diff --git a/runtime/docs/websocket_protocol.md b/runtime/docs/websocket_protocol.md
index 867a18290..588339e2d 100644
--- a/runtime/docs/websocket_protocol.md
+++ b/runtime/docs/websocket_protocol.md
@@ -45,7 +45,7 @@ Parameter explanation:
 `text`: the text output of speech recognition
 `is_final`: indicating the end of recognition
 `timestamp`：If AM is a timestamp model, it will return this field, indicating the timestamp, in the format of "[[100,200], [200,500]]"
-`stamp_sents`：If AM is a timestamp model, it will return this field, indicating the stamp_sents, in the format of "[{'text':'正 是 因 为','start':'430','end':'1130','ts_list':[[430,670],[670,810],[810,1030],[1030,1130]]}]"
+`stamp_sents`：If AM is a timestamp model, it will return this field, indicating the stamp_sents, in the format of "[{'text_seg':'正 是 因 为','punc':',','start':'430','end':'1130','ts_list':[[430,670],[670,810],[810,1030],[1030,1130]]}]"
 ```
 
 ## Real-time Speech Recognition
@@ -94,5 +94,5 @@ Parameter explanation:
 `text`: the text output of speech recognition
 `is_final`: indicating the end of recognition
 `timestamp`：If AM is a timestamp model, it will return this field, indicating the timestamp, in the format of "[[100,200], [200,500]]"
-`stamp_sents`：If AM is a timestamp model, it will return this field, indicating the stamp_sents, in the format of "[{'text':'正 是 因 为','start':'430','end':'1130','ts_list':[[430,670],[670,810],[810,1030],[1030,1130]]}]"
+`stamp_sents`：If AM is a timestamp model, it will return this field, indicating the stamp_sents, in the format of "[{'text_seg':'正 是 因 为','punc':',','start':'430','end':'1130','ts_list':[[430,670],[670,810],[810,1030],[1030,1130]]}]"
 ```
diff --git a/runtime/docs/websocket_protocol_zh.md b/runtime/docs/websocket_protocol_zh.md
index 98b5c2041..e00a989ef 100644
--- a/runtime/docs/websocket_protocol_zh.md
+++ b/runtime/docs/websocket_protocol_zh.md
@@ -46,7 +46,7 @@ message为（采用json序列化）
 `text`：表示语音识别输出文本
 `is_final`：表示识别结束
 `timestamp`：如果AM为时间戳模型，会返回此字段，表示时间戳，格式为 "[[100,200], [200,500]]"(ms)
-`stamp_sents`：如果AM为时间戳模型，会返回此字段，表示句子级别时间戳，格式为 "[{'text':'正 是 因 为','start':'430','end':'1130','ts_list':[[430,670],[670,810],[810,1030],[1030,1130]]}]"
+`stamp_sents`：如果AM为时间戳模型，会返回此字段，表示句子级别时间戳，格式为 "[{'text_seg':'正 是 因 为','punc':',','start':'430','end':'1130','ts_list':[[430,670],[670,810],[810,1030],[1030,1130]]}]"
 ```
 
 ## 实时语音识别
@@ -96,5 +96,5 @@ message为（采用json序列化）
 `text`：表示语音识别输出文本
 `is_final`：表示识别结束
 `timestamp`：如果AM为时间戳模型，会返回此字段，表示时间戳，格式为 "[[100,200], [200,500]]"(ms)
-`stamp_sents`：如果AM为时间戳模型，会返回此字段，表示句子级别时间戳，格式为 "[{'text':'正 是 因 为','start':'430','end':'1130','ts_list':[[430,670],[670,810],[810,1030],[1030,1130]]}]"
+`stamp_sents`：如果AM为时间戳模型，会返回此字段，表示句子级别时间戳，格式为 "[{'text_seg':'正 是 因 为','punc':',','start':'430','end':'1130','ts_list':[[430,670],[670,810],[810,1030],[1030,1130]]}]"
 ```
diff --git a/runtime/onnxruntime/src/util.cpp b/runtime/onnxruntime/src/util.cpp
index 661a50d8b..0fbec4f7b 100644
--- a/runtime/onnxruntime/src/util.cpp
+++ b/runtime/onnxruntime/src/util.cpp
@@ -584,7 +584,8 @@ std::string TimestampSentence(std::string &text, std::string &str_time){
                 }
             }
             // format
-            ts_sent += "{'text':'" + text_seg + "',";
+            ts_sent += "{'text_seg':'" + text_seg + "',";
+            ts_sent += "'punc':'" + characters[idx_str] + "',";
             ts_sent += "'start':'" + to_string(start) + "',";
             ts_sent += "'end':'" + to_string(end) + "',";
             ts_sent += "'ts_list':" + VectorToString(ts_seg) + "}";
@@ -620,7 +621,8 @@ std::string TimestampSentence(std::string &text, std::string &str_time){
             end = ts_seg[ts_seg.size()-1][1];
         }
         // format
-        ts_sent += "{'text':'" + text_seg + "',";
+        ts_sent += "{'text_seg':'" + text_seg + "',";
+        ts_sent += "'punc':'',";
         ts_sent += "'start':'" + to_string(start) + "',";
         ts_sent += "'end':'" + to_string(end) + "',";
         ts_sent += "'ts_list':" + VectorToString(ts_seg) + "}";