diff --git a/lib/ai/mcpcore/README2.md b/lib/ai/mcpcore/README2.md
deleted file mode 100644
index 7fa38724..00000000
--- a/lib/ai/mcpcore/README2.md
+++ /dev/null
@@ -1,3 +0,0 @@
-
-
-If logic is implemented in mcp module, than structure with folders logic and mcp, where logic residers in /logic and mcp related code (like tool and prompt handlers and server code) in /mcp
\ No newline at end of file
diff --git a/lib/core/code/improvements.md b/lib/core/code/improvements.md
new file mode 100644
index 00000000..f9d3f44b
--- /dev/null
+++ b/lib/core/code/improvements.md
@@ -0,0 +1,228 @@
+# Code Review and Improvement Plan for HeroLib Code Module
+
+## Overview
+The HeroLib `code` module provides utilities for parsing and generating V language code. It's designed to be a lightweight alternative to `v.ast` for code analysis and generation across multiple languages. While the module has good foundational structure, there are several areas that need improvement.
+
+## Issues Identified
+
+### 1. Incomplete TypeScript Generation Support
+- The `typescript()` method exists in some models but lacks comprehensive implementation
+- Missing TypeScript generation for complex types (arrays, maps, results)
+- No TypeScript interface generation for structs
+
+### 2. Template System Issues
+- Some templates are empty (e.g., `templates/function/method.py`, `templates/comment/comment.py`)
+- Template usage is inconsistent across the codebase
+- No clear separation between V and other language templates
+
+### 3. Missing Parser Documentation Examples
+- README.md mentions codeparser but doesn't show how to use the parser from this module
+- No clear examples of parsing V files or modules
+
+### 4. Incomplete Type Handling
+- The `parse_type` function doesn't handle all V language types comprehensively
+- Missing support for function types, sum types, and complex generics
+- No handling of optional types (`?Type`)
+
+### 5. Code Structure and Consistency
+- Some functions lack proper error handling
+- Inconsistent naming conventions in test files
+- Missing documentation for several key functions
+
+## Improvement Plan
+
+### 1. Complete TypeScript Generation Implementation
+
+**What needs to be done:**
+- Implement comprehensive TypeScript generation in `model_types.v`
+- Add TypeScript generation for all type variants
+- Create proper TypeScript interface generation in `model_struct.v`
+
+**Specific fixes:**
+```v
+// In model_types.v, improve the typescript() method:
+pub fn (t Type) typescript() string {
+  return match t {
+    Map { 'Record<string, ${t.typ.typescript()}>' }
+    Array { '${t.typ.typescript()}[]' }
+    Object { t.name }
+    Result { 'Promise<${t.typ.typescript()}>' } // Better representation for async operations
+    Boolean { 'boolean' }
+    Integer { 'number' }
+    Alias { t.name }
+    String { 'string' }
+    Function { '(...args: any[]) => any' } // More appropriate for function types
+    Void { 'void' }
+  }
+}
+
+// In model_struct.v, improve the typescript() method:
+pub fn (s Struct) typescript() string {
+  name := texttools.pascal_case(s.name)
+  fields := s.fields.map(it.typescript()).join('\n  ')
+  return 'export interface ${name} {\n  ${fields}\n}'
+}
+```
+
+### 2. Fix Template System
+
+**What needs to be done:**
+- Remove empty Python template files
+- Ensure all templates are properly implemented
+- Add template support for other languages
+
+**Specific fixes:**
+- Delete `templates/function/method.py` and `templates/comment/comment.py` if they're not needed
+- Add proper TypeScript templates for struct and interface generation
+- Create consistent template naming conventions
+
+### 3. Improve Parser Documentation
+
+**What needs to be done:**
+- Add clear examples in README.md showing how to use the parser
+- Document the parsing functions with practical examples
+
+**Specific fixes:**
+Add to README.md:
+```markdown
+## Parsing V Code
+
+The code module provides utilities to parse V code into structured models:
+
+```v
+import freeflowuniverse.herolib.core.code
+
+// Parse a V file
+content := os.read_file('example.v') or { panic(err) }
+vfile := code.parse_vfile(content) or { panic(err) }
+
+// Access parsed information
+println('Module: ${vfile.mod}')
+println('Number of functions: ${vfile.functions().len}')
+println('Number of structs: ${vfile.structs().len}')
+
+// Parse individual components
+function := code.parse_function(fn_code_string) or { panic(err) }
+struct_ := code.parse_struct(struct_code_string) or { panic(err) }
+```
+
+### 4. Complete Type Handling
+
+**What needs to be done:**
+- Extend `parse_type` to handle more complex V types
+- Add support for optional types (`?Type`)
+- Improve generic type parsing
+
+**Specific fixes:**
+```v
+// In model_types.v, enhance parse_type function:
+pub fn parse_type(type_str string) Type {
+  mut type_str_trimmed := type_str.trim_space()
+  
+  // Handle optional types
+  if type_str_trimmed.starts_with('?') {
+    return Optional{parse_type(type_str_trimmed.all_after('?'))}
+  }
+  
+  // Handle function types
+  if type_str_trimmed.starts_with('fn ') {
+    // Parse function signature
+    return Function{}
+  }
+  
+  // Handle sum types
+  if type_str_trimmed.contains('|') {
+    types := type_str_trimmed.split('|').map(parse_type(it.trim_space()))
+    return Sum{types}
+  }
+  
+  // Existing parsing logic...
