' })
+    fireEvent.click(screen.getByRole('button', { name: '下载模型' }))
+    await waitFor(() => expect(downloadRuntime).toHaveBeenCalledOnce())
+    fireEvent.click(screen.getByRole('button', { name: '绑定微信' }))
+    await waitFor(() => expect(rebind).toHaveBeenCalledOnce())
   })
 
-  it('offers a safe explicit rebind action', async () => {
-    render(<PersonalWechatSendDialog contact={contact} isGroupChat onClose={vi.fn()} />)
-    await screen.findByText('技术交流群')
-    fireEvent.click(screen.getByRole('button', { name: '尝试重新绑定' }))
-    await waitFor(() => expect(rebind).toHaveBeenCalledTimes(1))
+  it('shows a stable bound state after binding', async () => {
+    getRuntimeStatus.mockResolvedValue(readyRuntime)
+    getStatus
+      .mockResolvedValueOnce({
+        ...readyStatus,
+        state: 'stopped',
+        canSend: false,
+        canSendText: false,
+        canSendImage: false,
+        canSendVoice: false
+      })
+      .mockResolvedValue(readyStatus)
+    renderDialog()
+    await screen.findByRole('button', { name: '绑定微信' })
+    fireEvent.click(screen.getByRole('button', { name: '绑定微信' }))
+    expect(await screen.findByText('✓ 微信已绑定')).toBeInTheDocument()
+    expect(screen.queryByRole('button', { name: '绑定微信' })).not.toBeInTheDocument()
   })
 
-  it('blocks image sending until the media Hook and image directory are initialized', async () => {
+  it('treats image and voice readiness as one media capability', async () => {
     getStatus.mockResolvedValue({
       ...readyStatus,
-      state: 'hook_not_ready',
-      imageHookReady: false,
       canSendImage: false,
-      message: '请先手动发送图片'
+      canSendVoice: true
     })
-    render(<PersonalWechatSendDialog contact={contact} isGroupChat onClose={vi.fn()} />)
-    expect(await screen.findByText('已绑定，等待消息初始化')).toBeInTheDocument()
-    expect(screen.getByRole('button', { name: '测试发送图片到群聊' })).toBeDisabled()
+    renderDialog()
+    expect(await screen.findByText('图片和语音消息')).toBeInTheDocument()
+    expect(screen.getByText('微信消息发送已配置完成')).toBeInTheDocument()
+    expect(screen.getByRole('button', { name: '开始发送' })).toBeEnabled()
+    expect(screen.queryByText('图片消息')).not.toBeInTheDocument()
+    expect(screen.queryByText('语音消息')).not.toBeInTheDocument()
   })
 
-  it('closes with Escape or the overlay and restores focus to the opener', async () => {
-    const user = userEvent.setup()
-    const onClose = vi.fn()
-    render(<Harness onClose={onClose} />)
-    const opener = screen.getByRole('button', { name: '打开测试发送' })
-
-    await user.click(opener)
-    await screen.findByText('技术交流群')
-    await user.keyboard('{Escape}')
-    expect(screen.queryByRole('dialog', { name: '个人微信测试发送' })).not.toBeInTheDocument()
-    await waitFor(() => expect(opener).toHaveFocus())
-
-    await user.click(opener)
-    const dialog = await screen.findByRole('dialog', { name: '个人微信测试发送' })
-    await user.click(dialog.previousElementSibling as HTMLElement)
-    expect(screen.queryByRole('dialog', { name: '个人微信测试发送' })).not.toBeInTheDocument()
-    await waitFor(() => expect(opener).toHaveFocus())
-    expect(onClose).toHaveBeenCalledTimes(2)
+  it('keeps re-detection available and explains when no new messages are found', async () => {
+    getRuntimeStatus.mockResolvedValue(readyRuntime)
+    getStatus.mockResolvedValue({
+      ...readyStatus,
+      canSend: false,
+      canSendText: false,
+      canSendImage: false,
+      canSendVoice: false,
+      message: '等待消息初始化'
+    })
+    renderDialog()
+    const detect = await screen.findByRole('button', { name: '重新检测' })
+    expect(detect).toBeEnabled()
+    fireEvent.click(detect)
+    expect(
+      await screen.findByText(
+        '暂未检测到新的消息，请确认已在手机微信中发送文字和图片，然后再次检测。'
+      )
+    ).toBeInTheDocument()
+    expect(screen.getAllByText('未检测').length).toBe(2)
   })
 
