This project is a combination of Rust and Python applications that work together to provide a service. The service allows users to record audio, transcribe it into text, and then generate a shopping list item from the transcribed text.

1. The Rust application records audio from the user's microphone for a specified duration and saves it as a .wav file.
2. The Rust application sends a request to the Python application, passing the name of the .wav file as a parameter.
3. The Python application receives the file name, loads the file from disk, and transcribes the audio into text using the Whisper ASR model from OpenAI.
4. The transcribed text is then passed to a language model, which generates a new shopping list item.
5. The new shopping list item is returned to the Rust application, which then adds the item to the shopping list using the bring command.

This project only works on Windows. You will need to install my bring cli tool to run this project. You can find it here.

I am currently working on distributing part of the CLI as a library so that it can be used on other platforms and it doesn't have to be called as a subprocess.

To run the project, you need to start both the Rust and Python applications. The Python application should be started first, as it needs to be running when the Rust application sends the request.

To start the Python application, navigate to the Python directory and run the following command:

pip install -r requirements.txt
python python/main.py

To start the Rust application, navigate to the Rust directory and run the following command:

cargo run

The project has several dependencies, which are listed in the Cargo.lock file for the Rust application and in the requirements.txt file for the Python application. The dependencies include several crates and Python packages for handling audio, making HTTP requests, and working with machine learning models.

1. Export the functionality of the Bring! Client to Python so that it can be used on other platforms.
2. Scrap the Rust application and use Python for the whole project.
3. Add a Wakeword detection model to the project.