Commander is a lightweight, fully local fine-tuned Large Language Model (LLM)-powered terminal assistant. Designed for Unix systems, it lets you generate, explain, and run terminal commands through natural language — all from your terminal and without any need for cloud services or internet access.

Think of it as your AI command-line co-pilot for linux commands — private, fast, and efficient.

• Local LLM-Powered: Works entirely offline — your data stays on your machine.
• Interactive Shell: Start a session where you can chat with the LLM to generate commands.
• Quick Prompt Mode: Run one-off prompts directly from the command line.
• Lightweight: Minimal dependencies and fast startup.
• Unix-Focused: Tailored to understand and generate Unix-based commands.

You can install Commander locally using pip:

git clone https://github.com/sszxt/commander.git
cd commander
pip install .

Interactive Mode

commander

• Start an AI-powered interactive shell
• Then, start chatting with your LLM assistant to get help, generate commands, or run them directly.

Example:

> How do I find the largest file in a directory?
> find . -type f -exec du -h {} + | sort -rh | head -n 1

One - Line Prompt Mode

commander "how to list all open ports"

• Skips the interactive shell and get instant suggestions for a one-line prompt

Example :

commander "how to list all open ports"
assistant : sudo lsof -i -P -n | grep LISTEN

commander/
├── commander/              # Source code and model
│   ├── __main__.py         # CLI entry point
│   └── model/
│       └── checkpoint-750/ # Trained Local model files
├── build/                  # Build artifacts
├── requirements.txt        # Python dependencies
├── setup.py                # Installer
└── readme.md               # You're reading this right now.

Commander runs entirely locally. No commands, prompts, or data are sent to any server. It's your assistant, and your business stays yours.

Commander leverages powerful open-source LLMs and tools, and I welcome your contributions — feel free to open an issue or pull request for bug reports, suggestions, or new features!

Developed by Mohamed Sameer.