@proseprep/core

Pure TypeScript text processing engines powering ProsePrep.io. Zero external dependencies. Works in any browser or Node.js environment.

Install

npm install @proseprep/core
# or
pnpm add @proseprep/core

Modules

The library is organized into specialized modules, each focused on a specific text processing domain. All modules are exported from the root @proseprep/core package.

🧹 AI Sanitizer (sanitize)

Cleans and normalizes AI-generated text — removes hidden characters, normalizes smart quotes/dashes/ellipsis, strips markdown/HTML, and collapses whitespace.

import { sanitizeModule, purgeLLMTells, sanitizeMarkdown } from '@proseprep/core';

// Using the Module interface
const result = sanitizeModule.run(text, { removeAsterisks: true });

// Using specialized utilities
purgeLLMTells("Sure! Here's the answer..."); 
sanitizeMarkdown("**bold*"); // Fixes broken syntax

📝 LinkedIn Formatter (linkedin)

Converts text to Unicode mathematical styles (bold, italic, script, monospace, underline, strikethrough) and creates numbered circle lists for social media.

import { applyLinkedInStyle, applyLinkedInList, linkedinModule } from '@proseprep/core';

applyLinkedInStyle('Hello World', 'bold');   // 𝗛𝗲𝗹𝗹𝗼 𝗪𝗼𝗿𝗹𝗱
applyLinkedInList('Item one\nItem two');     // ① Item one\n② Item two

🔍 Robotic AI Detector (detector)

Scores text for AI-tell patterns including robotic buzzwords, passive voice ratio, and sentence length uniformity (pacing).

import { scanRoboticHeuristics } from '@proseprep/core';

const scan = scanRoboticHeuristics(text);
// scan.score (0-100), scan.grade ('Human'|'Robotic'), scan.highlightedHtml

✨ AI Humanizer (humanizer)

Transforms robotic AI-generated text into natural, human-like copy by replacing buzzwords with context-aware synonyms and rewriting passive voice.

import { humanizeText } from '@proseprep/core';

humanizeText(text, 'casual', 2); // tones: 'casual' | 'professional' | 'creative'

👁️ Invisible Character Scanner (invisible)

Detects and reports zero-width spaces (ZWSP), byte-order marks (BOM), and non-breaking spaces (NBSP).

import { scanInvisibleChars, getVisualDebuggerText } from '@proseprep/core';

const scan = scanInvisibleChars(text);
const debug = getVisualDebuggerText(text); // Visual markers for hidden characters

␠ Whitespace & Stripping (whitespace, stripper)

Utilities for collapsing excess whitespace, removing YAML/JSON frontmatter, and stripping HTML tags.

import { compressWhitespace, stripFrontmatter, stripHtml } from '@proseprep/core';

compressWhitespace(text, { collapseLines: true, trimTrailing: true });
stripFrontmatter(text); // Removes --- YAML --- blocks
stripHtml(text);        // Removes <tags>

⇄ Case & CSV Transformer (devcase)

Converts strings between common coding casings and transforms CSV data into Markdown tables.

import { caseTransform, csvToMarkdownTable } from '@proseprep/core';

caseTransform('hello world', 'camel');    // helloWorld
csvToMarkdownTable('Name,Age\nAlice,30'); // Markdown grid

📊 Quality Assurance (qa)

Provides live readability assessments, Flesch grade indexes, reading/speaking time estimations, and keyword density analysis.

import { calculateReadability, scanKeywordDensity } from '@proseprep/core';

const stats = calculateReadability(text);
const keywords = scanKeywordDensity(text);

✓ Grammar Scanner (grammar)

Detects and auto-fixes spelling mistakes, double words, spacing errors, a/an mismatches, and sentence capitalization.

import { scanGrammar, fixGrammar } from '@proseprep/core';

const issues = scanGrammar(text);
const fixed = fixGrammar(text);

✍️ Paraphraser (paraphraser)

Simplifies, expands, or rewrites sentences to change tone while preserving core meaning entirely on the client side.

import { paraphraseText } from '@proseprep/core';

paraphraseText(text, 'simplify'); // 'standard' | 'simplify' | 'expand'

👯 Duplicate Remover (duplicate)

Deduplicates list elements or text blocks with alphabetical sorting, whitespace trimming, and frequency analysis.

import { removeDuplicateLines } from '@proseprep/core';

const result = removeDuplicateLines(text, { sortOrder: 'asc' });

PrepModule Interface

First-class modules (sanitizeModule, linkedinModule) implement the standard PrepModule interface for uniform integration:

interface PrepModule<Options = any> {
  id: string;
  name: string;
  description: string;
  defaultOptions: Options;
  run(input: string, options?: Partial<Options>): {
    output: string;
    applied: boolean;
    changes: string[];
  };
}

License

MIT