Gitmit

🧠 Smart, AI-Enhanced Git Commit Message Generator

---

Gitmit is a high-performance CLI tool that takes the guesswork out of commit messages. It analyzes your staged changes and generates professional, context-aware messages following the Conventional Commits specification.

Whether you prefer deterministic heuristic rules or powerful local LLMs, Gitmit provides a seamless, offline-first experience to keep your project history clean and meaningful.

✨ Why Gitmit?

• Hybrid Intelligence: Combines fast, deterministic heuristics with optional Local AI (Ollama) for deep semantic understanding.
• Privacy First: Operates 100% locally. No API keys, no data leaving your machine.
• Project Aware: Automatically detects your project type (Go, Node.js, Python, Rust, etc.) and tailors suggestions accordingly.
• Seamless Workflow: Integrated interactive mode allows you to accept, edit, or regenerate suggestions instantly.

Screenshots

🚀 Quick Start

Installation

Using Go:

go install github.com/andev0x/gitmit@latest

From Source:

git clone https://github.com/andev0x/gitmit.git
cd gitmit
make build
sudo make install

Basic Usage

1. Stage your changes:

   git add .

2. Run Gitmit:

   gitmit

3. Profit! Accept the suggestion or interactively refine it.

🛠 Features

🔍 Intelligent Analysis

Uses git status --porcelain and git diff to understand exactly what changed. It's not just looking at filenames; it understands additions, deletions, and modifications at a structural level.

🤖 Hybrid Intelligence (Local AI)

Optionally use models like qwen2.5-coder via Ollama for human-like commit quality. If the AI is unavailable, Gitmit instantly falls back to its robust heuristic engine.

📋 Conventional Commit Standards

Supports feat, fix, refactor, chore, test, docs, style, perf, ci, build, security, and more, ensuring your team stays aligned with industry standards.

⚖️ Weighted Scoring Matrix

An advanced algorithm that aggregates signals from:

• Branch Names: Extracts intent from feature/auth or fix/bug-123.
• Keywords: Weighted analysis of code changes.
• Diff Stats: Understands the ratio of added/deleted lines to distinguish between features and refactors.

📦 Language & Ecosystem Awareness

• Symbol Extraction: Detects function, class, and variable names in Go, JS/TS, Python, and Java.
• Dependency Watcher: Identifies when you add or update libraries in go.mod, package.json, requirements.txt, etc.

🧠 Local AI Setup

To leverage the power of LLMs offline:

1. Install Ollama from ollama.com.
2. Pull a model: ollama pull qwen2.5-coder:7b
3. Initialize Gitmit config: gitmit init
4. Enable AI in .gitmit.json:

   {
     "engine": "ollama",
     "ollama": {
       "model": "qwen2.5-coder:7b"
     }
   }

⌨️ Command Reference

Command	Description
gitmit	Analyze changes and suggest a message interactively.
gitmit init	Create a local .gitmit.json configuration.
gitmit init --global	Create a global ~/.gitmit.json configuration.
gitmit propose --auto	Automatically commit with the best suggestion.
gitmit propose -s	Show multiple ranked suggestions.
gitmit --version	Show version information.

Interactive Actions:

• y: Accept and commit.
• n: Exit without committing.
• e: Edit the message manually.
• r: Regenerate a new suggestion.
• a: Upgrade to AI suggestion on-the-fly.

⚙️ Configuration

Gitmit works out of the box with zero configuration. For advanced users, it offers deep customization via a tiered config system (Local → Global → Default).

gitmit init --global

Key configuration options include topicMappings, keywordWeights, and diffStatThreshold. For a full deep dive, see docs/CONFIGURATION.md.

🤝 Contributing

Contributions are what make the open-source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (gitmit)
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Check out our CONTRIBUTING.md for more details.

⚖️ License

Distributed under the MIT License. See LICENSE for more information.

---

Built with ❤️ by andev0x