WhispyWyser 0.0.23

The WhispyWyser project is designed to provide a flexible and efficient implementation of Voice Assistant for Home Assistant based on Faster Whisper and Wyoming protocol. This project supports both CPU and CUDA architectures, allowing for optimized performance based on the available hardware.

WhispyWyser - Voice Assistant for Home Assistant

WhispyWyser is a flexible and efficient voice assistant for Home Assistant, leveraging Faster Whisper for speech recognition and the Wyoming protocol for seamless integration. It supports both CPU and CUDA architectures for optimized performance.

Features

• 🎙️ High-accuracy speech-to-text with Faster Whisper
• 🏠 Deep Home Assistant integration
• 🐳 Docker support for easy deployment
• 🚀 GPU acceleration support (CUDA)
• 🔌 Wyoming protocol for extensibility
• 🔒 Secure authentication with long-lived tokens

Prerequisites

• Docker and Docker Compose
• Home Assistant instance (local or remote)
• NVIDIA GPU with drivers (for CUDA support)

Quick Start

1. Clone the repository

   git clone https://github.com/yourusername/whispywyser.git
   cd whispywyser
   

2. Configure environment

   cp .env.example .env
   # Edit .env with your configuration
   nano .env
   

3. Start the service

   # For CPU
   docker-compose up -d whispywyser-cpu
   
   # For GPU (CUDA)
   docker-compose up -d whispywyser-gpu
   

Configuration

Environment Variables

Variable	Required	Description
HA_TOKEN	✅	Home Assistant long-lived access token
HOME_ASSISTANT_URL	✅	URL of your Home Assistant instance
HF_TOKEN	❌	Hugging Face Hub token (for private models)
MODEL	❌	Model name (default: openai/whisper-large-v3)
DEVICE	❌	cuda or cpu (auto-detected)
LANGUAGE	❌	Language code (e.g., en, de)

Generating a Long-Lived Access Token

1. In Home Assistant, click on your profile
2. Scroll down to "Long-Lived Access Tokens"
3. Click "Create Token"
4. Copy the token and add it to your .env file

Docker Compose

The project includes two services:

1. whispywyser-gpu: For systems with NVIDIA GPU
2. whispywyser-cpu: For CPU-only systems

Volumes

• /data: Stores models and configuration

Ports

• 10300: Wyoming protocol port

Development

Building the Image

# For CPU
docker-compose build whispywyser-cpu

# For GPU
docker-compose build whispywyser-gpu

Running Tests

docker-compose run whispywyser-cpu python -m pytest

Troubleshooting

Common Issues

1. CUDA errors

   • Ensure NVIDIA drivers are installed
   • Run nvidia-smi to verify GPU detection
   • Check Docker has access to GPU: docker run --gpus all nvidia/cuda:11.0-base nvidia-smi

2. Connection to Home Assistant

   • Verify HOME_ASSISTANT_URL is correct
   • Check if the token has the right permissions
   • Ensure Home Assistant is accessible from the container

License

MIT

Acknowledgements

• Faster Whisper
• Wyoming Protocol
• Home Assistant