OpenAI Chatbot with gpt-3.5-turbo model

This is a simple chatbot that uses OpenAI's GPT-3.5-turbo language model to generate responses to user input.

Installation

Can be installed on most Linux terminal and also Termux on android

1. Clone the repository: https://github.com/jacobhamilton12/chatgpt-on-terminal.git
2. Install the required packages: pip install -r requirements.txt

Note: This also works on termux on Android. If you have trouble installing dependencies on termux, read this termux/termux-packages#13803

3. Set up an OpenAI API key by following the instructions here
4. Add your API key as an environment variable export OPENAI_KEY=<Your key here>. Add this line to $PREFIX/etc/bash.bashrc to have it persist. Be sure to source after editing: source $PREFIX/etc/bash.bashrc

Note: Do not include the greater than less than signs.

Usage

To start the chatbot, run main.py using Python 3:

python main.py

Note: I've also added this as an alias to my bashrc alias chat="python ~/chatgpt-on-terminal/main.py"

The chatbot will prompt you to enter your input, and then it will generate a response using the GPT-3 model. The conversation history is stored in a list of dictionaries called message_log.

Promptbook

You can store system prompts in promptbook.json and select the prompt like so: python main.py "Default" replace Default with whatever key you made.

Format for promptbook.json

{
    "key": "prompt",
    "Default": "You are a helpful assistant",
    "Random": "Assistant is to respond to each prompt only with obscure incorrect facts"
}

To end the chatbot, type "q" at any time. Or ctrl+c

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• This chatbot was inspired by the OpenAI GPT-3 Playground
• The GPT-3 model is provided by OpenAI