products / practical tool

Git Commit Analyzer

Local AI Git plugin that turns staged diffs into clear Conventional Commits-compatible messages with llama.cpp and GGUF models.
Powered by Local AI - No Cloud Required

Generate consistent, meaningful Git commit messages automatically using local llama.cpp inference with GGUF models. All processing happens on your machine - your code never leaves your device.

AI-Powered

Local GGUF models

Conventional Commits

type(scope): subject

Interactive

Use, edit, or cancel

100% Private

Local processing

Quick Installation

🍺

Homebrew (Recommended)

Fast installation with pre-built binaries for macOS

$brew tap zh30/tap
$brew install git-ca
No Rust compilation required
Apple Silicon & Intel support

⚡ One-Click Install

bash -c "$(curl -fsSL https://sh.zhanghe.dev/install-git-ca.sh)"

Manual Download

GitHub Releases

System Requirements

Git (version 2.30+)
GGUF model (auto-downloaded)
macOS

Homebrew (Intel & Apple Silicon)

Windows

Download a pre-built binary from GitHub Releases

Linux

Pre-built releases are temporarily unavailable

Open Source

Fully open source on GitHub. Contributions welcome!

View on GitHub

Key Features

Smart Message Generation

Analyzes staged changes and generates contextual, meaningful commit messages using local GGUF models.

Conventional Commits

Produces type(scope): subject messages for a clean, maintainable commit history.

Interactive Mode

Choose to use, edit, or cancel the proposed commit message with full control.

100% Local & Private

All AI processing happens locally using llama.cpp - no data leaves your device.

How to Use

1
Stage changes: git add <files>
2
Run analyzer: git ca
3
Use, edit, or cancel the suggested message

Configuration

git ca modelSelect model
git ca languageEN/ZH
git ca doctorDiagnostics

Why Choose Git Commit Analyzer?

AI-Powered Intelligence

Leverages local llama.cpp with GGUF models for contextual understanding.

Consistent History

Ensures clean, maintainable commit history following Git Flow standards.

Developer Privacy

Complete privacy - no external servers or API calls.

Privacy & Security

Local Processing: All AI inference happens locally using llama.cpp with GGUF models.
No Cloud: Your code and commit messages never leave your device. Zero external API calls.
Open Source: The entire codebase is open source and auditable on GitHub.
No Telemetry: We do not collect any usage data or analytics from the CLI tool.
related workflows
XiaomaolvSelf-hosted Rust Telegram AI agent with MCP tools, MiniMax and OpenAI-compatible providers, and SQLite memoryBeejsOpen-source JavaScript and TypeScript runtime built with Rust and V8, with a focused v0.1 CLI and compatibility layer
official links

Official links

Use these canonical links for citations, installation, source review, privacy checks, and support.

WebsiteGitHubGitHub ReleasesSupport

Frequently Asked Questions