komit 0.5.1

AI-powered git commit message generator using local LLMs

🤖 komit

AI-powered git commit message generator using local LLMs via Ollama. No API keys, no internet required — runs completely locally and privately.

---

---

---

Features

• 🧠 Local LLM — uses Ollama, no API keys needed
• 🌿 Branch-Aware Context — infers conventional types and scopes directly from your active git branch
• 📝 Multiple commit styles — conventional, simple, detailed
• 🔍 Smart diff allocation — per-file context budget, code files prioritized over docs
• 🔎 Deep mode — summarize each file separately for better messages on large changesets
• 💡 Explain mode — summarize what changed without committing
• 🕐 Recent commits context — last 3 commits included in the prompt for better inference
• 🔄 Regenerate — not happy? generate a new message instantly
• ✏️ Edit before commit — open your editor to tweak the message
• ⚙️ CLI flags — control style, model, and more from the command line
• 📁 Config file — persist your preferences with komit init
• 🌍 Universal — works via pip, binary, or shell script
• 🪟 Cross-platform — Linux, macOS, Windows
• ⚡ Fast — runs on your machine, no network calls to external APIs
• 🔒 Ollama validation — checks if Ollama is running and model exists before generating
• 🔁 Smart updates — detects pip vs binary install and updates accordingly

---

Requirements

• Ollama installed and running
• A local model pulled (e.g. ollama pull qwen2.5:7b)

---

Recommended Models

Model	Size	Best for
qwen2.5:7b	4.7GB	Best quality
mistral:7b	4.1GB	Good balance
llama3.2:3b	2.0GB	Fastest

ollama pull qwen2.5:7b

---

Installation

Option 1 — One line install (Linux/macOS, no Python required)

curl -fsSL https://raw.githubusercontent.com/glemiu6/komit/master/scripts/install.sh | bash

Option 2 — Windows PowerShell

irm https://raw.githubusercontent.com/glemiu6/komit/master/scripts/install.ps1 | iex

Option 3 — pip

pip install komit

Option 4 — pipx (recommended for CLI tools)

pipx install komit

Option 5 — Download binary

Download the binary for your platform from GitHub Releases:

Platform	Binary
Linux x86_64	komit-linux-x86_64
macOS Apple Silicon	komit-macos-arm64
Windows	komit-windows-x86_64.exe

Intel Mac users: use pip install komit instead.

# Linux/macOS
chmod +x komit-*
sudo mv komit-* /usr/local/bin/komit

---

Setup

Git alias (recommended)

git config --global alias.ai '!komit'

Now you can use git ai as a shortcut.

Shell completion

# bash
echo 'eval "$(register-python-argcomplete komit)"' >> ~/.bashrc
source ~/.bashrc

# zsh
echo 'eval "$(register-python-argcomplete komit)"' >> ~/.zshrc
source ~/.zshrc

---

Usage

# stage your changes
git add .

# generate commit message
komit

# or via git alias
git ai

Subcommands & Maintenance

komit init       # create or overwrite your config file
komit update     # update to the latest version
komit uninstall  # remove komit from your system

CLI flags

Flag	Short	Type	Default	Description
--style	-s	conventional, simple, detailed	config	Commit message format
--model	-m	String	config	Ollama model to use
--ollama-url	-u	URL	config	Ollama API endpoint
--max_diff	—	Integer	config	Max diff length in characters
--timeout	—	Integer	config	LLM request timeout in seconds
--include_branch_name	-ib	Bool	config	Append branch name to commit message
--dry-run	-dr	Flag	False	Preview message without committing
--explain	—	Flag	False	Explain staged changes without committing
--deep	—	Flag	False	Summarize each file separately (slower, better on large diffs)
--config	—	Path	—	Path to a custom config file

Examples

# simple style
komit --style simple

# use a faster model
komit --model llama3.2:3b

# preview without committing
komit --dry-run

# explain what changed without committing
komit --explain

# better messages on large diffs (slower)
komit --deep

# include branch name in the message
komit --include_branch_name True

# remote Ollama instance
komit --ollama-url http://192.168.1.10:11434

---

Interactive Experience

╭──────────────────────── Staged files (3) ────────────────────────╮
│  • src/auth.py                                                    │
│  • tests/test_auth.py                                             │
│  • README.md                                                      │
╰───────────────────────────────────────────────────────────────────╯
Branch name: feat/auth-tokens
Model: qwen2.5:7b · Style: conventional

⠋ Generating commit message...

╭──────────────────── Suggested commit message ─────────────────────╮
│ feat: add JWT authentication token engine [feat/auth-tokens]      │
╰───────────────────────────────────────────────────────────────────╯

» Choose an action: (y)es, (n)o, (e)dit, (r)egenerate (y):

Option	Key	Behavior
Yes	y	Commit with the generated message
No	n	Cancel without committing
Edit	e	Open your editor to modify the message
Regenerate	r	Generate a new message

---

Deep Mode

For large changesets with many files, --deep summarizes each file individually before generating the final commit message. More accurate, but slower since it makes one LLM call per file.

komit --deep

---

Explain Mode

Explains your staged changes in plain English without committing — useful for code review prep or PR descriptions.

komit --explain

---

Branch Name Parsing

When --include_branch_name is enabled, komit reads your branch name to infer the commit type and scope automatically.

Branch	Inferred type	Inferred scope
feat/auth-login	feat	auth
fix/ui_button	fix	ui_button
refactor/api-layer	refactor	api

Supported types: feat, fix, chore, docs, style, refactor, test, ci, perf.

---

Configuration

Run komit init to create your config file at ~/.config/komit/config.toml:

model = "qwen2.5:7b"
style = "conventional"
ollama_url = "http://localhost:11434"
max_diff_length = 4000
timeout = 60
include_branch_name = false

CLI flags always override config file values.

---

Commit Style Outputs

Conventional (default)

feat: add user authentication
fix: resolve null pointer in response pipeline

Simple

Add user authentication
Fix null pointer in response pipeline

Detailed

feat: add user authentication

- Add JWT token generation pipeline
- Implement password hashing with bcrypt
- Add login and logout endpoints

---

Maintenance

komit update     # update to latest version
komit uninstall  # remove komit completely

---

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feat/my-cool-feature)
3. Commit your changes (komit 😉)
4. Push to the branch (git push origin feat/my-cool-feature)
5. Open a Pull Request

---

License

Apache License 2.0