simplechatgpt 1.0

An extremely simple interface to ChatGPT

Simple ChatGPT

An extremely simple programmatic interface to ChatGPT. The conversation automatically gets truncated (from the oldest message) if the token limit is reached.

The goal of this project is to remove the need for boilerplate code to create a chat bot in your program, allowing you to jump right into integrating ChatGPT into whatever you are making.

You can use this package like:

from simplechatgpt import Chat

import os

chat = Chat(os.environ["OPENAI_API_KEY"])

while True:
    prompt = input("You: ")
    print("Bot:", chat.send(prompt))

You can choose a model when initializing your Chat object like:

chat = Chat(os.environ["OPENAI_API_KEY"], model="gpt-4")

As of the time of this writing (07/08/2023), the following models are available for use:

• gpt-4
• gpt-4-32k
• gpt-3.5-turbo
• gpt-3.5-turbo-16k

See https://platform.openai.com/docs/models/overview for a full list of available models.

You can also customize the "system" message to initialize the chat bot to act in a certain way throughout the conversation. For example,

key = os.environ["OPENAI_API_KEY"]
chat = Chat(key, chatbot_description="You are Hucklyberry Finn.")

while True:
    prompt = input("You: ")
    print("Bot:", chat.send(prompt))

Would produce responses like:

You: Hey son, how are you?
Bot: Well, I reckon I'm doin' alright Pa. How 'bout yourself?