Creating a Chatbot with JavaScript and Gemini AI

So, how are you?

I was looking at some projects on my github and came across a chatbot I recently created using Google Gemini. The idea was to create a language assistant, where you could talk to the AI ​​to improve your skills in the language you wanted.

So I thought: " Why not share how I did this project with everyone? ". And that's why I'm writing here, to show you how I did each part. So let's start with the front-end of the application.

Starting a new project

Well, to justify some actions I'm going to take in the project, I'm going to say right away that we're going to create a "server" with express.js, where we'll provide an api route '/chat' that will be used for communication between the front-end and the Gemini API.

So, we will need to start our project with the npm init -y command. The result is a package.json file that looks something like this:

{
  "name": "chatbot-ia",
  "version": "1.0.0",
  "main": "index.js",
  "scripts": {
    "test": "echo \"Error: no test specified\" && exit 1"
  },
  "keywords": [],
  "author": "",
  "license": "ISC",
  "description": ""
}

And also, we will need to organize our files as follows:

public
  |__ index.html
  |__ style.css
  |__ script.js
package.json 

Once that's done, let's create the visual part of our chatbot. Let's go!

Creating the Chat look

As the idea was to create a project for a 1-hour live coding, I decided to create a very simple interface using HTML, CSS and JavaScript for the Chatbot. I'm pretty bad with design, so I chose the font and colors I liked the most. So let's go, starting with HTML.

  
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Chatbot Assistente de Idiomas</title>
    <link rel="preconnect" href="https://fonts.googleapis.com">
    <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
    <link href="https://fonts.googleapis.com/css2?family=Roboto&display=swap" rel="stylesheet">
    <link rel="stylesheet" href="style.css">
  
  
    <div>



<p>E agora o CSS da página<br>
</p>

<pre class="brush:php;toolbar:false">* {
  box-sizing: border-box;
  margin: 0;
  padding: 0;
  font-family: "Roboto", sans-serif;
}

body {
  display: flex;
  justify-content: center;
  align-items: center;
  min-height: 100vh;
  background-color: #f2f2f2;
}

.chat-container {
  width: 100%;
  max-width: 400px;
  background-color: #fff;
  border-radius: 10px;
  box-shadow: 0px 4px 10px rgba(0, 0, 0, 0.1);
  overflow: hidden;
}

.chat-box {
  height: 300px;
  max-height: 500px;
  overflow-y: auto;
  padding: 16px;
  display: flex;
  flex-direction: column;
}

.chat-form {
  width: 100%;
  display: flex;
  justify-content: space-between;
}

.message {
  padding: 10px;
  margin-bottom: 8px;
  border-radius: 20px;
  width: auto;
  display: inline-flex;
  max-width: 50%;
  word-wrap: break-word;
}

.model {
  background-color: #e0e0e0;
  color: #333;
  align-self: flex-start;
  justify-content: flex-start;
}

.user {
  background-color: #4caf50;
  color: white;
  align-self: flex-end;
  justify-content: flex-end;
  margin-left: auto;
}

.input-container {
  display: flex;
  padding: 10px;
  border-top: 1px solid #ddd;
}

#user-input {
  flex: 1;
  padding: 10px;
  border: 1px solid #ddd;
  border-radius: 20px;
  outline: none;
}

#send-button {
  margin-left: 10px;
  padding: 10px 15px;
  background-color: #4caf50;
  color: white;
  border: none;
  border-radius: 20px;
  cursor: pointer;
}

#send-button:hover {
  background-color: #45a049;
}

The result of this should be a screen similar to the one shown below:

Creating client logic

Our application is a chatbot that will communicate with the Gemini API. So, we need to create the logic that will make this communication. To make it clear what we should do, I will list it below:

• Get what is typed by the user
• Make a POST request to the '/chat' route that we will create
• Display the user and model (the AI) message on the chat screen

So here we go, first let's add an event listener to execute our logic only after the DOM content is fully loaded:

// script.js
document.addEventListener("DOMContentLoaded", () => {
  const chatForm = document.getElementById("chat-form");
  const chatWindow = document.getElementById("chat-window");
  const userInput = document.getElementById("user-input");

  // ...
});

We create constants to capture the elements that interest us, such as the input where the user types, the window where messages will appear and the form field, as we will listen when it is submitted and then execute our logic.

