@habibrosyad/docx-extractor

Extract DOCX documents to structured data and rebuild them programmatically. This library provides full round-trip conversion with preservation of formatting, styles, tables, images, and numbering.

Installation

npm install @habibrosyad/docx-extractor

Quick Start

import { DocxExtractor, DocxBuilder } from '@habibrosyad/docx-extractor';
import fs from 'fs';

// Extract DOCX to structured data
const extractor = new DocxExtractor();
const buffer = fs.readFileSync('document.docx');
const document = await extractor.extract(buffer);

// Modify the document structure
document.paragraphs.forEach(para => {
  if (para.runs) {
    para.runs.forEach(run => {
      // Modify text, formatting, etc.
    });
  }
});

// Rebuild DOCX
const builder = new DocxBuilder();
const newBuffer = await builder.build(document);
fs.writeFileSync('output.docx', newBuffer);

Features

• ✅ Full Structure Extraction: Paragraphs, tables, images, styles, numbering
• ✅ Formatting Preservation: Fonts, colors, spacing, alignment, borders
• ✅ Round-Trip Conversion: Extract → Modify → Rebuild without data loss
• ✅ Style Support: All paragraph and character styles with inheritance
• ✅ Table Support: Complete table structure with cell properties
• ✅ Image Support: Extract and embed images in documents
• ✅ Numbering Support: Bullet lists and numbered lists

API

DocxExtractor

Extract a DOCX file to structured data.

const extractor = new DocxExtractor();
const document = await extractor.extract(buffer);

Returns: ExtractedDocument with:

• paragraphs: Array of paragraphs
• tables: Array of tables
• body: Ordered sequence of paragraphs and tables
• styles: Map of style definitions
• defaults: Document-wide run formatting defaults
• paragraphDefaults: Document-wide paragraph defaults (spacing, etc.)
• numbering: Map of numbering definitions
• mediaFiles: Map of embedded images/media

DocxBuilder

Build a DOCX file from structured data.

const builder = new DocxBuilder();
const buffer = await builder.build(document);

Parameters: ExtractedDocument (from extractor or construct it yourself)

Returns: Promise<Uint8Array> - DOCX file buffer (compatible with Node.js, browsers, and edge runtimes)

Platform Compatibility

This library works across all JavaScript runtimes:

• ✅ Node.js: Full support (v18+)
• ✅ Cloudflare Workers: Full support
• ✅ Browsers: Full support
• ✅ Deno: Full support
• ✅ Bun: Full support

Using in Cloudflare Workers

export default {
  async fetch(request) {
    // Get DOCX file from request
    const arrayBuffer = await request.arrayBuffer();
    
    // Extract and process
    const extractor = new DocxExtractor();
    const document = await extractor.extract(arrayBuffer);
    
    // Modify document...
    
    // Rebuild
    const builder = new DocxBuilder();
    const outputBuffer = await builder.build(document);
    
    // Return as response
    return new Response(outputBuffer, {
      headers: {
        'Content-Type': 'application/vnd.openxmlformats-officedocument.wordprocessingml.document',
        'Content-Disposition': 'attachment; filename="output.docx"'
      }
    });
  }
};

Document Structure

Paragraph

• runs: Array of text runs with formatting
• spacing: Paragraph spacing (before, after, line)
• alignment: Left, center, right, justify
• indentation: Left, right, firstLine, hanging
• styleName: Applied paragraph style

Table

• rows: Array of table rows
• rows[i].cells: Array of table cells
• cells[i].runs: Text content with formatting
• cells[i].width, backgroundColor, borders: Cell properties

Run Formatting

• bold, italic, underline, strike
• fontSize, fontFamily, color
• highlight color

Requirements

• Modern JavaScript runtime with ES modules support (Node.js >= 18, Deno, Bun, Cloudflare Workers, or modern browsers)

License

MIT