repomix

📦 Repomix (Python Version)

English | 简体中文

🎯 1. Introduction

Repomix is a powerful tool that packs your entire repository into a single, AI-friendly file. It's perfect for when you need to feed your codebase to Large Language Models (LLMs) or other AI tools like Claude, ChatGPT, and Gemini.

The original Repomix is written in JavaScript, and this is the ported Python version.

⭐ 2. Features

• AI-Optimized: Formats your codebase in a way that's easy for AI to understand and process.
• Token Counting: Provides token counts for each file and the entire repository using tiktoken.
• Simple to Use: Pack your entire repository with just one command.
• Customizable: Easily configure what to include or exclude.
• Git-Aware: Automatically respects your .gitignore files.
• Security-Focused: Built-in security checks to detect and prevent the inclusion of sensitive information (powered by detect-secrets).
• Code Compression: Advanced code compression with multiple modes to reduce output size while preserving essential information.
• ⚡ Performance: Utilizes multiprocessing or threading for faster analysis on multi-core systems.
• ⚙️ Encoding Aware: Automatically detects and handles various file encodings (using chardet) beyond UTF-8, increasing robustness.

🚀 3. Quick Start

Quick Run with uvx (Recommended)

The easiest way to use Repomix is with uvx - no installation required:

uvx repomix

That's it! This will pack your current directory into an AI-friendly file.

More examples:

# Pack with JSON output
uvx repomix --style json

# Pack a remote repository
uvx repomix --remote https://github.com/username/repo

# Pack with specific patterns
uvx repomix --include "src/**/*.py" --ignore "tests/**"

# Use a specific version
uvx repomix@0.4.1

You can also use pipx: pipx run repomix

Install with pip

For frequent usage, you can install Repomix globally:

pip install repomix

Then run in any project directory:

repomix

Docker Usage

You can also use Repomix with Docker without installing it locally:

# Build the Docker image
docker build -t repomix .

# Run repomix on the current directory
docker run --rm -v "$(pwd)":/app repomix

# Run repomix with specific options
docker run --rm -v "$(pwd)":/app repomix --style markdown --output custom-output.md

# Run repomix on a different directory
docker run --rm -v "/path/to/your/project":/app repomix

Docker Benefits:

• Isolated Environment: Run repomix without installing Python dependencies on your host system
• Consistent Results: Ensures the same environment across different machines
• Easy Distribution: Share the exact repomix version and configuration with your team
• No Installation Required: Use repomix immediately without pip install

That's it! Repomix will generate a repomix-output.md file (by default) in your current directory, containing your entire repository in an AI-friendly format.

📖 4. Usage

4.1 Command Line Usage

To pack your entire repository:

repomix

To pack a specific directory:

repomix path/to/directory

To pack a remote repository:

repomix --remote https://github.com/username/repo

To pack a specific branch of a remote repository:

repomix --remote https://github.com/username/repo --branch feature-branch

To initialize a new configuration file:

repomix --init
# Use --global to create a global configuration file (see Configuration Options below)
repomix --init --global

4.2 Configuration Options

Create a repomix.config.json file in your project root for custom configurations. Repomix also automatically loads a global configuration file if it exists (e.g., ~/.config/repomix/repomix.config.json on Linux), merging it with lower priority than local config and CLI options.

{
  "output": {
    "file_path": "repomix-output.md",
    "style": "markdown",
    "header_text": "",
    "instruction_file_path": "",
    "remove_comments": false,
    "remove_empty_lines": false,
    "top_files_length": 5,
    "show_line_numbers": false,
    "copy_to_clipboard": false,
    "include_empty_directories": false,
    "calculate_tokens": false,
    "show_file_stats": false,
    "show_directory_structure": true,
    "include_full_directory_structure": false,
    "split_output": null,
    "token_count_tree": false,
    "git": {
      "sort_by_changes": true,
      "sort_by_changes_max_commits": 100,
      "include_diffs": false,
      "include_logs": false,
      "include_logs_count": 50
    }
  },
  "security": {
    "enable_security_check": true,
    "exclude_suspicious_files": true
  },
  "ignore": {
    "custom_patterns": [],
    "use_gitignore": true,
    "use_default_ignore": true
  },
  "compression": {
    "enabled": false,
    "keep_signatures": true,
    "keep_docstrings": true,
    "keep_interfaces": true
  },
  "remote": {
    "url": "",
    "branch": ""
  },
  "include": []
}

