GLIN PROFANITY

A multilingual profanity detection and filtering engine for modern applications — by GLINCKER

A multilingual profanity detection and filtering engine for modern applications — by GLINCKER

✨ Overview

Glin-Profanity is a high-performance JavaScript/TypeScript library built to detect, filter, and sanitize profane or harmful language in user-generated content. With support for over 20+ languages, configurable severity levels, obfuscation detection, and framework-agnostic design, it's perfect for developers who care about building safe, inclusive platforms.

Whether you're moderating chat messages, community forums, or content input forms, Glin-Profanity empowers you to:

• 🧼 Filter text with real-time or batch processing
• 🗣️ Detect offensive terms in 20+ human languages
• 💬 Catch obfuscated profanity like sh1t, f*ck, a$$hole
• 🎚️ Adjust severity thresholds (Exact, Fuzzy, Merged)
• 🔁 Replace bad words with symbols or emojis
• 🧩 Works in any JavaScript environment -
• 🛡️ Add custom word lists or ignore specific terms

🚀 Key Features

💡 Why glin-profanity?

✨ What's New in v3.0

• Leetspeak Detection — Catch f4ck, @ss, $h!t with 3 intensity levels
• Unicode Normalization — Detect Cyrillic/Greek lookalikes, full-width chars, zero-width spaces
• Result Caching — 800x speedup for repeated checks
• ML Integration — Optional TensorFlow.js toxicity model for nuanced detection
• Performance — Optimized for high-throughput production workloads

📚 Table of Contents

• 🚀 Key Features
• 📦 Installation
• 🌍 Supported Languages
• ⚙️ Usage
  • Basic Usage
  • Framework Examples
• 🧠 API
  • Core Functions
  • Filter Class
    • Constructor
    • FilterConfig Options
    • Methods
      • isProfane
      • checkProfanity
  • useProfanityChecker Hook
    • Parameters
    • Return Value
• ⚠️ Note
• 🛠 Use Cases
• 🔬 Advanced Features
  • Leetspeak Detection
  • Unicode Normalization
  • Result Caching
  • Configuration Management
  • ML-Based Detection
• 📊 Benchmarks
• 📄 License
  • MIT License

Installation

To install Glin-Profanity, use npm:

npm install glin-profanity

OR

yarn add glin-profanity

OR

pnpm add glin-profanity

Supported Languages

Glin-Profanity includes comprehensive profanity dictionaries for 23 languages:

🇸🇦 Arabic • 🇨🇳 Chinese • 🇨🇿 Czech • 🇩🇰 Danish • 🇬🇧 English • 🌍 Esperanto • 🇫🇮 Finnish • 🇫🇷 French • 🇩🇪 German • 🇮🇳 Hindi • 🇭🇺 Hungarian • 🇮🇹 Italian • 🇯🇵 Japanese • 🇰🇷 Korean • 🇳🇴 Norwegian • 🇮🇷 Persian • 🇵🇱 Polish • 🇵🇹 Portuguese • 🇷🇺 Russian • 🇪🇸 Spanish • 🇸🇪 Swedish • 🇹🇭 Thai • 🇹🇷 Turkish

Note: The JavaScript and Python packages maintain cross-language parity, ensuring consistent profanity detection across both ecosystems.

Usage

Basic Usage

Glin-Profanity now provides framework-agnostic core functions alongside React-specific hooks:

🟢 Node.js / Vanilla JavaScript

const { checkProfanity } = require('glin-profanity');

const text = "This is some bad text with damn words";
const result = checkProfanity(text, {
  languages: ['english', 'spanish'],
  replaceWith: '***'
});

console.log(result.containsProfanity); // true
console.log(result.profaneWords);      // ['damn']
console.log(result.processedText);     // "This is some bad text with *** words"

🔷 TypeScript

import { checkProfanity, ProfanityCheckerConfig } from 'glin-profanity';

const config: ProfanityCheckerConfig = {
  languages: ['english', 'spanish'],
  severityLevels: true,
  autoReplace: true,
  replaceWith: '🤬'
};

const result = checkProfanity("inappropriate text", config);

Framework Examples

⚛️ React

import React, { useState } from 'react';
import { useProfanityChecker, SeverityLevel } from 'glin-profanity';

const App = () => {
  const [text, setText] = useState('');
  
  const { result, checkText } = useProfanityChecker({
    languages: ['english', 'spanish'],
    severityLevels: true,
    autoReplace: true,
    replaceWith: '***',
    minSeverity: SeverityLevel.EXACT
  });

  return (
    <div>
      <input value={text} onChange={(e) => setText(e.target.value)} />
      <button onClick={() => checkText(text)}>Scan</button>
      
