This project is currently only tested on Linux, however it might run in other OS's ¯\(ツ)/¯

for chatgpt, set the OPENAI_API_KEY environment variable, for example:

export OPENAI_API_KEY="sk-the-rest-of-the-key"

You can find or create your key at the OpenAI website

The project expects your desired microphone to be the default one.

conda env create -f environent.yml

Run the project:

conda activate py-intercom
python main.py

env_name=".venv"
python -m venv $env_name
source $env_name/bin/activate # Will have to be run one every new shell

Install the python dependencies:

$env_name/bin/pip install -r requirments.txt

Run the project:

$env_name/bin/python main.py

A simple autorun script could be written like the following:

#!/bin/bash
cd py_intercom_dir
export OPENAI_API_KEY="sk-the-rest-of-the-key"
# In the case conda does not recognize the activate command, uncomment the below line
# eval "$(conda shell.bash hook)"
conda activate py-intercom
python main.py

After completing the setup, you can run the program.

The program will only respond when you call it by it's name (default - Intercom / "אינטרקום")

The program has a few commands that will call their own code, anything else will be sent directly to ChatGPT

The program can take as arguments a config file, and commands index, by default located in the project as config.yml and commands.json

python main.py -c my_config_file.yml -C my_commands_index.json