Git CMM — AI-Powered Git Commit Message Generator

A lightweight, cross-platform desktop app that generates Conventional Commits messages from your git diff using AI. Supports multiple LLM providers — local Ollama, OpenRouter, OpenAI, and any OpenAI-compatible API. Under 10 MB installer, zero telemetry, your code never leaves the machine (when using local models).

Built with Tauri v2 (Rust backend + native WebView) and React 19.

---

Features

• AI commit messages — generates Conventional Commits from staged diffs via streaming LLM
• Built-in Git operations — commit, push, pull, fetch with rebase/force-with-lease options
• Branch management — switch/create branches; handles uncommitted changes gracefully
• Commit history graph — interactive SVG multi-branch graph with merge connectors, per-branch colors, and timeline grouping
• Submodule support — sidebar lists submodules with status badges; click to operate independently
• Multi-provider LLM — switch between local Ollama, Ollama Cloud, OpenRouter, OpenAI, Groq, etc.
• Privacy-first — runs fully offline with local Ollama; no data sent anywhere

Not a full Git GUI. It does not resolve merge conflicts, delete/rename branches, or open PRs — use your IDE/CLI for those.

---

Screenshots

Main window

Settings

|

LLM Config - Compatible with OpenAI

---

Table of Contents

• Quick Start (Pre-built Installer)
• Build from Source
• Configuration
• Usage
• Tech Stack
• Project Structure
• CI / Releases
• License

---

Quick Start (Pre-built Installer)

Download the latest installer from the Releases page:

Platform	File	Architecture
macOS (Apple Silicon)	Git CMM_x.y.z_aarch64.dmg	ARM64 (M1/M2/M3/M4)
macOS (Intel)	Git CMM_x.y.z_x86_64.dmg	x86_64
Windows	Git CMM_x.y.z_x64_en-US.msi	x64
Windows (alt)	Git CMM_x.y.z_x64-setup.exe	x64 (NSIS)

macOS — bypass Gatekeeper (unsigned app)

xattr -d com.apple.quarantine "/Applications/Git CMM.app"

Windows — bypass SmartScreen (unsigned app)

Click More info → Run anyway on the blue SmartScreen dialog.

---

Build from Source

Prerequisites

Tool	Minimum Version	Install
Node.js	20+	nodejs.org
npm	10+	Bundled with Node.js
Rust	stable 1.77+	rustup.rs
Git	2.20+	Must be in PATH (also needed at runtime)

Platform-specific dependencies

macOS

xcode-select --install

Windows

• WiX Toolset 3.x — for .msi bundling:

  winget install WiXToolset.WiXToolset

• WebView2 Runtime — preinstalled on Windows 11; download for Windows 10

Linux (experimental)

sudo apt install libwebkit2gtk-4.1-dev libssl-dev libgtk-3-dev \
  libayatana-appindicator3-dev librsvg2-dev

Clone & Install

git clone https://github.com/baka3k/git-cmm.git
cd git-cmm
npm install

Development (hot reload)

npm run tauri dev

• Frontend changes (.tsx) — Vite HMR reloads instantly
• Rust changes (.rs) — Tauri recompiles and restarts the app
• Open DevTools: Cmd+Option+I (macOS) / Ctrl+Shift+I (Windows/Linux)

Production Build

npm run tauri build

This command:

1. Runs tsc && vite build → bundles frontend into dist/
2. Runs cargo build --release in src-tauri/
3. Runs Tauri bundler → creates platform installer

Output location: src-tauri/target/release/bundle/

OS	Format	Path
macOS	.app	macos/Git CMM.app
macOS	.dmg	dmg/Git CMM_x.y.z_<arch>.dmg
Windows	.msi	msi/Git CMM_x.y.z_x64_en-US.msi
Windows	.exe	nsis/Git CMM_x.y.z_x64-setup.exe
Linux	.deb / .AppImage / .rpm	deb/ · appimage/ · rpm/

First build takes 5–15 minutes (compiling all Rust dependencies). Subsequent incremental builds are seconds.