      {result && result.containsProfanity && (
        <p>Cleaned: {result.processedText}</p>
      )}
    </div>
  );
};

💚 Vue 3

<template>
  <div>
    <input v-model="text" @input="checkContent" />
    <p v-if="hasProfanity">{{ cleanedText }}</p>
  </div>
</template>

<script setup>
import { ref } from 'vue';
import { checkProfanity } from 'glin-profanity';

const text = ref('');
const hasProfanity = ref(false);
const cleanedText = ref('');

const checkContent = () => {
  const result = checkProfanity(text.value, {
    languages: ['english'],
    autoReplace: true,
    replaceWith: '***'
  });
  
  hasProfanity.value = result.containsProfanity;
  cleanedText.value = result.autoReplaced;
};
</script>

🔴 Angular

import { Component } from '@angular/core';
import { checkProfanity, ProfanityCheckResult } from 'glin-profanity';

@Component({
  selector: 'app-comment',
  template: `
    <textarea [(ngModel)]="comment" (ngModelChange)="validateComment()"></textarea>
    <div *ngIf="profanityResult?.containsProfanity" class="error">
      Please remove inappropriate language
    </div>
  `
})
export class CommentComponent {
  comment = '';
  profanityResult: ProfanityCheckResult | null = null;

  validateComment() {
    this.profanityResult = checkProfanity(this.comment, {
      languages: ['english', 'spanish'],
      severityLevels: true
    });
  }
}

🚂 Express.js Middleware

const express = require('express');
const { checkProfanity } = require('glin-profanity');

const profanityMiddleware = (req, res, next) => {
  const result = checkProfanity(req.body.message || '', {
    languages: ['english'],
    autoReplace: true,
    replaceWith: '[censored]'
  });
  
  if (result.containsProfanity) {
    req.body.message = result.autoReplaced;
  }
  
  next();
};

app.post('/comment', profanityMiddleware, (req, res) => {
  // Message is now sanitized
  res.json({ message: req.body.message });
});

API

🎯 Core Functions

checkProfanity

Framework-agnostic function for profanity detection.

checkProfanity(text: string, config?: ProfanityCheckerConfig): ProfanityCheckResult

checkProfanityAsync

Async version of checkProfanity.

checkProfanityAsync(text: string, config?: ProfanityCheckerConfig): Promise<ProfanityCheckResult>

isWordProfane

Quick check if a single word is profane.

isWordProfane(word: string, config?: ProfanityCheckerConfig): boolean

🔧 Filter Class

Constructor

new Filter(config?: FilterConfig);

FilterConfig Options:

Methods

isProfane

Checks if a given text contains profanities.

isProfane(value: string): boolean;

• value: The text to check.
• Returns: boolean - true if the text contains profanities, false otherwise.

checkProfanity

Returns details about profanities found in the text.

checkProfanity(text: string): CheckProfanityResult;

• text: The text to check.
• Returns: CheckProfanityResult
  • containsProfanity: boolean - true if the text contains profanities, false otherwise.
  • profaneWords: string[] - An array of profane words found in the text.
  • processedText: string - The text with profane words replaced (if replaceWith is specified).
  • severityMap: { [word: string]: number } - A map of profane words to their severity levels (if severityLevels is specified).

⚛️ useProfanityChecker Hook

A custom React hook for using the profanity checker.

Parameters

• config: An optional configuration object (same as ProfanityCheckerConfig).

Return Value

• result: The result of the profanity check.
• checkText: A function to check a given text for profanities.
• checkTextAsync: A function to check a given text for profanities asynchronously.
• reset: A function to reset the result state.
• isDirty: Boolean indicating if profanity was found.
• isWordProfane: Function to check if a single word is profane.

const { result, checkText, checkTextAsync, reset, isDirty, isWordProfane } = useProfanityChecker(config);

Note

⚠️ Glin-Profanity is a best-effort tool. Language evolves, and no filter is perfect. Always supplement with human moderation for high-risk platforms.

