Before installing, make sure Python 3.11 and git are installed
Download windows.zip from last release
Drop setup.bat in folder - script will make a folder with project inside the choosen folder
Drop Start.bat in project's folder and run it
git clone https://github.com/voidmolly/discord-ChatGPT.git
To install requirements use $ pip install -r requirements.txt in project's directory
Set up daemon:
sudo nano /etc/systemd/system/dchatgpt.service
[Unit]
Description=Discord ChatGPT
After=syslog.target
[Service]
Type=simple
User=root
Group=sudo
WorkingDirectory=/root/discord-ChatGPT/
ExecStart=/usr/bin/python3.11 main.py
[Install]
WantedBy=multi-user.target
Run: sudo systemctl start dchatgpt
Stop: sudo systemctl stop dchatgpt
Restart: sudo systemctl restart dchatgpt
Check logs: sudo systemctl status dchatgpt
Open config.json with any text editor
Then insert you openai and discord tokens in fields with such names
Optional:
1. You may change prefix in field named "prefix"
2. You may change max_tokens and temperature - ChatGPT's configs, but only if you know what you are doing
I use command hint to send info about commands in chat, because IDK how to make help command :)
Commands game, stream are made at the request of the customer. They are used for changing bot's status.
Permissions: Read Messages/View Channels and Send Messages
Intents: all
I use openai API to interact with ChatGPT. Actually, I have some problems with it, because sometimes ChatGPT sends a bit unexpected results when you use API.
In bot you can change the engine of openai Completion:
max_tokens is actually max number of symbols in ChatGPT's answer
temperature is something that you may change, but you shold read documentation before
You may change theese values in config.json
Also if you run bot in Russia, you should use server located in another country or use proxy, otherwise ChatGPT won't work.