Installation

Installation

Requirements

• macOS with Apple Silicon (M1/M2/M3/M4)
• Python 3.10+
• MLX 0.11+
• FFmpeg (for audio loading)

Install from PyPI

# Using uv (recommended)
uv pip install vayu-whisper

# Using pip
pip install vayu-whisper

Install from Source

git clone https://github.com/CodeWithBehnam/vayu.git
cd vayu

# Using uv (recommended)
uv pip install -e .

# Using pip
pip install -e .

Development Setup

git clone https://github.com/CodeWithBehnam/vayu.git
cd vayu

# Using uv (recommended)
uv pip install -e ".[dev]"

# Using pip
pip install -e ".[dev]"

This installs additional tools:

• pytest — testing
• black — code formatting
• isort — import sorting
• mypy — type checking

Notebook Support

For Jupyter notebook integration:

# Using uv
uv pip install vayu-whisper[notebook]

# Using pip
pip install vayu-whisper[notebook]

Dependencies

Package	Purpose
mlx>=0.11	Apple MLX framework for accelerated inference
numpy	Numerical computing
tqdm	Progress bars
tiktoken	OpenAI tokenizer
huggingface_hub	Model downloading from HuggingFace
numba	JIT compilation for word-level timestamps
scipy	Signal processing (DTW alignment)

Verifying Installation

from whisper_mlx import LightningWhisperMLX

whisper = LightningWhisperMLX(model="tiny", batch_size=12)
print("Vayu installed successfully!")

Or from the CLI:

vayu --help