diff --git a/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.IRGenerate.cs.old b/src/StgSharp.RegularAnalysis/.discard/RegexAnalyzer.IRGenerate.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.IRGenerate.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/RegexAnalyzer.IRGenerate.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.IROptimize.cs.old b/src/StgSharp.RegularAnalysis/.discard/RegexAnalyzer.IROptimize.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.IROptimize.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/RegexAnalyzer.IROptimize.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.PrefixEmit.cs.old b/src/StgSharp.RegularAnalysis/.discard/RegexAnalyzer.PrefixEmit.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.PrefixEmit.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/RegexAnalyzer.PrefixEmit.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexIR.Abstraction.cs.old b/src/StgSharp.RegularAnalysis/.discard/RegexIR.Abstraction.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/RegexIR.Abstraction.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/RegexIR.Abstraction.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexIR.Prefix.cs.old b/src/StgSharp.RegularAnalysis/.discard/RegexIR.Prefix.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/RegexIR.Prefix.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/RegexIR.Prefix.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexIR.cs.old b/src/StgSharp.RegularAnalysis/.discard/RegexIR.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/RegexIR.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/RegexIR.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexIRGenerator.cs.old b/src/StgSharp.RegularAnalysis/.discard/RegexIRGenerator.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/RegexIRGenerator.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/RegexIRGenerator.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/TextRegexSourceGen.GenClass.MultiIR.cs.old b/src/StgSharp.RegularAnalysis/.discard/TextRegexSourceGen.GenClass.MultiIR.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/TextRegexSourceGen.GenClass.MultiIR.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/TextRegexSourceGen.GenClass.MultiIR.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/TextRegexSourceGen.GenClass.Prefix.cs.old b/src/StgSharp.RegularAnalysis/.discard/TextRegexSourceGen.GenClass.Prefix.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/TextRegexSourceGen.GenClass.Prefix.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/TextRegexSourceGen.GenClass.Prefix.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Text/TextRegexSourceGen.GenClass.SingleIR.cs.old b/src/StgSharp.RegularAnalysis/.discard/TextRegexSourceGen.GenClass.SingleIR.cs.old
similarity index 100%
rename from src/StgSharp.RegularAnalysis/Text/TextRegexSourceGen.GenClass.SingleIR.cs.old
rename to src/StgSharp.RegularAnalysis/.discard/TextRegexSourceGen.GenClass.SingleIR.cs.old
diff --git a/src/StgSharp.RegularAnalysis/Abstraction/ISyntaxNode.cs b/src/StgSharp.RegularAnalysis/Abstraction/ISyntaxNode.cs
index 357d0013..36ea99fe 100644
--- a/src/StgSharp.RegularAnalysis/Abstraction/ISyntaxNode.cs
+++ b/src/StgSharp.RegularAnalysis/Abstraction/ISyntaxNode.cs
@@ -16,8 +16,9 @@ using System.Text;
 
 namespace StgSharp.RegularAnalysis.Abstraction
 {
-    public interface ITrinitySyntaxNode<TNode, TLabel> : ISyntaxNode<TNode, TLabel>
-        where TNode : ISyntaxNode<TNode, TLabel>
+    public interface ITrinitySyntaxNode<TNode, TPayload, TLabel> : ISyntaxNode<TNode, TPayload, TLabel>
+        where TNode : ISyntaxNode<TNode, TPayload, TLabel>
+        where TPayload : ISyntaxPayload<TPayload, TNode, TLabel>
         where TLabel : unmanaged
     {
 
@@ -25,7 +26,9 @@ namespace StgSharp.RegularAnalysis.Abstraction
 
     }
 
-    public interface ISyntaxNode<TNode, TLabel> where TNode : ISyntaxNode<TNode, TLabel>
+    public interface ISyntaxNode<TNode, TPayload, TLabel>
+        where TNode : ISyntaxNode<TNode, TPayload, TLabel>
+        where TPayload : ISyntaxPayload<TPayload, TNode, TLabel>
         where TLabel : unmanaged
     {
 
