This code snippet is used to detect toxic messages in a Discord server using Google's Perspective API. It checks the toxicity of a message and deletes it if it exceeds a certain threshold. The message author is then notified about the toxic content. The code uses regular expressions to remove emojis and emotes from the message content before ana…

ModerateMessageWithAI.js

/*
Author: Le Ngo Duc Manh
ModifiedDate: 2024-10-14
Description: This code snippet is used to detect toxic messages in a Discord server using Google's Perspective API. It checks the toxicity of a message and deletes it if it exceeds a certain threshold. The message author is then notified about the toxic content. The code uses regular expressions to remove emojis and emotes from the message content before analyzing it for toxicity. The Google API key is required to make requests to the Perspective API. The maximum allowed toxicity threshold can be adjusted as needed.
*/

const axios = require('axios');

const GOOGLE_API_KEY = "Your Google API Key";
const EMOJIS_REGEX = /<a?:([^:]+):(\d+)>|\p{Emoji_Presentation}|\p{Extended_Pictographic}/gmu;

/* This is the maximum threshold */
const MAX_ALLOWED_TOXICITY_THRESHOLD = 85;

/** Check if a string is empty or only contains emojis
 * 
 * @param {string} str The string to check
 * @returns {boolean} Whether the string is empty
 */
function isEmptyMessage(str) {
  return !str?.length;
}

/** Format the content of a message to remove emojis and emotes
 * 
 * @param {string} str the string to format
 * @returns {string} The formatted string
 */
function formatContent(str) {
  return str.replace(EMOJIS_REGEX, '').trim();
}

module.exports = async (message) => {
  // We skip checking bot message for less calling to the API
  if (message.author.bot) return;

  // This check if a message content is empty or not, if it is then quick return
  const messageContent = formatContent(message.content);
  if (isEmptyMessage(messageContent)) return;

  try {
    const response = await axios.post(
      'https://commentanalyzer.googleapis.com/v1alpha1/comments:analyze',
      {
        comment: { text: messageContent },
        languages: ['en'],
        requestedAttributes: { TOXICITY: {} }
      },
      {
        params: { key: GOOGLE_API_KEY }
      }
    );

    // round the score received to an int from 0 to 100
    const toxicity = Math.round(response.data.attributeScores.TOXICITY.summaryScore.value * 100);

    // if larger, do the executions you want.
    if (toxicity > MAX_ALLOWED_TOXICITY_THRESHOLD) {
      await message.delete();
      // send a warning message to the channel in which message was sent from
      message.channel.send(`Hey ${message.author}, please do not use toxic words here!`);
    }
  } catch (err) {
    console.error(`Error analyzing message: ${err.message}`);
  }
};

it works, thank you

Thank you very much for the feedback, wouldn't you mind leaving a star? :)

I dont know much about API keys, but you probably dont want to leave that API key as valid.

I dont know much about API keys, but you probably dont want to leave that API key as valid.

I can't share that API key, so I decided to insert it with a quote from our nations. :>
You can register your own Google API keys by searching Google. It's will be very helpful. 😍
Thank you for reading this. :3

Ah cool. I don't need the API key, just didnt want you to end up with a big bill for someone using it crypto mining.

Have a good one.

Just gave it a fresh update today. 😙