This project aims to provide a voice assistant that executes your commands to control Home Assistant.
It uses the KaldiRecognizer from Vosk to perform speech recognition.
A simple text-based lookup parser then parses the voice command. If it fails to find an entity and a state (on, off) a gpt-3.5-turbo model is asked for some function_calls.
Have a Home Assistant server running with API and HTTP enabled.
- Clone this Repo
- Run
pip3 install -r requirements.txt
- Create an openai API key and set it as an environment variable
- Create a config.ini and populate it
- Run
python3 main.py
- Say the assistant's activation word
- Say voice command