@@ -45,7 +48,7 @@ namespace StgSharp.RegularAnalysis.Abstraction
 
         static abstract TNode Empty { get; }
 
-        Token<TLabel> Source { get; }
+        TPayload Source { get; }
 
         protected internal int EnumState { get; set; }
 
diff --git a/src/StgSharp.RegularAnalysis/Abstraction/ISyntaxPayload.cs b/src/StgSharp.RegularAnalysis/Abstraction/ISyntaxPayload.cs
new file mode 100644
index 00000000..f6f1f25c
--- /dev/null
+++ b/src/StgSharp.RegularAnalysis/Abstraction/ISyntaxPayload.cs
@@ -0,0 +1,32 @@
+// -----------------------------------------------------------------------------
+// file="ISyntaxPayload"
+// Project: StgSharp
+// Copyright (c) Nitload.
+// SPDX-License-Identifier: MIT
+// -----------------------------------------------------------------------------
+
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Text;
+using System.Threading.Tasks;
+
+namespace StgSharp.RegularAnalysis.Abstraction
+{
+    // TODO 不对劲，TCarrier是个昏招，还是去掉吧
+    public interface ISyntaxPayload<TSelf, TCarrier, TLabel>
+        where TSelf : ISyntaxPayload<TSelf, TCarrier, TLabel>
+        where TCarrier : ISyntaxNode<TCarrier, TSelf, TLabel>
+        where TLabel : unmanaged
+    {
+
+        int Column { get; set; }
+
+        int Row { get; set; }
+
+        string Source { get; set; }
+
+        TLabel Flag { get; set; }
+
+    }
+}
diff --git a/src/StgSharp.RegularAnalysis/IR_RETIRED.md b/src/StgSharp.RegularAnalysis/IR_RETIRED.md
index 9f0f943b..b88bbe83 100644
--- a/src/StgSharp.RegularAnalysis/IR_RETIRED.md
+++ b/src/StgSharp.RegularAnalysis/IR_RETIRED.md
@@ -27,13 +27,37 @@ IR 层代码因此归档而非删除：将来做 stepping 解释器时是字节
 
 ## 需要在 AST 上补什么
 
+### 节点重设计（先于一切）
+
+`RegexAstNode` 不再挂载 `Token<RegexElementLabel>`：token 在 AST 构建时被**消费掉**，
+转换为结构化载荷。节点只剩三样东西：`Label`（原 `Source.Flag`）、`IPayload`、拓扑。
+这把 IR 层真正有价值的部分（类型化语义）请了回来，线性形态留在 `.old` 里。
+
+| Label | Payload | 转换时机 |
+| --- | --- | --- |
+| `UNIT` / `UNIT_SPAN` | `LiteralPayload(Text)`，**转义已消解** | AST 构建 |
+| `UNIT_SET` | `CharSetPayload(Set)` | AST 构建 |
+| `COUNT` | `CountPayload(Min, Max, IsGreedy)`，即 `{m,n}` 的唯一解析点 | AST 构建 |
+| `GROUP_BEGIN` | `GroupPayload(Name, Index)` | AST 构建 |
+| `CONCAT` / `ALT` | 无（null） | — |
+| `FIND`（新） | `FindPayload(Anchor, IsNearest)` | phase 6 融合 pass |
+
+规则：
+
+1. 载荷全部是不可变 record，深拷贝共享引用不复制
+2. `ReplaceBy` / `RemoveAndAsChild` 搬运 `Label + Payload`，不再拷 Source
+3. mask 决断载荷类型，取用走 `PayloadAs<T>()`，类型不符当场抛
+4. pass 局部属性（`FirstAtom`/`LastAtom`）不住这个槽，放融合 pass 的 side table
+
+改造点：`ASTGenerate` 构造处（token→payload 转换）、`ASTOptimize` 里读
+`Value`/`Source` 的消费点、两个变形 API、`EmptyRegexAstNode`（Label=NONE）。
+
 ### 基础件（不存在，先补）
 
-1. **子树深拷贝**——`(ab){2,3}` 展开需要克隆 body；配合破坏性移动可省一份
-2. **合成节点工厂**——展开要凭空造 CONCAT/COUNT 节点，
-   需要 `Token<RegexElementLabel>` 的合成构造入口（如 `Token.Synthetic(flag, value)`）
-3. **FIND 节点标签**——`RegexElementLabel` 新增 `FIND`，语义为"搜索连接"
-   （替代 IR 的 `FIND_COMPLEX`，见下）
+1. **子树深拷贝**——`(ab){2,3}` 展开需要克隆 body；配合破坏性移动可省一份；载荷共享不拷
+2. **FIND 节点标签**——`RegexElementLabel` 新增 `FIND`，语义为"搜索连接"
+   （替代 IR 的 `FIND_COMPLEX`，见下）；合成节点直接 `new(label, payload)`，
+   无需伪造 token
 
