reterm-ai 0.1.0

AI-powered terminal command recommender based on your shell history

 ___   ___   _____   ___   ___   __ __    __    _
| _ \ | __| |_   _| | __| | _ \ |  V  |  /  \  | |
| v / | _|    | |   | _|  | v / | \_/ | | /\ | | |
|_|_\ |___|   |_|   |___| |_|_\ |_| |_| |_||_| |_|

[reTermAI] Smart command assistant for your terminal 🧠💻

---

reTermAI

💡 A terminal command recommender powered by AI and your own shell history.

---

✨ Features

• 🔍 Recommends relevant terminal commands based on your past history using OpenAI or Gemini
• 🧠 Supports intelligent matching by keyword or partial input
• ⚡ Easy to install via pip
• 🐚 Supports zsh and bash shell history
• 🔐 API keys managed via .env

---

📦 Installation

pip install reterm-ai

Or, for local development:

git clone https://github.com/pie0902/reTermAI.git
cd reTermAI
pip install -e .

---

⚙️ Usage

🔮 AI-powered command suggestions:

reterm suggest

Options:

Option	Description
--history-limit, -hl	Number of recent commands to read (default: 300)
--context-limit, -cl	How many commands to feed into the LLM (default: 20)
--provider, -p	LLM to use (openai or gemini)

Examples:

reterm suggest -p gemini
reterm suggest --history-limit 500 --context-limit 30

---

🔎 Match past commands by keyword:

reterm match docker

Options:

Option	Description
--limit, -l	Max number of matching results to show (default: 5)

---

🔐 Configuration

Add a .env file in your home or project directory:

OPENAI_API_KEY=sk-xxxxxxxxxxxxxxxxxxxxx
GOOGLE_API_KEY=your-gemini-api-key

Key	Used for
OPENAI_API_KEY	Required if using --provider openai
GOOGLE_API_KEY	Required if using --provider gemini

You can also refer to the included .env.example.

---

📂 Project Structure

reterm/
├── cli.py          # Main CLI interface
├── llm.py          # LLM integration (OpenAI, Gemini)
├── shell.py        # Shell history detection and parsing
├── config.py       # API key loader
├── welcome.py      # ASCII and help panel

---

🧑‍💻 Contributing

We welcome contributions! Fork the repo, create a feature branch, and open a pull request.

git checkout -b feat/your-feature

Optional ideas:

• Add support for fish shell
• Add model options (e.g. gpt-4o, gemini-pro)
• Improve LLM prompt formatting

---

📄 License

Apache License 2.0