[!NOTE] Note on remove_comments: This feature is language-aware, correctly handling comment syntax for various languages like Python, JavaScript, C++, HTML, etc., rather than using a simple generic pattern.*

Remote Repository Configuration

The remote section allows you to configure remote repository processing:

• url: The URL of the remote Git repository to process
• branch: The specific branch, tag, or commit hash to process (optional, defaults to repository's default branch)

When a remote URL is specified in the configuration, Repomix will process the remote repository instead of the local directory. This can be overridden by CLI parameters.

Command Line Options

• repomix [directory]: Target directory (defaults to current directory).
• -v, --version: Show version.
• -o, --output <file>: Specify output file name.
• --style <style>: Specify output style (plain, xml, markdown, json).
• --remote <url>: Process a remote Git repository.
• --branch <name>: Specify branch for remote repository.
• --init: Initialize configuration file (repomix.config.json) in the current directory.
• --global: Use with --init to create/manage the global configuration file (located in a platform-specific user config directory, e.g., ~/.config/repomix on Linux). The global config is automatically loaded if present.
• --no-security-check: Disable security check.
• --include <patterns>: Comma-separated list of include patterns (glob format).
• -i, --ignore <patterns>: Additional comma-separated ignore patterns.
• -c, --config <path>: Path to a custom configuration file.
• --copy: Copy generated output to system clipboard.
• --top-files-len <number>: Max number of largest files to display in summary.
• --output-show-line-numbers: Add line numbers to output code blocks.
• --stdin: Read file paths from standard input (one per line) instead of discovering files automatically.
• --verbose: Enable verbose logging for debugging.
• --parsable-style: By escaping and formatting, ensure the output is parsable as a document of its type.
• --stdout: Output to stdout instead of writing to a file.
• --remove-comments: Remove comments from source code.
• --remove-empty-lines: Remove empty lines from source code.
• --truncate-base64: Enable truncation of base64 data strings.
• --include-empty-directories: Include empty directories in the output.
• --include-diffs: Include git diffs in the output.
• --include-logs: Include git log history in the output.
• --sort-by-changes: Sort files by git change frequency (most changed first).

4.3 Security Check

Repomix includes built-in security checks using the detect-secrets library to detect potentially sensitive information (API keys, credentials, etc.). By default (exclude_suspicious_files: true), detected files are excluded from the output.

Disable checks via configuration or CLI:

repomix --no-security-check

4.4 Code Compression

Repomix provides advanced code compression capabilities to reduce output size while preserving essential information. This feature is particularly useful when working with large codebases or when you need to focus on specific aspects of your code.

4.4.1 Compression Modes

Interface Mode (keep_interfaces: true)

• Preserves function and class signatures with their complete type annotations
• Keeps all docstrings for comprehensive API documentation
• Removes implementation details, replacing them with pass statements
• Perfect for generating API documentation or understanding code structure

Signature Mode (keep_signatures: true, keep_interfaces: false)

• Preserves function and class definitions
• Optionally keeps docstrings based on keep_docstrings setting
• Maintains full implementation code
• Useful for standard code compression while keeping functionality

Minimal Mode (keep_signatures: false)

• Removes all function and class definitions
• Keeps only global variables, imports, and module-level code
• Maximum compression for focusing on configuration and constants

4.4.2 Configuration Options

{
  "compression": {
    "enabled": false,           // Enable/disable compression
    "keep_signatures": true,    // Keep function/class signatures
    "keep_docstrings": true,    // Keep docstrings
    "keep_interfaces": true     // Interface mode (signatures + docstrings only)
  }
}

4.4.3 Usage Examples

Generate API Documentation:

# Create interface-only output for API documentation
repomix --config-override '{"compression": {"enabled": true, "keep_interfaces": true}}'

