diff --git a/src/wechat/media.ts b/src/wechat/media.ts
index fe053b4..807229b 100644
--- a/src/wechat/media.ts
+++ b/src/wechat/media.ts
@@ -73,9 +73,15 @@ export async function downloadImage(item: MessageItem): Promise<string | null> {
 
 /**
  * Extract text content from a message item.
- * Returns text_item.text or empty string.
+ * Handles text items and voice items (using ASR transcription).
  */
 export function extractText(item: MessageItem): string {
+  if (item.type === MessageItemType.VOICE) {
+    const voiceText = item.voice_item?.voice_text || item.voice_item?.text;
+    if (voiceText) {
+      return `[语音] ${voiceText}`;
+    }
+  }
   return item.text_item?.text ?? '';
 }
 
diff --git a/src/wechat/types.ts b/src/wechat/types.ts
index df135fe..46b255e 100644
--- a/src/wechat/types.ts
+++ b/src/wechat/types.ts
@@ -48,7 +48,9 @@ export interface ImageItem {
 
 export interface VoiceItem {
   cdn_media: CDNMedia;
+  media?: { encrypt_query_param: string; aes_key?: string };
   voice_text?: string;
+  text?: string;
 }
 
 export interface FileItem {