Automatic git commit message generation using AI with a beautiful terminal interface and comprehensive testing.
- Multi-Provider AI: Supports OpenRouter, OpenAI, and Anthropic backends.
- Beautiful UI: Rich terminal interface with colored output and styled boxes.
- Smart Parsing: Robust commit message parsing with markdown cleanup.
- Interactive: Preview, confirm, regenerate, or manually modify commit messages.
- Self-Testing: Built-in health checks and a comprehensive test suite.
- Conventional Commits: Enforces best practices for commit messages.
Ensure Python 3.11+ and Git are installed. You can use pyenv for version management:
# Install pyenv (if needed)
curl https://pyenv.run | bash
# Add to ~/.zshrc and source it
pyenv install 3.11.13
pyenv local 3.11.13# Clone repo
git clone https://github.com/rozeraf/autocommit.git
cd autocommit
# Install with pip
pip install -r requirements.txtCreate a .env file with your API keys. You only need to add the keys for the providers you intend to use.
# .env
OPENROUTER_API_KEY=sk-or-v1-your-key-here
OPENAI_API_KEY=sk-your-key-here
ANTHROPIC_API_KEY=sk-ant-your-key-hereGet your keys from:
- OpenRouter: https://openrouter.ai
- OpenAI: https://platform.openai.com/api-keys
- Anthropic: https://console.anthropic.com/settings/keys
⚠️ Important for OpenRouter Free Models: If you're using free models (e.g.,deepseek/deepseek-chat-v3.1:free), you must enable "Free model publication" in your OpenRouter privacy settings at https://openrouter.ai/settings/privacy. Without this setting, you'll get a "No endpoints found" error.
The application uses config.toml for all settings. A default configuration is used if the file is not present.
Create config.toml in the project root to customize settings. Example:
# Default provider to use
base_provider = "openrouter"
# Settings for each provider
[ai.providers.openrouter]
model = "deepseek/deepseek-chat-v3.1:free"
temperature = 0.3
env_key = "OPENROUTER_API_KEY"
[ai.providers.openai]
model = "gpt-4o-mini"
temperature = 0.3
env_key = "OPENAI_API_KEY"
[ai.providers.anthropic]
model = "claude-3-5-sonnet-20240620"
temperature = 0.3
env_key = "ANTHROPIC_API_KEY"
# Other settings
[format]
max_subject_length = 70
[context]
auto_detect = true# Stage changes
git add .
# Create an AI-generated commit message
python3 main.py
# --- Provider Management ---
# Use a specific provider (e.g., openai)
python3 main.py --provider openai
# List all available providers
python3 main.py --list-providers
# Test connectivity for all configured providers
python3 main.py --test-providers
# Get detailed info about a provider
python3 main.py --provider-info anthropic
# --- Other Options ---
# Skip confirmation prompt
python3 main.py -y
# Generate message without committing (dry run)
python3 main.py --dry-run
# Run comprehensive self-tests
python3 main.py --testThis project has undergone significant improvements with the latest v2.3 release:
- Multi-Provider AI: Switch between OpenRouter, OpenAI, and Anthropic.
- New CLI Commands: Manage and test providers directly from the command line.
- Enhanced Test Suite: Full test coverage for the new provider system.
For detailed information about all changes, see CHANGELOG.md.
Managed via pip in requirements.txt. Key packages:
requests- API communicationpython-dotenv- Environment variable managementcolorama- Cross-platform colored terminal outputhalo- Beautiful loading spinnersrich- Enhanced terminal formatting and UI
Development: ruff, pytest, pytest-cov.
autocommit/
├── src/
│ ├── api/ # AI provider implementations
│ ├── config/ # Configuration system
│ ├── context/ # Context detection module
│ ├── models/ # Data models (dataclasses)
│ ├── parsers/ # Message and diff parsing
│ ├── git_utils.py # Git operations
│ └── ui.py # Rich terminal interface
├── tests/
│ ├── test_providers.py # Tests for AI providers
│ ├── test_factory.py # Tests for ProviderFactory
│ ├── test_manager.py # Tests for AIProviderManager
│ └── ... # Other existing tests
├── main.py # CLI entrypoint
├── config.toml # Main configuration file
├── requirements.txt # pip dependencies
└── LICENSE
The project includes comprehensive testing:
# Run all tests
pytest
# Run with coverage
pytest --cov=src
# Run application self-tests
python3 main.py --test
# Test API connection
python3 main.py --test-apiTest coverage includes:
- AI response parsing edge cases
- Markdown and code block cleanup
- Complex AI response handling
- Whitespace and formatting
- Conventional commit detection
- Fork the repo
- Create feature branch:
git checkout -b feature/amazing-feature - Install dev deps:
pip install -r requirements.txt - Run tests:
pytest - Commit with AI:
git add . && python3 main.py - Push and create PR
Ensure Python 3.11+ and all tests pass before submitting.
For detailed information about changes and version history, see CHANGELOG.md.
GPL-3.0 - See LICENSE file for details.
The application features a beautiful terminal interface with:
- Styled commit preview boxes
- Interactive confirmation prompts
- Loading spinners with progress feedback
- Colored output and statistics
- Responsive terminal layouts
# Use different AI models
python3 main.py --model anthropic/claude-3.5-sonnet
python3 main.py --model openrouter/sonoma-dusk-alpha
# Debug mode for troubleshooting
python3 main.py --debug
# Skip confirmation
python3 main.py -y
# Generate message without committing
python3 main.py --dry-run- Fast: Optimized API calls with smart diff truncation
- Reliable: Comprehensive error handling and retries
- Smart: Context-aware parsing for complex AI responses
- Responsive: Terminal-adaptive UI that works on any screen size