I made a WhatsApp bot to solve Math problems

I am fascinated by chatbots. They can carry out any task automatically for you and can be invoked at a single command. I found bots on Discord and was intrigued to make one for WhatsApp. But, WhatsApp doesn’t share its official API with everyone. Fortunately, many workarounds exist using WhatsApp Web. I used whatsapp-web.js to ease authentication and messaging controls. It provides a simple WhatsApp Web client built using Node & Puppeteer.

Using this bot, you can add more functionality and automate anything you’d like. Maybe you want to monitor your servers or get random memes. I found WolframAlpha’s computational intelligence fascinating and hence decided to integrate its API with WhatsApp messaging.

Setup

Start a new NPM project in your favourite project directory

npm init

Install Whatsapp Web packages

We need whatsapp-web.js for WhatsApp client and qrcode-terminal to parse the auth information into visible QR code.

npm i whatsapp-web.js qrcode-terminal

Boilerplate code to get started

Create a new file index.js and add the following code.

const qrcode = require("qrcode-terminal");
const { Client } = require("whatsapp-web.js");

const puppeteerOptions = {
  headless: true,
  args: ["--no-sandbox"],
};

const client = new Client({
  puppeteer: puppeteerOptions,
});

client.on("qr", (qr) => {
  qrcode.generate(qr, { small: true });
});

client.on("ready", () => {
  console.log("Client is ready!");
});

client.on("message_create", (message) => {
  const messageBody = message.body;
  console.log(messageBody);
  if (messageBody == "!ping") {
    message.reply("pong");
  }
});

client.initialize();

This will initialize a client instance. We pass puppeteerOptions to client to disable a GUI. Notice the message_create event. The client listens for any new messages. Right now, we will listen to only !ping and respond with pong.

Add the following start script to package.json:

"scripts": {
  "start": "node index.js"
}

Start the node app with npm start. You should see a QR code displayed in the terminal. Scan this QR code with the WhatsApp Web option in the WhatsApp menu. Once scanned successfully, you should see Client is ready! in the console.

Note: WhatsApp only allows web use from a single session. If already web sessions are active (for e.g.: on a browser), they will automatically become inactive.

Responding to Math problems

To solve Math problems, we are going to use the Wolfram|Alpha API. Sign up for a free account and create a new app on the Developer Portal by clicking on Get an App ID button.

Create a new file called .env and paste the APPID:

APPID=<App ID from Wolfram|Alpha here>

Install dotenv:

npm i dotenv

Then initialize dotenv by adding this line at the top of index.js:

require("dotenv").config();

Download WolframAlphaAPI.js to your project directory. Initialize the instance:

const appid = process.env.APPID;
const WolframAlphaAPI = require('./WolframAlphaAPI.js');
let wraAPI = WolframAlphaAPI(appid);
const invokeKey = '!b';

Update the message handler:

client.on("message_create", (message) => {
  const messageBody = message.body;
  if (messageBody.startsWith(invokeKey)) {
    messageHandler(message);
  }
});

const messageHandler = (message) => {
  let query = message.body.substring(invokeKey.length + 1);
  console.log(`Querying result for ${query}`);
  handleImage(message, query, wraAPI);
};

Add image handling:

const parseDataUrl = require("parse-data-url");
const { MessageMedia } = require("whatsapp-web.js");

const handleImage = (message, text, wraAPI) => {
  try {
    wraAPI
      .getSimple({ i: text, timeout: 5 })
      .then((res) => {
        const parsed = parseDataUrl(res);
        message.reply(new MessageMedia(parsed.contentType, parsed.data));
      })
      .catch((err) => {
        message.reply(String(err));
      });
  } catch (err) {
    console.log(err);
  }
};

And … we’re done!

Start the node app again with npm start. Once you have authenticated, type !b integrate x^2 in any chat and the bot should reply with the correct image output.

The complete code is available on GitHub.

What’s next

Now you have a way to programmatically reply to WhatsApp messages. What you can create with it is limitless:

• Lots of prebuilt WolframAlpha APIs — text & audio responses, steps for a problem, conversational API
• Google search responses
• Random joke or roast
• Server monitoring alerts

Right now, you need to scan the QR code for every run but the session can be stored in a JSON file to resume it again. Read more on Resuming Sessions.

Thank you for reading this tutorial to the very end!