Metadata-Version: 2.4
Name: bayan-cli
Version: 1.3.1
Summary: Arabic coding assistant for your terminal — read, write, and refactor code with an AI that works directly in your project directory.
Requires-Python: >=3.8
Description-Content-Type: text/markdown
Provides-Extra: pc
Requires-Dist: pyautogui>=0.9.54; extra == "pc"
Requires-Dist: mss>=9.0.1; extra == "pc"
Requires-Dist: Pillow>=10.0.0; extra == "pc"
Requires-Dist: pytesseract>=0.3.10; extra == "pc"
Requires-Dist: pyperclip>=1.8.2; extra == "pc"

# Bayan CLI

Arabic coding assistant with agent tools, for your terminal.

## Install

```bash
pipx install bayan-cli
```

## Usage

```bash
bayan "explain decorators"         # single question, then exit
cat main.py | bayan "review this"  # read from stdin
bayan -m bayan-pro                 # pick a model
bayan --yolo                       # full access, no confirmations
```

## PC control (optional) — `pip install "bayan-cli[pc]"`

Teach Bayan to see and use your desktop: `screenshot`, OCR clicks
(`click_text`), `type_text`, `press_key`, clipboard access, and
`open_app`. A vision-capable model is recommended.

OCR additionally needs the tesseract binary on your system:

- macOS: `brew install tesseract`
- Linux: `sudo apt install tesseract-ocr`
- Windows: download from https://github.com/UB-Mannheim/tesseract/wiki

With the extras installed:

```bash
bayan "look at my screen and click the save button"
```

PyAutoGUI's failsafe is left on — slam the mouse into a screen corner to
abort an out-of-control automation.

## Setup

First run will ask for your Bayan API key. Get one at
[bayan-beta.vercel.app/dashboard](https://bayan-beta.vercel.app/dashboard).

## Build & publish (maintainers)

```bash
# Build
pip install build twine
python -m build

# Upload to TestPyPI
twine upload --repository testpypi dist/*

# Install from TestPyPI
pip install --index-url https://test.pypi.org/simple/ bayan-cli[pc]

# Then upload to real PyPI
twine upload dist/*
```

## License

MIT