🛠 Use Cases

• 🔐 Chat moderation in messaging apps
• 🧼 Comment sanitization for blogs or forums
• 🕹️ Game lobbies & multiplayer chats
• 🤖 AI content filters before processing input

🔬 Advanced Features

Leetspeak Detection

Detect and normalize leetspeak variations like f4ck, @ss, $h!t:

import { Filter } from 'glin-profanity';

const filter = new Filter({
  languages: ['english'],
  detectLeetspeak: true,
  leetspeakLevel: 'moderate', // 'basic' | 'moderate' | 'aggressive'
});

filter.isProfane('f4ck');   // true
filter.isProfane('@ss');    // true
filter.isProfane('$h!t');   // true
filter.isProfane('f u c k'); // true (spaced characters)

Leetspeak Levels:

• basic: Numbers only (0→o, 1→i, 3→e, 4→a, 5→s)
• moderate: Basic + common symbols (@→a, $→s, !→i)
• aggressive: All known substitutions including rare ones

Unicode Normalization

Detect homoglyphs and Unicode obfuscation:

import { Filter } from 'glin-profanity';

const filter = new Filter({
  languages: ['english'],
  normalizeUnicode: true, // enabled by default
});

// Detects various Unicode tricks:
filter.isProfane('fυck');   // true (Greek upsilon υ → u)
filter.isProfane('fᴜck');   // true (Small caps ᴜ → u)
filter.isProfane('f​u​c​k'); // true (Zero-width spaces removed)
filter.isProfane('ｆｕｃｋ'); // true (Full-width characters)

Result Caching

Enable caching for high-performance repeated checks:

import { Filter } from 'glin-profanity';

const filter = new Filter({
  languages: ['english'],
  cacheResults: true,
  maxCacheSize: 1000, // LRU eviction when full
});

// First call computes result
filter.checkProfanity('hello world'); // ~0.04ms

// Subsequent calls return cached result
filter.checkProfanity('hello world'); // ~0.00005ms (800x faster!)

// Cache management
console.log(filter.getCacheSize()); // 1
filter.clearCache();

Configuration Management

Export and import filter configurations for sharing between environments:

import { Filter } from 'glin-profanity';

const filter = new Filter({
  languages: ['english', 'spanish'],
  detectLeetspeak: true,
  leetspeakLevel: 'aggressive',
  cacheResults: true,
});

// Export configuration
const config = filter.getConfig();
// Save to file: fs.writeFileSync('filter.config.json', JSON.stringify(config));

// Later, restore configuration
// const savedConfig = JSON.parse(fs.readFileSync('filter.config.json'));
// const restoredFilter = new Filter(savedConfig);

// Get dictionary size for monitoring
console.log(filter.getWordCount()); // 406

ML-Based Detection

Optional TensorFlow.js-powered toxicity detection for context-aware filtering:

# Install optional dependencies
npm install @tensorflow/tfjs @tensorflow-models/toxicity

import { HybridFilter } from 'glin-profanity/ml';

const filter = new HybridFilter({
  languages: ['english'],
  detectLeetspeak: true,
  enableML: true,
  mlThreshold: 0.85,
  combinationMode: 'or', // 'or' | 'and' | 'ml-override' | 'rules-first'
});

// Initialize ML model (async)
await filter.initialize();

// Hybrid check (rules + ML)
const result = await filter.checkProfanityAsync('you are terrible');
console.log(result.isToxic);          // true
console.log(result.mlResult?.matchedCategories); // ['insult', 'toxicity']
console.log(result.confidence);       // 0.92

// Sync rule-based check (fast, no ML)
filter.isProfane('badword'); // true

ML Categories Detected:

• toxicity - General toxic content
• insult - Insults and personal attacks
• threat - Threatening language
• obscene - Obscene/vulgar content
• identity_attack - Identity-based hate
• sexual_explicit - Sexually explicit content
• severe_toxicity - Highly toxic content

📊 Benchmarks

Performance benchmarks on a MacBook Pro (M1):

Key Findings:

• Leetspeak and Unicode normalization add minimal overhead
• Caching provides 800x speedup for repeated checks
• Multi-language support scales well

Run benchmarks yourself:

npm run benchmark

License

This software is also available under the GLINCKER LLC proprietary license. The proprietary license allows for use, modification, and distribution of the software with certain restrictions and conditions as set forth by GLINCKER LLC.

You are free to use this software for reference and educational purposes. However, any commercial use, distribution, or modification outside the terms of the MIT License requires explicit permission from GLINCKER LLC.

By using the software in any form, you agree to adhere to the terms of both the MIT License and the GLINCKER LLC proprietary license, where applicable. If there is any conflict between the terms of the MIT License and the GLINCKER LLC proprietary license, the terms of the GLINCKER LLC proprietary license shall prevail.

MIT License

GLIN PROFANITY is MIT licensed.