Let’s write some code!

We’re going to build a simple chat app that you can use in the terminal.

1. Imports:

import 'dotenv/config'
import readline from 'node:readline'
import OpenAI from 'openai'

• The code imports necessary modules.
  • dotenv/config: This loads environment variables from a .env file.
  • readline: Provides an interface for reading data from a readable stream.
  • openai: The SDK for interacting with the OpenAI API.

1. Initialize OpenAI API:

const openai = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY,
})

• An instance of the OpenAI SDK is created using the API key loaded from the environment variables.

1. Setting up User Interface:

const rl = readline.createInterface({
  input: process.stdin,
  output: process.stdout,
})

• readline is set up to read input from the terminal (stdin) and output back to it (stdout).

1. Message Handling:

const newMessage = async (history, message) => {
  const chatCompletion = await openai.chat.completions.create({
    messages: [...history, message],
    model: 'gpt-3.5-turbo',
  })

  return chatCompletion.choices[0].message
}

• This function sends the chat history and the latest user message to OpenAI, gets the AI's response, and returns it.

1. Formatting User Message:

const formatMessage = (userInput) => ({ role: 'user', content: userInput })

• Formats the user's input into the required structure for OpenAI.

1. Main Chat Function:

const chat = () => {
  //...
}