Useful Build Commands

# macOS universal binary (Apple Silicon + Intel in one .dmg)
rustup target add aarch64-apple-darwin x86_64-apple-darwin
npm run tauri build -- --target universal-apple-darwin

# Single target (Apple Silicon only)
npm run tauri build -- --target aarch64-apple-darwin

# Binary only, skip bundling (faster iteration)
npm run tauri build -- --no-bundle

# Rust type-check only (no compile)
cargo check --manifest-path src-tauri/Cargo.toml

---

Configuration

LLM Provider Setup

You need at least one LLM provider. Options:

Option A: Local Ollama (recommended — fully offline, private)

1. Install Ollama from ollama.com
2. Pull a model:

   ollama pull qwen2.5-coder:7b

3. Verify it's running:

   curl http://localhost:11434
   # → "Ollama is running"

4. In the app: Settings → Add Provider → Ollama → URL: http://localhost:11434 → Test → Save

Option B: OpenRouter / OpenAI / Groq (cloud)

1. Get an API key from your provider:
   • OpenRouter: openrouter.ai/keys
   • OpenAI: platform.openai.com/api-keys
2. In the app: Settings → Add Provider → OpenAI-compatible
   • Base URL: e.g. https://openrouter.ai/api/v1 or https://api.openai.com/v1
   • API Key: paste your key
   • Click Fetch Models → select a model
   • Test → Save

Option C: Ollama Cloud

• Kind: Ollama
• URL: https://ollama.com (or per Ollama Cloud docs)
• API Key: from Ollama Cloud dashboard

Switching Providers

Multiple providers can be configured. Select the active one via the radio button in Settings. The app uses whichever is active for all generate operations.

---

Usage

Daily Workflow

1. Add a repository — click the folder icon in the sidebar, select a folder containing .git/
2. Stage changes — use git add in your terminal or IDE
3. Generate message — click Generate; the AI streams a Conventional Commit message
4. Review & commit — edit if needed, then click Commit (or Copy to paste elsewhere)
5. Push — click Push in the header (auto-pulls with rebase if behind remote)

Other Operations

Action	How
Fetch remote updates	Click Fetch in the header
Pull changes	Click Pull (optional rebase toggle)
Switch branch	Use the branch dropdown in the header
Create branch	Branch dropdown → + New branch
View history	Click the History tab
Manage submodules	Expand the submodule tree in the sidebar

---

Tech Stack

Layer	Technology
Desktop runtime	Tauri v2 (Rust)
Frontend	React 19 + TypeScript 5.8 + Vite 7
Styling	Tailwind CSS 3
Icons	Lucide React
LLM communication	HTTP streaming (Ollama NDJSON / OpenAI SSE) via Rust reqwest
Persistent storage	tauri-plugin-store

---

Project Structure

git-cmm/
├── src/                    # React frontend (TypeScript)
│   ├── App.tsx             # Root component
│   ├── components/         # UI components (Sidebar, MainPanel, Settings, etc.)
│   ├── hooks/              # Custom React hooks
│   ├── lib/                # Utility functions
│   ├── types/              # TypeScript type definitions
│   └── styles/             # CSS / Tailwind
├── src-tauri/              # Tauri / Rust backend
│   ├── src/
│   │   ├── main.rs         # Entry point
│   │   ├── lib.rs          # Plugin registration & command handler
│   │   └── commands/       # Tauri commands
│   │       ├── git.rs      # Git operations (status, commit, push, pull, etc.)
│   │       └── llm.rs      # LLM streaming & model listing
│   ├── Cargo.toml          # Rust dependencies
│   └── tauri.conf.json     # Tauri app configuration
├── public/                 # Static assets
├── docs/                   # Documentation & screenshots
├── plans/                  # Feature implementation plans
├── package.json            # Node.js dependencies & scripts
├── vite.config.ts          # Vite configuration
├── tailwind.config.js      # Tailwind configuration
└── tsconfig.json           # TypeScript configuration

---

CI / Releases

A GitHub Actions workflow (.github/workflows/release.yml) builds for macOS (aarch64 + x86_64) and Windows (x64) on every v* tag push.

# Create a release
git tag v0.2.0
git push origin v0.2.0

You can also trigger manually from Actions → Release → Run workflow on GitHub.

---

License

MIT

---

Acknowledgments

• Conventional Commits specification
• Tauri team for an excellent cross-platform toolkit
• Ollama for making local LLMs accessible