Compress Implementation Details:

# Keep signatures but remove implementation for code overview
repomix --config-override '{"compression": {"enabled": true, "keep_interfaces": false, "keep_signatures": true, "keep_docstrings": false}}'

Extract Configuration Only:

# Keep only global variables and constants
repomix --config-override '{"compression": {"enabled": true, "keep_signatures": false}}'

4.4.4 Language Support

Currently, advanced compression features are fully supported for:

• Python: Complete AST-based compression with all modes
• JavaScript/TypeScript: Tree-sitter based compression
• Go: Tree-sitter based compression
• Java: Tree-sitter based compression
• C/C++: Tree-sitter based compression
• C#: Tree-sitter based compression
• Rust: Tree-sitter based compression
• Ruby: Tree-sitter based compression
• PHP: Tree-sitter based compression
• Swift: Tree-sitter based compression
• CSS: Tree-sitter based compression
• Other Languages: Basic compression with warnings (future enhancement planned)

4.4.5 Example Output

Original Python Code:

def calculate_sum(a: int, b: int) -> int:
    """
    Calculate the sum of two integers.
    
    Args:
        a: First integer
        b: Second integer
        
    Returns:
        The sum of a and b
    """
    if not isinstance(a, int) or not isinstance(b, int):
        raise TypeError("Both arguments must be integers")
    
    result = a + b
    print(f"Calculating {a} + {b} = {result}")
    return result

Interface Mode Output:

def calculate_sum(a: int, b: int) -> int:
    """
    Calculate the sum of two integers.
    
    Args:
        a: First integer
        b: Second integer
        
    Returns:
        The sum of a and b
    """
    pass

4.5 Ignore Patterns

Repomix provides multiple methods to set ignore patterns for excluding specific files or directories during the packing process:

Priority Order

Ignore patterns are applied in the following priority order (from highest to lowest):

• Custom patterns in configuration file (ignore.custom_patterns)
• .repomixignore file
• .gitignore file (if ignore.use_gitignore is true)
• Default patterns (if ignore.use_default_ignore is true)

Ignore Methods

.gitignore

By default, Repomix uses patterns listed in your project's .gitignore file. This behavior can be controlled with the ignore.use_gitignore option in the configuration file:

{
  "ignore": {
    "use_gitignore": true
  }
}

Default Patterns

Repomix includes a default list of commonly excluded files and directories (e.g., __pycache__, .git, binary files). This feature can be controlled with the ignore.use_default_ignore option:

{
  "ignore": {
    "use_default_ignore": true
  }
}

The complete list of default ignore patterns can be found in default_ignore.py.

.repomixignore

You can create a .repomixignore file in your project root to define Repomix-specific ignore patterns. This file follows the same format as .gitignore.

Custom Patterns

Additional ignore patterns can be specified using the ignore.custom_patterns option in the configuration file:

{
  "ignore": {
    "custom_patterns": [
      "*.log",
      "*.tmp",
      "tests/**/*.pyc"
    ]
  }
}

Notes

• Binary files are not included in the packed output by default, but their paths are listed in the "Repository Structure" section of the output file. This provides a complete overview of the repository structure while keeping the packed file efficient and text-based.
• Ignore patterns help optimize the size of the generated pack file by ensuring the exclusion of security-sensitive files and large binary files, while preventing the leakage of confidential information.
• All ignore patterns use glob pattern syntax similar to .gitignore.

🔒 5. Output File Format

Repomix generates a single file with clear separators between different parts of your codebase. To enhance AI comprehension, the output file begins with an AI-oriented explanation, making it easier for AI models to understand the context and structure of the packed repository.

5.1 Plain Text Format (default)

This file is a merged representation of the entire codebase, combining all repository files into a single document.

