🦜 LLM-Auto-Context

A CLI tool for generating code snapshots for AI code assistants such as ChatGPT and Claude.

Installation

Using UV (recommended):

uv pip install llm-auto-context

Using pip:

pip install llm-auto-context

Usage

Basic usage:

llm_auto_context

This will use the default config file .codesnapshot.json in your current directory.

Configuration

Create a .codesnapshot.json file in your project root:

{
    "directories": ["src", "lib"],
    "output_file": "code_snapshot.md",
    "include_extensions": [".py", ".js", ".ts"],
    "exclude_dirs": ["node_modules", ".git", "build"],
    "exclude_files": ["secrets.env"]
}

CLI Options

• --config: Path to config file (default: .codesnapshot.json)
• -d, --directory: Override directories to scan (can be used multiple times)
• -o, --output: Override output file path
• --include: Override file extensions to include (can be used multiple times)
• --exclude-dir: Additional directories to exclude (can be used multiple times)
• --exclude-file: Additional files to exclude (can be used multiple times)

Example:

llm_auto_context -d src -d lib -o snapshot.md --exclude-dir tests

Development

1. Clone the repository
2. Create a virtual environment:

uv venv

3. Install dependencies:

uv pip install -e ".[dev]"

4. Run tests:

pytest

License

MIT