mcp-server-code-extractor 0.2.3

A Model Context Protocol (MCP) server that provides precise code extraction tools using tree-sitter parsing

MCP Server Code Extractor

A Model Context Protocol (MCP) server that provides precise code extraction tools using tree-sitter parsing. Extract functions, classes, and code snippets from 30+ programming languages without manual parsing.

Why MCP Server Code Extractor?

When working with AI coding assistants like Claude, you often need to:

• Extract specific functions or classes from large codebases
• Get an overview of what's in a file without reading the entire thing
• Retrieve precise code snippets with accurate line numbers
• Avoid manual parsing and grep/sed/awk gymnastics

MCP Server Code Extractor solves these problems by providing structured, tree-sitter-powered code extraction tools directly within your AI assistant.

Features

• 🎯 Precise Extraction: Uses tree-sitter parsing for accurate code boundary detection
• 🌍 30+ Languages: Supports Python, JavaScript, TypeScript, Go, Rust, Java, C/C++, and many more
• 📍 Line Numbers: Every extraction includes precise line number information
• 🔍 Code Discovery: List all functions and classes in a file before extracting
• 🌐 URL Support: Fetch and extract code from GitHub, GitLab, and direct file URLs
• 🔄 Git Integration: Extract code from any git revision, branch, or tag
• ⚡ Fast & Lightweight: Efficient caching and minimal dependencies
• 🤖 AI-Optimized: Designed specifically for use with AI coding assistants

Installation

Quick Start with UV (Recommended)

# Install UV if you haven't already
curl -LsSf https://astral.sh/uv/install.sh | sh

# Clone this repository
git clone https://github.com/ctoth/mcp_server_code_extractor
cd mcp_server_code_extractor

# Run directly with UV (no installation needed!)
uv run mcp_server_code_extractor.py

Traditional Installation

pip install mcp[cli] tree-sitter-languages tree-sitter==0.21.3

Configure with Claude Desktop

Add to your Claude Desktop configuration:

{
  "mcpServers": {
    "mcp-server-code-extractor": {
      "command": "uv",
      "args": ["run", "/path/to/mcp_server_code_extractor.py"]
    }
  }
}

Or with traditional Python:

{
  "mcpServers": {
    "mcp-server-code-extractor": {
      "command": "python",
      "args": ["/path/to/mcp_server_code_extractor.py"]
    }
  }
}

Available Tools

1. get_symbols - Discover Code Structure

List all functions, classes, and other symbols in a file.

Returns:
- name: Symbol name
- type: function/class/method/etc
- start_line/end_line: Line numbers
- preview: First line of the symbol

2. get_function - Extract Complete Functions

Extract a complete function with all its code.

Parameters:
- file_path: Path to the source file
- function_name: Name of the function to extract

Returns:
- code: Complete function code
- start_line/end_line: Precise boundaries
- language: Detected language

3. get_class - Extract Complete Classes

Extract an entire class definition including all methods.

Parameters:
- file_path: Path to the source file
- class_name: Name of the class to extract

Returns:
- code: Complete class code
- start_line/end_line: Precise boundaries
- language: Detected language

4. get_lines - Extract Specific Line Ranges

Get exact line ranges when you know the line numbers.

Parameters:
- file_path: Path to the source file
- start_line: Starting line (1-based)
- end_line: Ending line (inclusive)

Returns:
- code: Extracted lines
- line numbers and metadata

5. get_signature - Get Function Signatures

Quickly get just the function signature without the body.

Parameters:
- file_path: Path to the source file
- function_name: Name of the function

Returns:
- signature: Function signature only
- start_line: Where the function starts

Usage Examples

Example 1: Exploring Local Files

# First, see what's in the file
symbols = get_symbols("src/main.py")
# Returns: List of all functions and classes with line numbers

# Extract a specific function
result = get_function("src/main.py", "process_data")
# Returns: Complete function code with line numbers

# Get just a function signature
sig = get_signature("src/main.py", "process_data")
# Returns: "def process_data(input_file: str, output_dir: Path) -> Dict[str, Any]:"

Example 2: Working with URLs

# Explore a GitHub file
symbols = get_symbols("https://raw.githubusercontent.com/user/repo/main/src/api.py")

# Extract function from GitLab
result = get_function("https://gitlab.com/user/project/-/raw/main/utils.py", "helper_func")

# Get lines from any URL
lines = get_lines("https://example.com/code/script.py", 10, 25)

Example 3: Working with Classes

# Extract an entire class (local file)
result = get_class("models/user.py", "User")
# Returns: Complete User class with all methods

# Extract class from URL
result = get_class("https://raw.githubusercontent.com/user/repo/main/models.py", "DatabaseModel")

# Get specific lines (e.g., just the __init__ method)
lines = get_lines("models/user.py", 10, 25)
# Returns: Lines 10-25 of the file

Example 4: Multi-Language Support

// Works with JavaScript/TypeScript
symbols = get_symbols("app.ts")
func = get_function("app.ts", "handleRequest")

// Works with Go
symbols = get_symbols("main.go")
method = get_function("main.go", "ServeHTTP")

Supported Languages

• Python, JavaScript, TypeScript, JSX/TSX
• Go, Rust, C, C++, C#, Java
• Ruby, PHP, Swift, Kotlin, Scala
• Bash, PowerShell, SQL
• Haskell, OCaml, Elixir, Clojure
• And many more...

Best Practices

1. Always use get_symbols first when exploring a new file
2. Use get_function/get_class instead of reading entire files
3. Use get_lines when you know exact line numbers
4. Use get_signature for quick API exploration

Why Not Just Use Read?

Traditional file reading tools require you to:

• Read entire files (inefficient for large files)
• Manually parse code to find functions/classes
• Count lines manually for extraction
• Deal with complex syntax and edge cases

MCP Server Code Extractor:

• ✅ Extracts exactly what you need
• ✅ Provides structured data with metadata
• ✅ Handles complex syntax automatically
• ✅ Works across 30+ languages consistently

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT License - see LICENSE file for details.

Acknowledgments

• Built on tree-sitter for robust parsing
• Uses tree-sitter-languages for language support
• Implements the Model Context Protocol specification