================================================================
File Summary
================================================================
(Metadata and usage AI instructions)

================================================================
Repository Structure
================================================================
src/
  cli/
    cliOutput.py
    index.py
  config/
    configLoader.py

(...remaining directories)

================================================================
Repository Files
================================================================

================
File: src/index.py
================
# File contents here

================
File: src/utils.py
================
# File contents here

(...remaining files)

================================================================
Statistics
================================================================
(File statistics and metadata)

5.2 Markdown Format

To generate output in Markdown format, use the --style markdown option:

python -m repomix --style markdown

The Markdown format structures the content in a readable manner:

# File Summary
(Metadata and usage AI instructions)

# Repository Structure
```
src/
  cli/
    cliOutput.py
    index.py
```

# Repository Files

## File: src/index.py
```python
# File contents here
```

## File: src/utils.py
```python
# File contents here
```

# Statistics
- Total Files: 19
- Total Characters: 37377
- Total Tokens: 11195

5.3 XML Format

To generate output in XML format, use the --style xml option:

python -m repomix --style xml

The XML format structures the content in a hierarchical manner:

<?xml version="1.0" encoding="UTF-8"?>
<repository>
<repository_structure>
(Directory and file structure)
</repository_structure>

<repository_files>
<file>
  <path>src/index.py</path>
  <stats>
    <chars>1234</chars>
    <tokens>567</tokens>
  </stats>
  <content>
    # File contents here
  </content>
</file>
(...remaining files)
</repository_files>

<statistics>
  <total_files>19</total_files>
  <total_chars>37377</total_chars>
  <total_tokens>11195</total_tokens>
</statistics>
</repository>

5.4 JSON Format

To generate output in JSON format, use the --style json option:

python -m repomix --style json

The JSON format provides machine-readable structured output:

{
  "summary": {
    "total_files": 19,
    "total_chars": 37377,
    "total_tokens": 11195,
    "generation_date": "2025-01-28T12:00:00"
  },
  "file_tree": {
    "src": {
      "index.py": "",
      "utils.py": ""
    }
  },
  "files": [
    {
      "path": "src/index.py",
      "content": "# File contents here",
      "chars": 1234,
      "tokens": 567
    }
  ]
}

JSON format is ideal for:

• Integration with other tools and scripts
• Programmatic processing of codebase analysis
• Building custom pipelines and workflows

🛠️ 6. Advanced Usage

6.1 Library Usage

You can use Repomix as a Python library in your projects. Here's a basic example:

from repomix import RepoProcessor

# Basic usage
processor = RepoProcessor(".")
result = processor.process()

# Process remote repository with specific branch
processor = RepoProcessor(repo_url="https://github.com/username/repo", branch="feature-branch")
result = processor.process()

# Access processing results
print(f"Total files: {result.total_files}")
print(f"Total characters: {result.total_chars}")
print(f"Total tokens: {result.total_tokens}")
print(f"Output saved to: {result.config.output.file_path}")

6.2 Advanced Configuration

from repomix import RepoProcessor, RepomixConfig

# Create custom configuration
config = RepomixConfig()

# Output settings
config.output.file_path = "custom-output.md"
config.output.style = "markdown"  # supports "plain", "markdown", and "xml"
config.output.show_line_numbers = True

# Security settings
config.security.enable_security_check = True
config.security.exclude_suspicious_files = True

# Compression settings
config.compression.enabled = True
config.compression.keep_signatures = True
config.compression.keep_docstrings = True
config.compression.keep_interfaces = True  # Interface mode for API documentation

# Include/Ignore patterns
config.include = ["src/**/*", "tests/**/*"]
config.ignore.custom_patterns = ["*.log", "*.tmp"]
config.ignore.use_gitignore = True

# Remote repository configuration
config.remote.url = "https://github.com/username/repo"
config.remote.branch = "feature-branch"

# Process repository with custom config
processor = RepoProcessor(".", config=config)
result = processor.process()

6.2.1 Compression Examples

from repomix import RepoProcessor, RepomixConfig

# Example 1: Generate API documentation (Interface Mode)
config = RepomixConfig()
config.compression.enabled = True
config.compression.keep_interfaces = True  # Keep signatures + docstrings only
config.output.file_path = "api-documentation.md"

processor = RepoProcessor(".", config=config)
result = processor.process()
print(f"API documentation generated: {result.config.output.file_path}")