Continuing, let's move on to the second step, which is to make the request for the route that we will create by sending the user's message.

{
  "name": "chatbot-ia",
  "version": "1.0.0",
  "main": "index.js",
  "scripts": {
    "test": "echo \"Error: no test specified\" && exit 1"
  },
  "keywords": [],
  "author": "",
  "license": "ISC",
  "description": ""
}

In this code, we are listening to the submit event on the form element. So, at first we use preventDefault to prevent the page from reload whenever we send a message. Then we take what the user typed, removing the blank spaces from the message, from the beginning and end with trim() and check if the message is not empty, blank. If the message is empty, we will stop our process right there.

Now, if we have the user's message, we display it on the screen using the addMessage() function. This function is defined as follows:

public
  |__ index.html
  |__ style.css
  |__ script.js
package.json 

Basically, it receives who sent the message and the text of the message and displays this information in the chat, adding the correct styles of the user and model, the AI ​​model.

Okay, now going back to the logic of our request, if we have a user message we need to make a POST request using the fetch API, and the body of this request is the user message.

Finally, if we receive a response to this request, we will display the model's message in the chat. Otherwise, we take the error and display it in the console, with console.error() or display a message in the chat itself in a customized way. And to improve chat usability, we cleaned the user message input with userInput.value = "";.

The script.js file looks like this:

  
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Chatbot Assistente de Idiomas</title>
    <link rel="preconnect" href="https://fonts.googleapis.com">
    <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
    <link href="https://fonts.googleapis.com/css2?family=Roboto&display=swap" rel="stylesheet">
    <link rel="stylesheet" href="style.css">
  
  
    <div>



<p>E agora o CSS da página<br>
</p>

<pre class="brush:php;toolbar:false">* {
  box-sizing: border-box;
  margin: 0;
  padding: 0;
  font-family: "Roboto", sans-serif;
}

body {
  display: flex;
  justify-content: center;
  align-items: center;
  min-height: 100vh;
  background-color: #f2f2f2;
}

.chat-container {
  width: 100%;
  max-width: 400px;
  background-color: #fff;
  border-radius: 10px;
  box-shadow: 0px 4px 10px rgba(0, 0, 0, 0.1);
  overflow: hidden;
}

.chat-box {
  height: 300px;
  max-height: 500px;
  overflow-y: auto;
  padding: 16px;
  display: flex;
  flex-direction: column;
}

.chat-form {
  width: 100%;
  display: flex;
  justify-content: space-between;
}

.message {
  padding: 10px;
  margin-bottom: 8px;
  border-radius: 20px;
  width: auto;
  display: inline-flex;
  max-width: 50%;
  word-wrap: break-word;
}

.model {
  background-color: #e0e0e0;
  color: #333;
  align-self: flex-start;
  justify-content: flex-start;
}

.user {
  background-color: #4caf50;
  color: white;
  align-self: flex-end;
  justify-content: flex-end;
  margin-left: auto;
}

.input-container {
  display: flex;
  padding: 10px;
  border-top: 1px solid #ddd;
}

#user-input {
  flex: 1;
  padding: 10px;
  border: 1px solid #ddd;
  border-radius: 20px;
  outline: none;
}

#send-button {
  margin-left: 10px;
  padding: 10px 15px;
  background-color: #4caf50;
  color: white;
  border: none;
  border-radius: 20px;
  cursor: pointer;
}

#send-button:hover {
  background-color: #45a049;
}

---

And with that we finish the front-end part of the chatbot. The next step will be to create our "server", communicate with the Gemini API and talk to it about life, the universe and everything else!

See you next time!