thomwebb/gac

What it solves

Git Auto Commit (gac) solves the tedious task of writing descriptive, high-quality commit messages. Instead of generic messages like "fixed bug" or "updated files," it uses Large Language Models (LLMs) to analyze your code changes and generate contextual messages that explain the "why" behind the changes.

How it works

gac acts as a CLI tool that integrates with Git. It analyzes the staged changes in your repository, sends the relevant code structure and logic to a supported LLM provider, and generates a commit message based on theon desired format. Users can then review, edit, or reroll the generated message before confirming the commit.

Who it’s for

Developers who want to automate their commit history with professional, semantic commit messages without spending time writing them manually.

Highlights

• Extensive Provider Support: Supports over 33 AI providers, including OpenAI, Anthropic, Gemini, DeepSeek, and local options like Ollama.
• Smart Analysis: Recognizes refactoring, bug fixes, and features, while filtering out noise like generated files and dependencies.
• Multiple Formats: Offers one-liners, standard bulleted summaries, verbose explanations, and adherence to the 50/72 rule.
• Interactive Feedback: Allows users to reroll messages, edit them in-place or via an external editor, or provide direct feedback to the LLM to refine the result.
• Multilingual Support: Generates messages in over 25 languages.
• Built-in Security: Scans for API keys and passwords before committing to prevent accidental secret leaks.
• MCP Server: Can be run as a Model Context Protocol server to expose commit tools to other AI agents.
• Usage Statistics: Tracks local stats on token usage, latency, and commit streaks.