# Example 2: Code overview without implementation details
config = RepomixConfig()
config.compression.enabled = True
config.compression.keep_signatures = True
config.compression.keep_docstrings = False
config.compression.keep_interfaces = False  # Keep full signatures but remove docstrings
config.output.file_path = "code-overview.md"

processor = RepoProcessor(".", config=config)
result = processor.process()

# Example 3: Extract only configuration and constants
config = RepomixConfig()
config.compression.enabled = True
config.compression.keep_signatures = False  # Remove all functions/classes
config.output.file_path = "config-only.md"

processor = RepoProcessor(".", config=config)
result = processor.process()

For more example code, check out the examples directory:

• basic_usage.py: Basic usage examples
• custom_config.py: Custom configuration examples
• security_check.py: Security check feature examples
• file_statistics.py: File statistics examples
• remote_repo_usage.py: Remote repository processing examples
• json_output.py: JSON output format examples
• git_integration.py: Git diff, log, and sorting examples
• output_split.py: Output splitting for large codebases
• token_count_tree.py: Token distribution visualization
• full_directory_structure.py: Full directory tree display
• tree_sitter_compression.py: Tree-sitter compression examples

6.3 Environment Variables

• REPOMIX_COCURRENCY_STRATEGY: Set to thread or process to manually control the concurrency strategy used for file processing (default is process, but thread might be used automatically in environments like AWS Lambda or if set explicitly).
• REPOMIX_LOG_LEVEL: Set the logging level. Available values are TRACE, DEBUG, INFO, SUCCESS, WARN, and ERROR (default is INFO). This setting controls the verbosity of log output regardless of the --verbose flag.

🤖 7. AI Usage Guide

7.1 Prompt Examples

Once you have generated the packed file with Repomix, you can use it with AI tools like Claude, ChatGPT, and Gemini. Here are some example prompts to get you started:

Code Review and Refactoring

For a comprehensive code review and refactoring suggestions:

This file contains my entire codebase. Please review the overall structure and suggest any improvements or refactoring opportunities, focusing on maintainability and scalability.

Documentation Generation

To generate project documentation:

Based on the codebase in this file, please generate a detailed README.md that includes an overview of the project, its main features, setup instructions, and usage examples.

Test Case Generation

For generating test cases:

Analyze the code in this file and suggest a comprehensive set of unit tests for the main functions and classes. Include edge cases and potential error scenarios.

Code Quality Assessment

Evaluate code quality and adherence to best practices:

Review the codebase for adherence to coding best practices and industry standards. Identify areas where the code could be improved in terms of readability, maintainability, and efficiency. Suggest specific changes to align the code with best practices.

Library Overview

Get a high-level understanding of the library

This file contains the entire codebase of library. Please provide a comprehensive overview of the library, including its main purpose, key features, and overall architecture.

API Documentation Review

For reviewing API interfaces (when using interface mode compression):

This file contains the API interfaces of my codebase with all implementation details removed. Please review the API design, suggest improvements for consistency, and identify any missing documentation or unclear method signatures.

Code Architecture Analysis

For analyzing code structure (when using signature mode compression):

This file contains the code structure with function signatures but minimal implementation details. Please analyze the overall architecture, identify design patterns used, and suggest improvements for better modularity and separation of concerns.

Configuration Analysis

For analyzing configuration and constants (when using minimal mode compression):

This file contains only the configuration, constants, and global variables from my codebase. Please review these settings, identify potential configuration issues, and suggest best practices for configuration management.

Feel free to modify these prompts based on your specific needs and the capabilities of the AI tool you're using.

7.2 MCP (Model Context Protocol) Server

Repomix can run as an MCP server, allowing AI assistants like Claude to directly interact with your codebase without manual file preparation.