 ### 树上规范化 pass（`OptimizeTree` 扩到六阶段）
 
@@ -81,7 +105,9 @@ GROUP 的捕获终点落在锚起点**之前**。
 | 文件 | 错误原因 | 处置 |
 | --- | --- | --- |
 | `RegexAnalyzer.cs` | `TextRegexSource` 记录携带 `List<RegexIR>` 字段，`Analyze` 调 `GenerateIR`/`ScanAndOptimizeIR`/`EmitPrefix` | 砍掉 IR 字段与三个调用，管线止于 `OptimizeTree` |
-| `TextRegexSourceGen.GenClass.cs` | `GenerateSource` 读 `source.IRs`，`GenerateBlock`/`GenerateRegion`/`MeasureRegionExtents`/`FindRegionEnd` 全是 IR 机器 | 重写为读 `source.Ast` 的节点递归；保留 `FormatStringLiteral`/`FormatCharLiteral`/charset 解析 |
+| `TextRegexSourceGen.GenClass.cs` | `GenerateSource` 读 `source.IRs`，`GenerateBlock`/`GenerateRegion`/`MeasureRegionExtents`/`FindRegionEnd` 全是 IR 机器 | 重写为读 `source.Ast` 的节点递归；保留 `FormatStringLiteral`/`FormatCharLiteral`；charset 解析迁入 `CharSetPayload` 转换 |
+
+节点重设计落地时，`ASTGenerate`/`ASTOptimize` 会新增一批预期内错误（`Source`/`Value` 消费点），同批修复。
 
 ## `.old` 文件清单与参考价值
 
diff --git a/src/StgSharp.RegularAnalysis/StgSharp.RegularAnalysis.csproj b/src/StgSharp.RegularAnalysis/StgSharp.RegularAnalysis.csproj
index 0ab69b4c..87c61924 100644
--- a/src/StgSharp.RegularAnalysis/StgSharp.RegularAnalysis.csproj
+++ b/src/StgSharp.RegularAnalysis/StgSharp.RegularAnalysis.csproj
@@ -1,4 +1,4 @@
-<Project Sdk="Microsoft.NET.Sdk">
+<Project Sdk="Microsoft.NET.Sdk">
 
   <PropertyGroup>
     <IsRoslynComponent>true</IsRoslynComponent>
@@ -34,6 +34,16 @@
     <None Include="$(OutputPath)StgSharp.Common.dll" Pack="true" PackagePath="analyzers/dotnet/cs" Visible="false" Condition="Exists('$(OutputPath)StgSharp.Common.dll')" />
     <None Remove="bin\x64\Debug\net8.0\StgSharp.Common.pdb" />
     <None Include="$(OutputPath)StgSharp.Common.pdb" Pack="true" PackagePath="analyzers/dotnet/cs" Visible="false" Condition="Exists('$(OutputPath)StgSharp.Common.pdb')" />
+    <None Include=".discard\RegexAnalyzer.IRGenerate.cs.old" />
+    <None Include=".discard\RegexAnalyzer.IROptimize.cs.old" />
+    <None Include=".discard\RegexAnalyzer.PrefixEmit.cs.old" />
+    <None Include=".discard\RegexIR.Abstraction.cs.old" />
+    <None Include=".discard\RegexIR.cs.old" />
+    <None Include=".discard\RegexIR.Prefix.cs.old" />
+    <None Include=".discard\RegexIRGenerator.cs.old" />
+    <None Include=".discard\TextRegexSourceGen.GenClass.MultiIR.cs.old" />
+    <None Include=".discard\TextRegexSourceGen.GenClass.Prefix.cs.old" />
+    <None Include=".discard\TextRegexSourceGen.GenClass.SingleIR.cs.old" />
   </ItemGroup>
 
 </Project>
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.ASTNodeOptimize.cs b/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.ASTNodeOptimize.cs
new file mode 100644
index 00000000..1105fa9f
--- /dev/null
+++ b/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.ASTNodeOptimize.cs
@@ -0,0 +1,25 @@
+// -----------------------------------------------------------------------------
+// file="RegexAnalyzer.ASTNodeOptimize"
+// Project: StgSharp
+// Copyright (c) Nitload.
+// SPDX-License-Identifier: MIT
+// -----------------------------------------------------------------------------
+
+using StgSharp.RegularAnalysis.Abstraction;
+using System;
+
+namespace StgSharp.RegularAnalysis.Text
+{
+    internal partial class RegexAnalyzer
+    {
+
+        private static bool TryFlattenCount(
+                            RegexAstNode node
+        )
+        {
+            // flatten count node if it matches raw string pattern
+            throw new NotImplementedException();
+        }
+
+    }
+}
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.cs b/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.cs
index 826e791e..a48ea10f 100644
--- a/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.cs
+++ b/src/StgSharp.RegularAnalysis/Text/RegexAnalyzer.cs
@@ -44,10 +44,11 @@ namespace StgSharp.RegularAnalysis.Text
         public string Pattern { get; }
 
