Getting Started with GPT-4 API and JavaScript

GPT-4 launched this week!

There's a waitlist for GPT-4 API access, but invites have already started going out. Once you get access, it's dead simple to get started building.

Here's a command line chatbot powered by node.js and the GPT-4 API. The only difference between this code and a Javascript bot powered by ChatGPT is that we've changed the model parameter to gpt-4.

const { Configuration, OpenAIApi } = require("openai");
const configuration = new Configuration({
     apiKey: process.env.OPENAI_API_KEY,
});
const openai = new OpenAIApi(configuration);

var rl = require('readline-sync');

async function startAssistant() {
   var messages = [];

   var input = rl.question('What type of chatbot would you like to create? > ');
   messages.push({"role": "system", "content": "Stay in character. You are " + input});

   input = rl.question("Say hello to your new chatbot > ");
   while(true) {
       messages.push({"role": "user", "content": input});

       console.log("...");
       const completion = await openai.createChatCompletion({
         model: "gpt-4",
         messages: messages,
       });

       var reply = completion.data.choices[0].message.content;
       messages.push({"role": "assistant", "content": reply});
       console.log(reply);
       input = rl.question("> ");
   }
}

startAssistant();

Here's what that code looks like in action:

If you want to run this code yourself there are three key things you’ll need to do:

1. Put the above code in a file code index.js
2. Use these instructions from OpenAI to save your OpenAI API key as an environment variable.
3. Install these 2 dependencies:

npm install openai
npm install readline-sync

Then you can run node index.js to interact with GPT-4 from the command line!

Have fun!