Get started from scratch with a simplified API request, whether you’re launching a new project or testing the OpenAI API.
Before we jump in. 👇 Here’s the source code for your reference:
Start By Creating an open OpenAI account and getting your API key:
- Visit https://platform.openai.com/ and sign up for an account.
- Click on the account dropdown menu and choose ‘Manage Account’.
- Locate ‘API Keys’ and generate your key.
Setting Up Your Project:
Create your project folder and add the following files:
- package.json
- .env
- index.js
Here’s a snapshot of my folder…
Copy and paste the following code into your .env file, and replace YOUR_API_KEYS
with the key you created on your OpenAI account. 🔥
# Do not share your OpenAI API key with anyone! It should remain a secret.
OPENAI_API_KEY=YOUR_KEY
- Make sure to include a
.gitignore
file and add.env
to it if you're publishing to GitHub.
Now, copy and paste the following code into your package.json
file:
{
"name": "openai_bare_bones",
"version": "0.1.0",
"private": true,
"openai": "^4.0.0",
"type": "module",
"scripts": {
"start": "node index.js"
},
"dependencies": {
"dotenv": "^16.3.1",
"openai": "^3.3.0"
},
"engines": {
"node": ">=14.6.0"
}
}
Finally, copy and paste the following code into your index.js
file…
import { Configuration, OpenAIApi } from "openai"
import * as dotenv from 'dotenv'
dotenv.config()
const configuration = new Configuration({ apiKey: process.env.OPENAI_API_KEY })
const openai = new OpenAIApi(configuration)
const start = async function() {
const response = await openai.createCompletion({
model: "gpt-3.5-turbo-instruct",
prompt: "What chemical compounds are computers mostly made from?",
temperature: 0.6,
max_tokens: 256,
});
console.log('Response Text', response.data.choices[0].text)
}
start()
Time to Get It Running!
First install the node dependencies…🏗️
npm install
And, Run the code….🚀
npm start
You should now see a response for your prompt. If you didn’t change the code, it’ll be info about the chemical compounds contained in a computer. 🤓
What’s Happening Here? Let’s talk about the code
index.js
The meat of what we’re doing is in the request/response section of this file where we’re making a request to OpenAI:
const response = await openai.createCompletion({
model: "gpt-3.5-turbo-instruct",
prompt: "What chemical compounds are computers mostly made from?",
temperature: 0.6,
max_tokens: 256,
});
- The
temperature
parameter affects randomness. A higher value (closer to 1) will make outputs more random, whereas a lower value (closer to 0) will make outputs more deterministic. - The
max_tokens
parameter limits the length of the response to 256 tokens.
Package.json
- name: 📛 Whatever we decided to call our project “openai_bare_bones”.
- version: 🏷️ Keeps track of what version were on.
- private: 🤫 Set to
true
, which means npm will refuse to publish it. - openai: 🤖 version of Open AI we’re using.
- type: 📘 By setting this to “module”, we’re stepping into the modern ECMAScript module world. Fancy!
- scripts: 🚀 command-line shortcut!
- dependencies: 🧩 source code we need to include to run our application:
dotenv
: Environment variables gatekeeper, handling.env
file withprocess.env
.openai
: 🧠 Version "^3.3.0". openai version. - engines: 🚂 Make sure we’re on Node.js version “14.6.0” or a newer, model.