diff --git a/__init__.py b/__init__.py
new file mode 100644
index 0000000..cac219e
--- /dev/null
+++ b/__init__.py
@@ -0,0 +1 @@
+from .logpp import *
diff --git a/logpp.py b/logpp.py
new file mode 100644
index 0000000..06082be
--- /dev/null
+++ b/logpp.py
@@ -0,0 +1,113 @@
+# encoding:utf-8
+
+import plugins
+from bridge.context import ContextType
+from bridge.reply import Reply, ReplyType
+from common.log import logger
+from plugins import *
+import datetime
+import tiktoken
+import csv
+from datetime import datetime
+
+def create_csv_file(file_name):
+    with open(file_name, mode='w', newline='', encoding='utf-8') as file:
+        writer = csv.writer(file)
+        writer.writerow(["时间", "用户名", "群聊", "输入token数", "输出token数", "实际请求内容", "装饰后输出内容"])
+
+def add_row_to_csv(file_name, current_time, user, group, in_token, out_token, input_text, reply):
+    with open(file_name, mode='a', newline='', encoding='utf-8') as file:
+        writer = csv.writer(file)
+        writer.writerow([current_time, user, group, in_token, out_token, input_text, reply])
+
+def num_tokens_from_string(string: str, encoding) -> int:
+    """Returns the number of tokens in a text string."""
+    num_tokens = len(encoding.encode(string))
+    return num_tokens
+
+
+@plugins.register(
+    name="logpp",
+    desire_priority=5,
+    hidden=True,
+    desc="A plugin that the usage of each user(group) as a csv file",
+    version="0.1",
+    author="loping151",
+)
+class logpp(Plugin):
+    def __init__(self):
+        super().__init__()
+        if not os.path.exists(os.path.join(os.path.dirname(__file__), 'logs')):
+            os.makedirs(os.path.join(os.path.dirname(__file__), 'logs'))
+            
+        self.handlers[Event.ON_DECORATE_REPLY] = self.on_decorate_reply # 输出的部分，在其他插件装饰回复前就计算
+        self.encoding = tiktoken.get_encoding("cl100k_base") # support gpt-4, gpt-3.5-turbo, text-embedding-ada-002. Davinci(GPT-3) not supported
+        
+        logger.info("[logpp] inited")
+        
+    @property 
+    def csv(self):
+        if not os.path.exists(os.path.join(os.path.dirname(__file__), 'logs', self.today+".csv")):
+            create_csv_file(os.path.join(os.path.dirname(__file__), 'logs', self.today+".csv"))
+        return os.path.join(os.path.dirname(__file__), 'logs', self.today+".csv")
+    
+    @property
+    def today(self): 
+        return datetime.now().strftime("%Y-%m-%d")
+     
+    @property
+    def now(self):
+        return datetime.now().strftime("%H:%M:%S")
+           
+    def on_decorate_reply(self, e_context: EventContext):
+        
+        if e_context["context"].type not in [
+            ContextType.TEXT,
+            ContextType.IMAGE
+        ]:
+            return
+        
+        if e_context["reply"].type not in [
+            ReplyType.TEXT,
+            ReplyType.IMAGE
+        ]:
+            return
+            
+        content = e_context['context'].content
+        is_group = e_context['context'].kwargs['msg'].is_group # certain group
+        group = e_context['context'].kwargs['msg'].other_user_nickname
+        if is_group:
+            user = e_context['context'].kwargs['msg'].actual_user_nickname
+        else: 
+            user = e_context['context'].kwargs['msg'].from_user_nickname
+        reply = e_context['reply']
+        
+        if e_context["context"].type == ContextType.TEXT:
+            if reply.type == ReplyType.TEXT:
+                if is_group:
+                    add_row_to_csv(self.csv, self.now, user, group, num_tokens_from_string(content, self.encoding), num_tokens_from_string(reply.content, self.encoding), content, reply.content)
+                else:
+                    add_row_to_csv(self.csv, self.now, user, "", num_tokens_from_string(content, self.encoding), num_tokens_from_string(reply.content, self.encoding), content, reply.content)
+            elif reply.type == ReplyType.IMAGE:
+                if is_group:
+                    add_row_to_csv(self.csv, self.now, user, group, num_tokens_from_string(content, self.encoding), 0, content, "image")
+                else:
+                    add_row_to_csv(self.csv, self.now, user, "", num_tokens_from_string(content, self.encoding), 0, content, "image")
+            return
+
+        if e_context["context"].type == ContextType.IMAGE:
+            if reply.type == ReplyType.TEXT:
+                if is_group:
+                    add_row_to_csv(self.csv, self.now, user, group, 0, num_tokens_from_string(reply.content, self.encoding), "image", reply.content)
+                else:
+                    add_row_to_csv(self.csv, self.now, user, "", 0, num_tokens_from_string(reply.content, self.encoding), "image", reply.content)
+            elif reply.type == ReplyType.IMAGE:
+                if is_group:
+                    add_row_to_csv(self.csv, self.now, user, group, 0, 0, "image", "image")
+                else:
+                    add_row_to_csv(self.csv, self.now, user, "", 0, 0, "image", "image")
+            return
+
+    def get_help_text(self, **kwargs):
+        help_text = f"后台统计用户请求与回复token数\n"
+        return help_text