diff --git a/wechat_ipad/client/message.py b/wechat_ipad/client/message.py
index 7fcacf2..517c19a 100644
--- a/wechat_ipad/client/message.py
+++ b/wechat_ipad/client/message.py
@@ -216,6 +216,9 @@ class MessageMixin(WechatAPIClientBase):
                     ValueError: 视频或图片参数都为空或都不为空时
                     根据error_handler处理错误
                 """
+        has_image = False
+        if not image:
+            image = Path(os.path.join(Path(__file__).resolve().parent, "fallback.png"))
 
         # get video base64 and duration
         if isinstance(video, str):
@@ -236,9 +239,11 @@ class MessageMixin(WechatAPIClientBase):
             file_len = len(video_bytes)
             vid_base64 = base64.b64encode(video_bytes).decode()
             media_info = MediaInfo.parse(video_path)
-            first_frame = _get_first_frame(video_path, f"dify_frame_{int(time.time())}.jpg")
-            if first_frame:
-                image = first_frame
+            # 如果没有传入首帧，则自己提取一次
+            if not has_image:
+                first_frame = _get_first_frame(video_path, f"dify_frame_{int(time.time())}.jpg")
+                if first_frame:
+                    image = first_frame
         else:
             raise ValueError("video should be str, bytes, or path")
             # 获取视频时长
@@ -250,8 +255,6 @@ class MessageMixin(WechatAPIClientBase):
         if duration is None:
             duration = 1
             self.logging.error(f"无法从视频文件获取时长: {video}")
-        if not image:
-            image = Path(os.path.join(Path(__file__).resolve().parent, "fallback.png"))
         # get image base64
         if isinstance(image, str):
             image_base64 = image