         internal AbstractSyntaxTree<RegexAstNode, RegexElementLabel> Ast { get; }
-
+        /*
         internal List<RegexIR> BaseIR { get; }
 
         internal List<RegexIR> IRs { get; }
+        /**/
 
         internal RegexInfo Info { get; }
 
@@ -76,10 +77,11 @@ namespace StgSharp.RegularAnalysis.Text
             {
                 return new TextRegexSource(regex, null, null, null, info, ex);
             }
-            List<RegexIR> base_ir;
+            /*
+            // List<RegexIR> base_ir;
             try
             {
-                base_ir = GenerateIR(tree, ref info);
+                // base_ir = GenerateIR(tree, ref info);
             }
             catch (Exception ex)
             {
@@ -87,8 +89,8 @@ namespace StgSharp.RegularAnalysis.Text
             }
             try
             {
-                ScanAndOptimizeIR(base_ir, out List<RegexIR> opt_ir);
-                List<RegexIR> prefix_ir = EmitPrefix(opt_ir, ref info);
+                // ScanAndOptimizeIR(base_ir, out List<RegexIR> opt_ir);
+                // List<RegexIR> prefix_ir = EmitPrefix(opt_ir, ref info);
                 return new TextRegexSource(regex, require_ast ? tree : null!,
                                            require_origin_ir ? base_ir : null!, prefix_ir, info);
             }
@@ -97,6 +99,7 @@ namespace StgSharp.RegularAnalysis.Text
                 return new TextRegexSource(regex, require_ast ? tree : null!,
                                            require_origin_ir ? base_ir : null!, null!, info, ex);
             }
+            /**/
 #pragma warning restore CA1031 
         }
 
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexAnanlyzer.ASTNodeOptimize.cs b/src/StgSharp.RegularAnalysis/Text/RegexAnanlyzer.ASTNodeOptimize.cs
new file mode 100644
index 00000000..e69de29b
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexAstNode.AdditionalPayload.cs b/src/StgSharp.RegularAnalysis/Text/RegexAstNode.AdditionalPayload.cs
new file mode 100644
index 00000000..12301acd
--- /dev/null
+++ b/src/StgSharp.RegularAnalysis/Text/RegexAstNode.AdditionalPayload.cs
@@ -0,0 +1,34 @@
+// -----------------------------------------------------------------------------
+// file="RegexAstNode.AdditionalPayload"
+// Project: StgSharp
+// Copyright (c) Nitload.
+// SPDX-License-Identifier: MIT
+// -----------------------------------------------------------------------------
+
+using StgSharp.RegularAnalysis.Abstraction;
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Text;
+using System.Threading.Tasks;
+
+namespace StgSharp.RegularAnalysis.Text
+{
+    public class RegexCountPayload : ISyntaxPayload<RegexCountPayload, RegexAstNode, RegexElementLabel>
+    {
+
+        public int Min { get; set; }
+
+        public int Max { get; set; }
+
+        public string Source { get; set; }
+
+    }
+
+    public class RegexCommonPayload : IRegexAstNodeAdditionalPayload
+    {
+
+        public string Source { get; set; } = string.Empty;
+
+    }
+}
diff --git a/src/StgSharp.RegularAnalysis/Text/RegexAstNode.cs b/src/StgSharp.RegularAnalysis/Text/RegexAstNode.cs
index 1ca9d4ae..80af460f 100644
--- a/src/StgSharp.RegularAnalysis/Text/RegexAstNode.cs
+++ b/src/StgSharp.RegularAnalysis/Text/RegexAstNode.cs
@@ -10,6 +10,8 @@ using System.Runtime.CompilerServices;
 
 namespace StgSharp.RegularAnalysis.Text
 {
+    // TODO 现在存在的问题是，AstNode过于依赖NodeFlag表达其内部语义，这导致进行精细优化时识别节点语义复杂，
+    // TODO 当前Ast的类型差异过小，不利于保存特征数据，应当争取进一步细化，或者说这是AstNode一开始为了简化架构设计引入的妥协
     internal class RegexAstNode : ISyntaxNode<RegexAstNode, RegexElementLabel>
     {