diff --git a/app/models/__init__.py b/app/models/__init__.py
index 545fc6f..45d7ee4 100644
--- a/app/models/__init__.py
+++ b/app/models/__init__.py
@@ -5,7 +5,6 @@ from .follow import Follow
 from .workspace import Workspace, WorkspaceMember
 from .project import ProjectNode
 from .library import LibraryItem
-# 引入新交互表
 from .interaction import ProjectLike, ProjectFavorite, ProjectShare
 
 __all__ = [
diff --git a/app/models/base.py b/app/models/base.py
index eb5c8e1..9113d4a 100644
--- a/app/models/base.py
+++ b/app/models/base.py
@@ -9,16 +9,22 @@ Base = declarative_base()
 ALPHABET = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
 
 
-def encode_base62(num: int) -> str:
-    """将一个整数编码为 Base62 字符串"""
+def encode_base62(num: int, length: int = None) -> str:
+    """将一个整数编码为 Base62 字符串，支持左侧补 0 固定长度"""
     if num == 0:
-        return ALPHABET[0]
-    arr = []
-    base = len(ALPHABET)
-    while num:
-        num, rem = divmod(num, base)
-        arr.append(ALPHABET[rem])
-    return "".join(reversed(arr))
+        res = ALPHABET[0]
+    else:
+        arr = []
+        base = len(ALPHABET)
+        while num:
+            num, rem = divmod(num, base)
+            arr.append(ALPHABET[rem])
+        res = "".join(reversed(arr))
+    
+    if length:
+        # 如果太长就截取，太短就用第一个字符(0)补齐
+        return res.rjust(length, ALPHABET[0])[-length:]
+    return res
 
 
 def encode_tablename_to_prefix(tablename: str, length: int = 3) -> str:
@@ -33,7 +39,7 @@ def encode_tablename_to_prefix(tablename: str, length: int = 3) -> str:
     num = int(hex_digest, 16)
     
     # 转为 Base62 字符串
-    b62_str = encode_base62(num)
+    b62_str = encode_base62(num, length=length)
     
     # 截取固定长度（例如 3 位）作为该表的专属代码
     return b62_str[:length]
@@ -49,17 +55,31 @@ class AutoID:
     """
     全自动 Convex 风格 ID 注入器
     不再需要手动定义任何 prefix，全自动根据 __tablename__ 编码生成
+    支持子类自定义长度
     """
+    # 默认配置（如果子类不定义，则使用这些默认值）
+    _id_prefix_len = 3   # 前缀长度
+    _id_payload_len = 22 # 主体长度（ULID 转 Base62 最大约 22 位）
+    
     @declared_attr
     def id(cls):
         # 运行时动态获取子类的 __tablename__
         tablename = cls.__tablename__
+        # 获取子类配置的长度
+        p_len = cls._id_prefix_len
+        s_len = cls._id_payload_len
         
         # 将表名编码为固定前缀（例如 "users" -> "7X1"）
-        table_prefix = encode_tablename_to_prefix(tablename, length=3)
+        hasher = hashlib.md5(tablename.encode("utf-8"))
+        table_num = int(hasher.hexdigest(), 16)
+        table_prefix = encode_base62(table_num, length=p_len)
         
         # 核心工厂函数：拼接【编码后的表前缀】和【递增Payload】
         def id_factory():
-            return f"{table_prefix}_{gen_convex_payload()}"
+            # 生成固定长度的 Payload
+            u = ulid.new()
+            payload = encode_base62(u.int, length=s_len)
+            # 直接拼接，不带下划线
+            return f"{table_prefix}{payload}"
 
-        return Column(String(32), primary_key=True, default=id_factory)
\ No newline at end of file
+        return Column(String(p_len + s_len), primary_key=True, default=id_factory)
\ No newline at end of file
diff --git a/app/models/user.py b/app/models/user.py
index 91b6490..8c5578d 100644
--- a/app/models/user.py
+++ b/app/models/user.py
@@ -5,6 +5,10 @@ from .base import AutoID, Base
 
 class User(Base, AutoID):
     __tablename__ = "users"
+    
+    # 自定义配置
+    # _id_prefix_len = 4
+    # _id_payload_len = 16
 
     google_id = Column(String, unique=True, index=True)
     email = Column(String, unique=True, index=True)