-  it('keeps the dialog open while a message is being sent', async () => {
+  it('keeps the dialog and controls stable while sending', async () => {
     const user = userEvent.setup()
     sendMessage.mockImplementation(() => new Promise(() => undefined))
-    render(<Harness />)
-
-    await user.click(screen.getByRole('button', { name: '打开测试发送' }))
-    await screen.findByText('技术交流群')
-    await user.click(screen.getByRole('button', { name: '选择图片' }))
-    await user.click(screen.getByRole('button', { name: '测试发送图片到群聊' }))
+    renderDialog()
+    await startComposer()
+    await screen.findByRole('textbox', { name: '消息内容' })
+    await user.type(screen.getByRole('textbox', { name: '消息内容' }), '发送中')
+    await user.click(screen.getByRole('button', { name: '发送消息' }))
     expect(screen.getByRole('button', { name: '正在发送…' })).toBeDisabled()
     expect(screen.getByRole('radio', { name: '图片' })).toBeDisabled()
-    expect(screen.getByRole('radio', { name: '语音' })).toBeDisabled()
-
-    await user.keyboard('{Escape}')
-    expect(screen.getByRole('dialog', { name: '个人微信测试发送' })).toBeVisible()
-    await user.click(
-      screen.getByRole('dialog', { name: '个人微信测试发送' }).previousElementSibling as HTMLElement
-    )
-    expect(screen.getByRole('dialog', { name: '个人微信测试发送' })).toBeVisible()
   })
 })
diff --git a/tests/component/report-features.test.tsx b/tests/component/report-features.test.tsx
index cd551fd..f900df3 100644
--- a/tests/component/report-features.test.tsx
+++ b/tests/component/report-features.test.tsx
@@ -608,7 +608,7 @@ describe('daily report controls', () => {
 
     fireEvent.click(screen.getByRole('button', { name: '发送到当前群聊' }))
 
-    expect(await screen.findByRole('dialog', { name: '个人微信测试发送' })).toBeVisible()
+    expect(await screen.findByRole('dialog', { name: '测试群' })).toBeVisible()
     expect(screen.getByRole('radio', { name: '图片' })).toHaveAttribute('aria-checked', 'true')
     expect(screen.getByText('测试群日报.png')).toBeVisible()
     expect(screen.getByText('/Users/fixture/测试群日报.png')).toBeVisible()
diff --git a/tests/e2e/app.spec.ts b/tests/e2e/app.spec.ts
index 0a665b4..9184469 100644
--- a/tests/e2e/app.spec.ts
+++ b/tests/e2e/app.spec.ts
@@ -190,11 +190,10 @@ test('CHAT-02 personal WeChat send dialog is keyboard-safe and fits the viewport
     await fixture.page.getByText('产品测试群', { exact: true }).click()
     const trigger = fixture.page.getByRole('button', { name: '发送消息' })
     await trigger.click()
-    const dialog = fixture.page.getByRole('dialog', { name: '个人微信测试发送' })
+    const dialog = fixture.page.getByRole('dialog', { name: '产品测试群' })
     await expect(dialog).toBeVisible()
-    await expect(
-      dialog.locator('.personal-wechat-send-status strong').filter({ hasText: '个人微信已绑定' })
-    ).toBeVisible()
+    const startSending = dialog.getByRole('button', { name: '开始发送' })
+    if (await startSending.isVisible()) await startSending.click()
     expect(
       await fixture.page.evaluate(() => document.documentElement.scrollWidth <= window.innerWidth)
     ).toBe(true)
@@ -208,7 +207,7 @@ test('CHAT-02 personal WeChat send dialog is keyboard-safe and fits the viewport
     expect(pageErrors).toEqual([])
 
