diff --git a/back/__pycache__/main.cpython-314.pyc b/back/__pycache__/main.cpython-314.pyc
new file mode 100644
index 0000000..eebc52e
Binary files /dev/null and b/back/__pycache__/main.cpython-314.pyc differ
diff --git a/back/main.py b/back/main.py
index 5df142f..a8bcb3e 100644
--- a/back/main.py
+++ b/back/main.py
@@ -44,7 +44,6 @@ class Submission(db.Model):
     __tablename__ = 'submissions'
     id = db.Column(db.Integer, primary_key=True, autoincrement=True)
     content = db.Column(db.Text, nullable=False)
-    hashtopic = db.Column(db.Text) # JSON string
     identity_token = db.Column(db.String(36), nullable=True)
     status = db.Column(db.String(20), default='Pending')
     created_at = db.Column(db.DateTime, default=lambda: datetime.now())
@@ -58,13 +57,19 @@ class Comment(db.Model):
     id = db.Column(db.Integer, primary_key=True, autoincrement=True)
     submission_id = db.Column(db.Integer, db.ForeignKey('submissions.id'), nullable=False)
     content = db.Column(db.Text, nullable=False)
-    hashtopic = db.Column(db.Text) # JSON string
     identity_token = db.Column(db.String(36), nullable=True)
     created_at = db.Column(db.DateTime, default=lambda: datetime.now())
     upvotes = db.Column(db.Integer, default=0)
     downvotes = db.Column(db.Integer, default=0)
     parent_comment_id = db.Column(db.Integer, db.ForeignKey('comments.id'), nullable=True)
 
+class Hashtag(db.Model):
+    __tablename__ = 'hashtags'
+    id = db.Column(db.Integer, primary_key=True, autoincrement=True)
+    type = db.Column(db.Integer, nullable=False) # 0: Submission, 1: Comment
+    target_id = db.Column(db.Integer, nullable=False)
+    name = db.Column(db.String(255), nullable=False)
+
 class DenyWord(db.Model):
     __tablename__ = 'deny_words'
     id = db.Column(db.Integer, primary_key=True, autoincrement=True)
@@ -88,6 +93,17 @@ def load_config():
         except Exception as e:
             print(f"Warning: Failed to load settings: {e}")
 
+def load_deny_words():
+    global DENY_WORDS_CACHE
+    with app.app_context():
+        try:
+            words = db.session.query(DenyWord.word).all()
+            # words 是 list of tuples [('word1',), ('word2',)]
+            DENY_WORDS_CACHE = [w[0] for w in words]
+            print(f"Loaded {len(DENY_WORDS_CACHE)} deny words.")
+        except Exception as e:
+            print(f"Warning: Failed to load deny words: {e}")
+
 # --- 用户普通api端点 ---
 @app.route('/api/settings', methods=['GET'])
 def get_settings():
@@ -185,10 +201,9 @@ def submit_post():
 
         identity_token = data.get('identity')
         
-        # 违禁词检测
-        deny_words = DenyWord.query.all()
-        for dw in deny_words:
-            if dw.word in content:
+        # 违禁词检测 (使用内存缓存)
+        for word in DENY_WORDS_CACHE:
+            if word in content:
                 return jsonify({"code": 2005, "data": "提交内容包含违禁词"})
 
         # Identity 验证
@@ -201,13 +216,23 @@ def submit_post():
         # 保存
         new_post = Submission(
             content=content,
-            hashtopic=json.dumps(hashtopic) if hashtopic else '[]',
             identity_token=identity_token,
             status='Pending' if NEED_AUDIT else 'Pass'
         )
         db.session.add(new_post)
         db.session.commit()
         
+        # 保存 Hashtags
+        if hashtopic:
+            for tag in hashtopic:
+                new_tag = Hashtag(
+                    type=0, # 0 for Submission
+                    target_id=new_post.id,
+                    name=tag
+                )
+                db.session.add(new_tag)
+            db.session.commit()
+        
         code = 1002 if new_post.status == 'Pending' else 1001
         return jsonify({"code": code, "data": {"id": new_post.id}})
     except Exception as e:
@@ -243,12 +268,22 @@ def submit_comment():
         new_comment = Comment(
             submission_id=submission_id,
             content=content,
-            hashtopic=json.dumps(hashtopic) if hashtopic else '[]',
             identity_token=identity_token
         )
         db.session.add(new_comment)
         db.session.commit()
         
+        # 保存 Hashtags
+        if hashtopic:
+            for tag in hashtopic:
+                new_tag = Hashtag(
+                    type=1, # 1 for Comment
+                    target_id=new_comment.id,
+                    name=tag
+                )
+                db.session.add(new_tag)
+            db.session.commit()
+        
         return jsonify({"code": 1000, "data": ""})
     except Exception as e:
         return jsonify({"code": 2003, "data": f"评论失败: {str(e)}"})
@@ -337,4 +372,5 @@ def get_statics():
 if __name__ == '__main__':
     init_db()
     load_config()
+    load_deny_words()
     app.run(debug=True, port=5000)
diff --git a/front/src/components/CreatePost.tsx b/front/src/components/CreatePost.tsx
index 63d72ee..caa860b 100644
--- a/front/src/components/CreatePost.tsx
+++ b/front/src/components/CreatePost.tsx
@@ -1,6 +1,6 @@
 // 我草，react-md-editor这么好用，无语了，早知道v1也用这个编辑器了.....
 // 不过居然没有汉化...有点可惜
-// 我撤回刚才那句话，编辑器有提供中文指令集，我是sb。
+// 我撤回刚才那句话，编辑器有提供中文指令集，我是sb，没发现。
 import React, { useState, useEffect, useRef, useCallback } from 'react';
 import MDEditor from '@uiw/react-md-editor';
 // 引入中文指令集
diff --git a/front/src/components/StatusDisplay.tsx b/front/src/components/StatusDisplay.tsx
index 00d04e5..6b20342 100644
--- a/front/src/components/StatusDisplay.tsx
+++ b/front/src/components/StatusDisplay.tsx
@@ -1,3 +1,4 @@
+// v1的时候就连统计信息也5秒获取一次，不妥，所以改成逻辑触发刷新
 import React, { useState, useEffect, useCallback } from 'react';
 import {
   makeStyles,