GitGenie 🧞‍♂️

AI-powered CLI tool that generates commit messages and PR descriptions using local LLMs.

Features

• 🎯 Smart Commit Messages - Analyzes staged changes and generates conventional commit messages
• 📝 PR Descriptions - Creates comprehensive PR descriptions from commit history
• 🔄 Regenerate Option - Don't like the suggestion? Regenerate with one keystroke
• 🤖 Local AI - Uses Ollama for private, offline AI generation
• ⚡ Streaming Output - See AI responses generate in real-time

Demo

$ git add .
$ gitgenie commit

Generating commit message...
feat(auth): add JWT token validation and refresh mechanism

[c]ommit, [r]egenerate, [q]uit: c
✓ Committed successfully!

Prerequisites

1. Python 3.9+
2. Ollama - Install from ollama.com
3. llama3 model

Install Ollama and llama3

# Install Ollama (macOS)
brew install ollama

# Or download from https://ollama.com

# Start Ollama
ollama serve

# Pull the llama3 model (in a new terminal)
ollama pull llama3

Installation

# Clone the repository
git clone https://github.com/mhs170/gitgenie.git
cd gitgenie

# Create a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install the package
pip install -e .

Usage

Generate Commit Messages

# Stage your changes
git add .

# Generate a commit message
gitgenie commit

# Options:
# [c] - Commit with the generated message
# [r] - Regenerate a new message
# [q] - Cancel and quit

Generate PR Descriptions

# Generate PR description (compares against main)
gitgenie pr

# Compare against a different branch
gitgenie pr master

How It Works

1. Commit Messages: GitGenie reads your staged git diff, sends it to llama3, and generates a conventional commit message following best practices
2. PR Descriptions: GitGenie analyzes your commit history between branches and creates a structured PR description with overview, changes, and testing notes

Configuration

GitGenie uses Ollama running locally on localhost:11434. Make sure Ollama is running before using GitGenie:

# Check if Ollama is running
ollama list

# Start Ollama if needed
ollama serve

Commit Message Format

GitGenie follows Conventional Commits:

type(scope): description

Examples:
feat(auth): add user login
fix(api): resolve timeout issue
docs(readme): update installation steps

PR Description Format

Generated PR descriptions include:

• Title: Brief summary
• Overview: What and why
• Changes: Bullet-pointed key changes
• Testing: How to test the changes

Troubleshooting

Error: "Ollama might not be running"

# Start Ollama
ollama serve

Error: "Model not found"

# Pull the llama3 model
ollama pull llama3

Error: "Not a git repository"

• Make sure you're inside a git repository
• Run git init if needed

Contributing

Contributions are welcome! Feel free to open issues or submit pull requests.

License

MIT

Author

Mohammed Salama - GitHub