prompt-bouncer/dist/types.d.ts

A lightweight, customizable content moderation library for AI applications. Filters profanity, explicit content, and inappropriate prompts for text-to-image generation.

/**
 * Configuration interface for the content filter
 */
export interface FilterConfig {
    /** Enable profanity detection */
    enableProfanityFilter?: boolean;
    /** Enable explicit content detection */
    enableExplicitFilter?: boolean;
    /** Enable violence/harmful content detection */
    enableViolenceFilter?: boolean;
    /** Enable self-harm content detection */
    enableSelfHarmFilter?: boolean;
    /** Enable drug-related content detection */
    enableDrugsFilter?: boolean;
    /** Enable hate speech detection */
    enableHateSpeechFilter?: boolean;
    /** Enable mild content detection (gaming, technical terms, mild expressions) */
    enableMildFilter?: boolean;
    /** Custom words to add to the filter */
    customBannedWords?: string[];
    /** Words to allow (whitelist) */
    allowedWords?: string[];
    /** Case sensitive matching */
    caseSensitive?: boolean;
    /** Enable word boundary detection (prevents partial matches) */
    strictWordBoundaries?: boolean;
}
/**
 * Result of content moderation
 */
export interface ModerationResult {
    /** Whether the content is safe */
    isSafe: boolean;
    /** Reason for flagging if not safe */
    reason?: string;
    /** Specific words that were flagged */
    flaggedWords: string[];
    /** Categories of violations found */
    categories: string[];
    /** Confidence score (0-1) */
    confidence: number;
    /** Original text */
    originalText: string;
    /** Cleaned version of the text */
    cleanedText: string;
    /** Severity level */
    severity: "low" | "medium" | "high";
}
/**
 * Detection category
 */
export interface DetectionCategory {
    name: string;
    keywords: string[];
    severity: "low" | "medium" | "high";
    description: string;
}
//# sourceMappingURL=types.d.ts.map