#git-commit #ollama #artificial-intelligence #git #commit #cli

app gcomm

Generate AI-powered Git commit messages from staged changes using a local Ollama model

8 releases

new 0.2.2 Mar 29, 2025
0.2.1 Mar 29, 2025
0.1.4 Mar 29, 2025

#7 in #ollama

MIT license

17KB
164 lines

gcomm

AI-powered Git commit message generator using your local Ollama model.

This CLI tool analyzes staged Git changes and generates a structured commit message using a locally running LLM (e.g. gemma3, llama3, etc.).


🚀 Getting Started

Install via cargo

cargo install gcomm

Alternative: Build from source

1. Install Rust (if you haven't yet)

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

2. Clone and build the project

git clone https://github.com/bryson-tang/gcomm
cd gcomm
cargo build

3. (Optional) Install globally

cargo install --path .

Or add a shortcut:

alias gcomm="./target/debug/gcomm"

Now you can use gcomm from anywhere.

🧠 Usage

This uses your staged changes only (like git diff --cached), so make sure you've staged what you want:

git add .

Generate a commit message (copies to clipboard):

gcomm

Specify a different Ollama model (default is gemma3:latest):

gcomm --model codellama

Run git add . first:

gcomm --add

Generate, edit in your default editor, and commit:

gcomm --edit

Generate and commit immediately without editing:

gcomm --yolo

🧩 Requirements

  • Ollama running locally
  • A model pulled (e.g. ollama run gemma3 or ollama run llama3)

✨ Output Format

Add analytics and fix layout issues

Added

- Event tracking for login

Changed

- Layout of settings page
- Refactored token logic

Removed

- Unused `tracking.js`

📄 License

MIT

💖 Acknowledgements

Built with a lot of love and help from ChatGPT.

Dependencies

~5–17MB
~238K SLoC