     const imageMode = dialog.getByRole('radio', { name: '图片' })
-    const voiceMode = dialog.getByRole('radio', { name: '语音' })
+    const voiceMode = dialog.getByRole('radio', { name: '语音', exact: true })
     await imageMode.focus()
     await fixture.page.keyboard.press('ArrowRight')
     await expect(voiceMode).toBeFocused()
diff --git a/tests/e2e/visual.spec.ts b/tests/e2e/visual.spec.ts
index 6d034f0..2bc9e20 100644
--- a/tests/e2e/visual.spec.ts
+++ b/tests/e2e/visual.spec.ts
@@ -387,7 +387,7 @@ for (const appearanceTheme of ['light', 'dark'] as const) {
       await fixture.setWindowContentSize(visualViewport)
       await fixture.page.getByText('产品测试群', { exact: true }).click()
       await fixture.page.getByRole('button', { name: '发送消息' }).click()
-      const dialog = fixture.page.getByRole('dialog', { name: '个人微信测试发送' })
+      const dialog = fixture.page.getByRole('dialog', { name: '产品测试群' })
       await expect(dialog).toBeVisible()
       await expect(fixture.page.locator('html')).toHaveAttribute('data-theme', appearanceTheme)
       expect(
@@ -406,9 +406,11 @@ for (const appearanceTheme of ['light', 'dark'] as const) {
         caret: 'hide'
       })
 
+      const startSending = dialog.getByRole('button', { name: '开始发送' })
+      if (await startSending.isVisible()) await startSending.click()
       await dialog.getByRole('radio', { name: '语音' }).click()
-      await expect(dialog.getByRole('textbox', { name: '要生成的文字' })).toBeVisible()
-      await dialog.locator('footer').scrollIntoViewIfNeeded()
+      await expect(dialog.getByRole('textbox', { name: '语音文字' })).toBeVisible()
+      await dialog.locator('.personal-wechat-composer').scrollIntoViewIfNeeded()
       await clearScreenshotFocus(fixture.page)
       await expect(fixture.page).toHaveScreenshot(
         `personal-wechat-voice-controls-${appearanceTheme}.png`,
diff --git a/tests/unit/wcdb-close.test.ts b/tests/unit/wcdb-close.test.ts
index a7e0bb1..be7d871 100644
--- a/tests/unit/wcdb-close.test.ts
+++ b/tests/unit/wcdb-close.test.ts
@@ -6,6 +6,25 @@ function setPrivate(target: object, key: string, value: unknown): void {
 }
 
 describe('Wcdb4Client shutdown', () => {
+  it('clears folded contact status snapshots when sessions are invalidated', () => {
+    const client = Object.create(Wcdb4Client.prototype) as Wcdb4Client
+    const statusCache = new Map([['group@chatroom', { isFolded: true, isMuted: false }]])
+    setPrivate(client, 'sessionCacheGeneration', 3)
+    setPrivate(client, 'cachedSessions', [{ username: 'group@chatroom' }])
+    setPrivate(client, 'cachedChatTables', [{ name: 'Chat_group', db_number: 'group@chatroom' }])
+    setPrivate(client, 'sessionStatusCache', statusCache)
+    setPrivate(client, 'sessionStatusesUpdatedAt', Date.now())
+
+    client.invalidateSessionCache()
+
+    expect(statusCache.size).toBe(0)
+    expect(client).toMatchObject({
+      cachedSessions: null,
+      cachedChatTables: null,
+      sessionStatusesUpdatedAt: 0
+    })
+  })
+
   it('waits for tracked Koffi calls before shutting down the native runtime', async () => {
     const client = Object.create(Wcdb4Client.prototype) as Wcdb4Client
     const shutdown = vi.fn(() => 0)