📦 Installation Required: Before using MCP features, make sure you have installed repomix: pip install repomix

Starting the MCP Server

# Start the MCP server (detailed logs output to stderr)
repomix --mcp

After starting, you'll see logs like:

📦 Repomix v0.3.0

Starting Repomix MCP Server...
🔧 Creating MCP server...
📦 Registering MCP tools...
  ✅ pack_codebase
  ✅ pack_remote_repository
  ✅ read_repomix_output
  ✅ grep_repomix_output
  ✅ file_system_read_file
  ✅ file_system_read_directory
  ✅ generate_skill
🎯 Repomix MCP Server configured with 7 tools
🚀 Starting Repomix MCP Server on stdio transport...
📡 Waiting for MCP client connections...
💡 Use Ctrl+C to stop the server
──────────────────────────────────────────────────

Configuring in AI Assistants

Claude Desktop Add to claude_desktop_config.json:

{
  "mcpServers": {
    "repomix": {
      "command": "repomix",
      "args": ["--mcp"],
      "cwd": "/path/to/your/project"
    }
  }
}

VS Code / Cline Add to cline_mcp_settings.json:

{
  "mcpServers": {
    "repomix": {
      "command": "repomix", 
      "args": ["--mcp"],
      "cwd": "/path/to/your/project"
    }
  }
}

Claude Code

# From any directory (after installing repomix)
claude mcp add repomix -- repomix --mcp

About the cwd Parameter

The cwd (current working directory) parameter in MCP configuration determines where the repomix command runs from. Here are the recommended settings:

• For general use: Set cwd to your home directory or any stable location like "/Users/yourusername" (macOS) or "/home/yourusername" (Linux)
• For specific projects: Set cwd to your main project directory that you frequently analyze
• For development: You can use any directory since repomix can process any path you specify in the tool calls

Examples:

// General use - works from anywhere
"cwd": "/Users/yourusername"

// Project-specific - convenient for frequent analysis  
"cwd": "/Users/yourusername/projects/my-main-project"

// Development - flexible starting point
"cwd": "/Users/yourusername/dev"

💡 Pro Tip: The MCP tools allow you to specify target directories in the tool parameters, so the cwd is just the starting location. You can analyze any accessible directory regardless of where the server starts.

Available MCP Tools

• pack_codebase - Package local codebase into XML format

  • Parameters: directory, compress, include_patterns, ignore_patterns, top_files_length

• read_repomix_output - Read generated output files

  • Parameters: output_id, start_line, end_line

• grep_repomix_output - Search within output files

  • Parameters: output_id, pattern, context_lines, ignore_case

• file_system_read_file - Read files from filesystem

  • Parameters: path

• file_system_read_directory - List directory contents

  • Parameters: path

• pack_remote_repository - Package remote repositories

  • Parameters: remote, compress, include_patterns, ignore_patterns

• generate_skill - Generate Claude Agent Skills from codebase

  • Parameters: directory, skill_name, include_patterns, ignore_patterns

Tool Call Logs

When AI assistants call tools, you'll see detailed logs in the server terminal:

🔨 MCP Tool Called: pack_codebase
   📁 Directory: /path/to/project
   🗜️ Compress: false
   📊 Top files: 10
   🏗️ Creating workspace...
   📝 Output will be saved to: /tmp/repomix_mcp_xxx/repomix-output.xml
   🔄 Processing repository...
   ✅ Processing completed!
   📊 Files processed: 45
   📝 Characters: 125,432
   🎯 Tokens: 0
   🎉 MCP response generated successfully

Features

• ✅ Complete MCP protocol support
• ✅ Detailed operation logging
• ✅ Security file checking
• ✅ Multiple output formats
• ✅ File search and reading
• ✅ Temporary file management
• 🔄 Remote repository support (in development)
• 🔄 Code compression features (in development)

7.3 Best Practices

• Be Specific: When prompting the AI, be as specific as possible about what you want. The more context you provide, the better the results will be.
• Iterate: Don't be afraid to iterate on your prompts. If you don't get the results you want on the first try, refine your prompt and try again.
• Combine with Manual Review: While AI can be a powerful tool, it's not perfect. Always combine AI-generated output with manual review and editing.
• Security First: Always be mindful of security when working with your codebase. Use Repomix's built-in security checks and avoid sharing sensitive information with AI tools.

📄 8. License

This project is licensed under the MIT License.

For more detailed information, please visit the repository.