We are going to apply custom styles and behaviors to our chatbot using HTML, TailwindCSS, and JavaScript.
First, create a new JavaScript file named template.js in the src directory. This file will serve as the HTML template for our chatbot's user interface.
export const html = ``;You will need to import this template into your index.js file like so:
import { html } from './template.js';Next, let's modify our server's routing logic. We want to show a chat interface when the user accesses the root path /, and we will set up a separate /ai path for handling message exchange with the LLM.
const url = new URL(request.url);
if (request.method === 'POST' && url.pathname === '/ai') {
// rest of the logic goes here
// ...
}We will wrap the existing logic in a new conditional statement that checks for a POST request to the /ai endpoint.
Then, we will send the HTML content from template.js to the browser as a response.
return new Response(html, {
headers: {
'content-type': 'text/html;charset=UTF-8',
},
});With these changes, visiting / will display an empty page, while /ai will show the output that we previously configured.
Now, let's update template.js with an HTML boilerplate for our chatbot:
export const html = `
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>ChatGPT Clone App</title>
<link rel="icon" type="image/x-icon" href="https://openmoji.org/data/color/svg/1F913.svg" />
<link href="https://cdn.jsdelivr.net/npm/tailwindcss@2.2.19/dist/tailwind.min.css" rel="stylesheet" />
</head>
<body class="bg-gray-700 text-white h-screen overflow-hidden flex">
<div class="bg-gray-900 w-1/5 p-6 overflow-auto">
<h1 class="text-2xl font-bold">ChatGPT Clone App</h1>
<p class="mt-2 text-gray-400 text-lg">
Welcome! This app is your personal assistant that is available 24/7 to help with your questions.
</p>
</div>
<div class="flex-1 flex flex-col">
<div id="message-container" class="flex-1 max-w-full h-full overflow-y-auto"></div>
<div id="input-container" class="border-t border-gray-600 p-4 flex">
<textarea
placeholder="Type your message"
rows="1"
class="flex-1 bg-gray-700 max-h-[200px] resize-none rounded p-2 text-white border-none focus:ring-2 text-lg focus:ring-blue-500 outline-none"
style="height: 50px"
></textarea>
<button class="bg-blue-500 hover:bg-blue-600 text-white rounded px-4 ml-4">
<svg
stroke="currentColor"
fill="currentColor"
stroke-width="0"
viewBox="0 0 24 24"
height="1em"
width="1em"
xmlns="http://www.w3.org/2000/svg"
>
<path fill="none" d="M0 0h24v24H0z"></path>
<path d="M2.01 21L23 12 2.01 3 2 10l15 2-15 2z"></path>
</svg>
</button>
</div>
</div>
</body>
</html>
`This template serves as a starting point for your chat interface and can be customized to your liking. To make our chatbot interactive, let's add JavaScript functionality within a <script> tag at the bottom of the template:
<script>
document.addEventListener('DOMContentLoaded', function () {
const messageContainer = document.getElementById('message-container');
const inputContainer = document.getElementById('input-container');
const textBox = inputContainer.querySelector('textarea');
function appendUserMessage(text) {
const messageDiv = document.createElement('div');
messageDiv.className = 'bg-gray-800 text-white text-xl p-16 mx-auto';
messageDiv.textContent = text;
messageContainer.appendChild(messageDiv);
messageContainer.scrollTop = messageContainer.scrollHeight;
}
function appendSystemMessage(text) {
const messageDiv = document.createElement('div');
messageDiv.className = 'bg-gray-700 text-white text-xl p-16 mx-auto';
messageDiv.textContent = text;
messageContainer.appendChild(messageDiv);
messageContainer.scrollTop = messageContainer.scrollHeight;
}
function handleMessageSend() {
const message = textBox.value.trim();
if (message) {
appendUserMessage(message);
sendMessage(message);
textBox.value = '';
}
}
inputContainer.querySelector('button').addEventListener('click', handleMessageSend);
textBox.addEventListener('keypress', function (e) {
if (e.key === 'Enter') {
e.preventDefault();
handleMessageSend();
}
});
async function sendMessage(message) {
try {
const response = await fetch('/ai', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({ message }),
});
if (!response.ok) {
throw new Error('Failed to send message');
}
const data = await response.json();
appendSystemMessage(data.response);
textBox.value = '';
} catch (error) {
appendSystemMessage(error.message);
}
}
});
</script>The provided JavaScript code handles user input, message sending, and displaying responses in the chat interface. Here's a breakdown of its functionality:
- DOMContentLoaded event setup
- Message container and input box selection
- Added a
appendUserMessageandappendSystemMessageFunctions to append messages to the chat - Event listeners for message sending
- The
sendMessagefunction that handles the fetch request to the/aiendpoint and processes the response
And now in your index.js file add the following code:
const body = await request.json();
const userMessage = body.message;
const messages = [
{ role: 'system', content: 'You are an friendly assistant.' },
{ role: 'user', content: userMessage },
]
const message = await ai.run('@cf/meta/llama-2-7b-chat-int8', {
messages,
});After integrating the template, launch the application with the following command:
wrangler dev --remoteYou should now see a chat interface where you can type messages and view responses.
Congratulations! You've successfully built a ChatGPT clone using Cloudflare Workers. The next step is to deploy your application and have it ready to play with.