+}
+```
+
+### 5. Code Structure Improvements
+
+**What needs to be done:**
+- Add proper error handling to all parsing functions
+- Standardize naming conventions
+- Improve documentation consistency
+
+**Specific fixes:**
+- Add error checking in `parse_function`, `parse_struct`, and other parsing functions
+- Ensure all public functions have clear documentation comments
+- Standardize test function names
+
+## Module Generation to Other Languages
+
+### Current Implementation
+The current code shows basic TypeScript generation support, but it's incomplete. The generation should:
+
+1. **Support multiple languages**: The code structure allows for multi-language generation, but only TypeScript has partial implementation
+2. **Use templates consistently**: All language generation should use the template system
+3. **Separate language-specific code**: Each language should have its own generation module
+
+### What Needs to Move to Other Modules
+
+**TypeScript Generation Module:**
+- Move all TypeScript-specific generation code to a new `typescript` module
+- Create TypeScript templates for structs, interfaces, and functions
+- Add proper TypeScript formatting support
+
+**Example Structure:**
+```
+lib/core/code/
+├── model_types.v          # Core type models (language agnostic)
+├── model_struct.v         # Core struct/function models (language agnostic)
+└── typescript/            # TypeScript-specific generation
+    ├── generator.v        # TypeScript generation logic
+    └── templates/         # TypeScript templates
+```
+
+### Parser Usage Examples (to add to README.md)
+
+```v
+// Parse a V file into a structured representation
+content := os.read_file('mymodule/example.v') or { panic(err) }
+vfile := code.parse_vfile(content)!
+
+// Extract all functions
+functions := vfile.functions()
+println('Found ${functions.len} functions')
+
+// Extract all structs
+structs := vfile.structs()
+for s in structs {
+  println('Struct: ${s.name}')
+  for field in s.fields {
+    println('  Field: ${field.name} (${field.typ.symbol()})')
+  }
+}
+
+// Find a specific function
+if greet_fn := vfile.get_function('greet') {
+  println('Found function: ${greet_fn.name}')
+  println('Parameters: ${greet_fn.params.map(it.name)}')
+  println('Returns: ${greet_fn.result.typ.symbol()}')
+}
+
+// Parse a function from string
+fn_code := '
+pub fn add(a int, b int) int {
+  return a + b
+}
+'
+function := code.parse_function(fn_code)!
+println('Parsed function: ${function.name}')
+```
+
+## Summary of Required Actions
+
+1. **Implement complete TypeScript generation** across all model types
+2. **Remove empty template files** and organize templates properly
+3. **Enhance type parsing** to handle optional types, function types, and sum types
+4. **Add comprehensive parser documentation** with practical examples to README.md
+5. **Create language-specific generation modules** to separate concerns
+6. **Improve error handling** in all parsing functions
+7. **Standardize documentation and naming** conventions across the module
+
+These improvements will make the code module more robust, easier to use, and better prepared for multi-language code generation.
\ No newline at end of file