Learning Generative AI

Make ChatGPT API As A More Programmable Interface

Have ChatGPT API be a more meaningful interface for developers

Photo by Emiliano Vittoriosi on Unsplash

In my previous share, we can understand the basics of ChatGPT API. In a simplified view, it looks like something below. A chatbot.

In this article, I will expand on how it can be more useful in producing developer-interested interfaces, which eventually turns it into something below

Pre-requisite

The blog here assumes you have set up your ChatGPT API Key as below

import openai
import os

from dotenv import load_dotenv, find_dotenv
_ = load_dotenv(find_dotenv())

openai.api_key  = os.getenv('OPENAI_API_KEY')

And with two functions, one accepting a prompt, and the other a full messages

def get_completion(prompt, model="gpt-3.5-turbo", temperature=0):
    messages = [{"role": "user", "content": prompt}]…