This is a Python voice assistant that uses both OpenAI's 3.5 turbo model and Bing's model with access to the internet for searchable results. All speech-to-text transcription is done locally using OpenAI's Whisper running on the local machine, and text-to-speech is done with AWS Polly.

Note: You will need to have access to the Bing AI search function to utilize this feature.

• Setup
  • Quick Setup

The setup for this project is relatively simple and has a detailed youtube tutorial, which briefly explains how the program works as well as providing a detailed installation guide.

demonstration

1. install ffmpeg from their website, or by running sudo apt update && sudo apt install ffmpeg on Linux

2. Click the green Code button then Download ZIP. After downloading, unzip the file and copy the path to the folder that has the main.py file in it. Alternatively run git clone https://github.com/Ai-Austin/Bing-GPT-Voice-Assistant.git if you have installed git on your device,

3. To set up the AI's for response generation, use the steps outlined from the video to setup the cookies.json file and follow the steps for creating an OpenAI key and putting it in the program where prompted.

4. Install python if it is not already on your system. Scroll down to files and download and install the version compatible for your computer. Windows x86-64 executable installer should work for most windows users (note: The link is to the 3.10.0 version, the latest release 3.11 is not compatible with Whisper)

5. Open a new terminal window and run the command cd path\you\copied using the path from step 1. Then run pip install -r requirements.txt to install the required packages

Simply open a terminal and move to the folder with the code cd path\you\copied or use the terminal from step 5. Then run python main.py to start the assitant. To exit the assistant, press ctrl+c to stop the program (you may need to press it twice if it doesn't shut down).