Installation

npm install curse-filter

Loading the module

ESM / Typescript

This module is first thought for Typescript/ESM, so this is the recommended way to load it.

import { filter } from 'curse-filter';

Note

Profanity language is used in this README for demonstration purposes only. Please be respectful.

Usage

supportedLangs

import { supportedLangs } from 'curse-filter';

// This will log an array of supported languages
console.log(supportedLangs);

filter()

The filter() function asynchronously replaces curse words in a string with asterisks ("***") or a custom placeholder.

import { filter } from 'curse-filter';

// This must be used in an async function or top-level await context

await filter('fuck you');
// result: '*** you'

await filter('fuck you', { lang: 'en' });
// result: '*** you'

await filter('fuck you, coglione', { lang: ['en', 'it'] });
// result: '*** you, ***'

await filter('fuck you, coglione', { lang: ['en', 'it'], placeholder: 'customPlaceholder' });
// result: 'customPlaceholder you, customPlaceholder'

Parameters

• str (string) – The input string to filter.

• options (optional) – An object with:

  • lang (string | string[] | true): One or more language codes (e.g., 'en', 'it', or ['en', 'it']). If omitted, all supported languages will be used. Note that the fewer languages you pass to the function's options object, the faster the filtering will be.

  • placeholder (string): The replacement string. Defaults to '\*\*\*'.

  • customKeywords (Set<string>): A Set to add custom words to look for in the string.

Returns

A Promise<string> with the filtered string.

detect()

The detect() function asynchronously detects whether or not curse words are in a string.

import { detect } from 'curse-filter';

// This must be used in an async function or top-level await context

await detect('fuck you');
// result: true

await detect('fuckyou');
// false (no space, not detected by default)

await detect('fuckyou', { rigidMode: true });
// true (rigid mode allows substring matching)

await detect('fuck you, coglione', { lang: ['en', 'it'] });
// true

Parameters

• str (string) – The input string to scan.

• options (optional) – An object with:

  • lang (string | string[]): One or more language codes (e.g., 'en', 'it', or ['en', 'it']). If omitted, all supported languages will be used.

  • rigidMode (boolean): If true, performs substring detection (e.g., matches "fuckyou"). Defaults to false.

  • processedChunkSize (number): Optional internal chunk size for performance control (default is 100).

  • customKeywords (Set<string>): A Set to add custom words to look for in the string.

Returns

A Promise<boolean> – Resolves to true if profanity is detected, otherwise false.

Express.js middlewares

curse-filter comes with built-in express.js middlewares.

detectMiddleware middleware

The detectMiddleware middleware analizes the whole req.body object for curse words.

import express, { Request, Response } from 'express';
import { detectMiddleware } from 'curse-filter';
import { registerUserToDB } from './db';

const app = express();

app.use(express.json());

app.post('/register', detectMiddleware, async (req: Request, res: Response) => {
    /* If the request body contains curse words, the middleware will automatically 
    send a 422 response with a message containing the detected curse words.
    If no curse words are detected, the middleware will call the next() function. */

    await registerUserToDB(req.body);

    res.status(200).json({ message: 'User registered!' });
});

It is possible to configure the middleware with the following options:

// Class for configuring the middleware
import { MiddlewaresConfig } from 'curse-filter';

// Default values:

MiddlewareConfig.onError = null; // Called when a curse word is detected, before sending the response

MiddlewareConfig.detectOptions = {
    lang: 'en', // Language(s) to use for detection
    rigidMode: false, // If true, performs substring detection (e.g., matches "fuckyou")
    processedChunkSize: 100, // Optional internal chunk size for performance control
    customKeywords: new Set() // A Set to add custom words to look for in the string
}; // Options for the detect function

MiddlewareConfig.errorMessage = 'Not allowed content detected.'; // Message sent in the response

MiddlewareConfig.statusCode = 422; // Status code sent in the response

Typescript

SupportedLang type

import { SupportedLang } from 'curse-filter';

const lang: SupportedLang = 'en';

New Version v6.0.0 → v7.0.0

This version drops the CustomKeywords Set-like object to introduce a quicker way of adding custom words to look for and drops support for the synchronous versions of the functions, in favor of promise-based functions. This means that the filter() and detect() functions and the detectMiddleware middleware are now only asynchronous and return a promise for increased performance.

License

MIT