terma-cli 1.0.0

A lightweight CLI tool that converts natural language instructions into terminal commands.

Terma

A lightweight Python CLI tool that converts natural language instructions into terminal commands.

Overview

Terma lets you describe what you want to do in plain English, and your configured AI provider (OpenAI, Anthropic, or Google GenAI) will translate it into a valid terminal command.

Features

• Convert natural language directly into executable commands
• Multiple AI provider support:
  • OpenAI
  • Anthropic
  • Google GenAI
• AI agent configuration via terma config
• Persistent settings stored automatically at ~/.config/terma/config.json
• Simple CLI tool created with Python + Typer

Installation

Globally

You can install Terma through PyPI:

pip install terma

Locally

git clone https://github.com/ShlokShar/terma.git
cd terma
pip install -e .

Usage

Basic Syntax

terma <command> <args>

Configuration

Terma uses a simple config system to set your AI provider, API key, and model. Configuration is required prior to executing the natural language command.

Supported config commands

Command	Description
terma config provider <provider>	Set the AI provider (e.g., openai, anthropic, google) & API key
terma config provider <provider> --api-key <key>	Set provider and API key in one line
terma config api-key	Set the API key separately
terma config model <model-name>	Sets the model (e.g., gpt-4-nano, claude-haiku-4-5, etc.)
terma config peek	View Terma's current configuration

Natural Language Execution

Translate any natural language instruction into a real command:

terma exec <instruction>

Example Commands

Input:

terma exec check node version

Output:

node --version

---

Input:

terma exec make a new directory called logs and move all .txt files into it

Output:

mkdir logs && mv *.txt logs/

(Actual output may differ depending on the AI provider/model.)

Config File Path

Terma automatically stores configuration settings at:

~/.config/terma/config.json

You do not need to create this manually.

License

This project is licensed under the MIT License.

Contributing

Feel free to open issues, submit PRs, or propose features.

Roadmap

• Command history log for past AI-generated commands
• More provider-specific settings
• Optional safety/output confirmation mode

Author

Shlok Sharma
GitHub: @ShlokShar

Related Projects

• insectool — A lightweight Python web scanner for identifying common web vulnerabilities