A tiny GPT CLI tool. You can chat with the GPT model developped by OpenAI and save the conversation as json.

Assuming you have an environment variable with key named OPENAI_API_KEY. If you don't have a OpenAI API key visit here, generate one and add it as an environment variable

export OPENAI_API_KEY=<YOUR-OPENAI-API-KEY>

You can install the package using pip:

$ pip install oregpt

Once you have installed oregpt, you can run it by typing:

$ oregpt

There are a few options:

$ oregpt --help
Usage: oregpt [OPTIONS]

Options:
  -V, -v, --version          Show the version and exit.
  -m, --model_name TEXT      Model name in OpenAI (e.g, gpt-3.5-turbo, gpt-4)
  -a, --assistant_role TEXT  Role setting for Assistant (AI)
  --help                     Show this message and exit.

Commands such as saving and loading conversations are available as the following:

You can specify the place of conversation log, style (color etc) and the model supported in /v1/chat/completions endpoint provided by OpenAI in ~/.config/oregpt/config.yml

❯ cat ~/.config/oregpt/config.yml
log: /tmp/oregpt/
openai:
    model: gpt-3.5-turbo
# You can also specify OpenAI's API key here
#     api_key: <your-api-key>
character:
    user:
        name: Me
        style: "#00BEFE"
    assistant:
        name: AI
        style: "#87CEEB"
        role: "You are a chat bot"
    system:
        name: System
        style: "#cc0000"