Repomix

Core Features#

• Multiple Output Formats: XML (recommended for Claude), Markdown, JSON, Plain Text
• Code Compression: Tree-sitter based extraction of key code elements (classes, functions, interfaces), ~70% token reduction
• Token Management: Precise token counting and statistics tree visualization to manage LLM context limits
• Security Check: Integrated Secretlint for detecting API keys, passwords, and sensitive information
• Remote Processing: Process GitHub repositories without cloning, supports branch specification
• MCP Integration: Model Context Protocol support for AI tools like Claude and Cursor
• Output Splitting: Split large outputs into multiple files (e.g., --split-output 1mb)
• Git History: Optional inclusion of git commit history and diff information

Quick Start#

# Method 1: No installation needed, use npx (recommended)
npx repomix@latest

# Method 2: npm global install
npm install -g repomix

# Method 3: Homebrew (macOS/Linux)
brew install repomix

# Method 4: Docker
docker run -v .:/app -it --rm ghcr.io/yamadashy/repomix

Common Commands#

repomix                                    # Pack current directory
repomix path/to/directory                  # Pack specific directory
repomix --include "src/**/*.ts,**/*.md"    # Include specific files
repomix --ignore "**/*.log,tmp/"           # Exclude specific files
repomix --remote https://github.com/user/repo  # Process remote repository
repomix --style markdown                   # Specify output format
repomix --compress                         # Enable code compression
repomix --split-output 20mb                # Split large output
repomix --token-count-tree                 # Show token statistics tree
repomix --mcp                              # Run as MCP server

Configuration Example#

{
  "$schema": "https://repomix.com/schemas/latest/schema.json",
  "output": {
    "filePath": "repomix-output.xml",
    "style": "xml",
    "compress": false,
    "removeComments": false,
    "tokenCountTree": false
  },
  "ignore": {
    "useGitignore": true,
    "useDefaultPatterns": true,
    "customPatterns": ["additional-folder", "**/*.log"]
  },
  "security": {
    "enableSecurityCheck": true
  }
}

Typical Use Cases#

• Code review and quality assessment: Send entire codebase to AI for structure and quality analysis
• Refactoring suggestions and architecture analysis: Get AI-powered refactoring and optimization recommendations
• Documentation and test case auto-generation: Generate README, technical docs, and unit tests from codebase
• Quick understanding of third-party library architecture
• Context preparation for AI coding assistants

Key Dependencies#

• @modelcontextprotocol/sdk - MCP server support
• tiktoken - OpenAI official token counting
• web-tree-sitter - Code structure parsing
• @secretlint/core - Security scanning
• commander - CLI framework
• globby - File pattern matching

Requirements#

• Node.js >= 20.0.0
• Or Docker environment