@torqlab/check-forbidden-content

A utility library for detecting forbidden content patterns in text. Used to validate prompts and descriptions before AI image generation to ensure compliance with content policies.

Installation

npm install @torqlab/check-forbidden-content

Or with Bun:

bun add @torqlab/check-forbidden-content

Usage

import checkForbiddenContent from '@torqlab/check-forbidden-content';

const hasForbidenContent = checkForbiddenContent('A photo of a person running');
console.log(hasForbidenContent); // true

const isSafe = checkForbiddenContent('An athlete running through a scenic landscape');
console.log(isSafe); // false

Forbidden Content

This library detects the following categories of forbidden content:

Real Persons or Identifiable Individuals

• Keywords: person, people, individual, human, man, woman, child, kid, baby, face, portrait, photo, picture, image

Political or Ideological Symbols

• Keywords: political, politics, government, president, election, vote, democracy, republican, democrat, flag, banner, symbol, emblem, crest

Violence or Combat

• Keywords: violence, violent, fight, war, battle, weapon, gun, knife, sword, attack, kill, death, blood, combat, military, soldier, army, navy

Sexual Content

• Keywords: sexual, sex, nude, naked, explicit, adult, porn

Typography and Text Instructions

• Keywords: text, word, letter, alphabet, caption, typography, font

API

checkForbiddenContent(text: string): boolean

Checks if the provided text contains any forbidden content patterns.

Parameters:

• text (string): The text to check for forbidden content.

Returns:

• (boolean): true if forbidden content is detected, false otherwise.

License

MIT