pdf2square

Convert PDF pages to exactly square (NxN) images with letterboxing, plus extract text from each page. Available as both a CLI tool and a JavaScript library.

Features

• 📄 Convert PDF pages to square images (letterboxed to maintain aspect ratio)
• 🖼️ Support for PNG and JPEG output formats
• 📝 Extract text from each page
• 🎯 High DPI rendering for crisp text (configurable)
• ⚡ Concurrent processing for better performance
• 🎨 Customizable background colors (including transparency for PNG)
• 📚 Available as both CLI tool and programmatic library
• 🔤 Base64 encoded output for library usage

Requirements

No system dependencies required - uses PDF.js for PDF processing and Sharp for image manipulation.

Installation

npm install pdf2square

Or for global CLI usage:

npm install -g pdf2square

CLI Usage

Basic Usage

# Convert PDF to 896x896 PNG images + text files  
pdf2square input.pdf

# Specify output prefix
pdf2square input.pdf output/page

# Convert with custom options
pdf2square input.pdf --size 512 --dpi 300 --format jpg

CLI Options

| Option | Description | Default | | ----------------------- | ----------------------------------------------- | --------- | | -n, --max-pages <int> | Maximum pages to convert | 10 | | -s, --size <int> | Target square size in pixels | 896 | | --dpi <int> | Render DPI (higher = crisper text) | 700 | | --first <int> | First page to convert (1-based) | 1 | | --format <fmt> | Output format: png or jpg | png | | --bg <color> | Background color (#RRGGBB[AA] or 'transparent') | #ffffffff | | --concurrency <int> | Max parallel processes | 4 | | --keep-intermediate | Keep intermediate renders | false |

CLI Examples

# Convert first 5 pages to 512x512 JPEG with white background
pdf2square document.pdf --max-pages 5 --size 512 --format jpg --bg "#ffffff"

# Convert pages 3-7 with transparent background (PNG only)
pdf2square document.pdf --first 3 --max-pages 5 --bg transparent

# High DPI conversion for crisp text
pdf2square document.pdf --dpi 1000 --size 1024

# Process with higher concurrency
pdf2square document.pdf --concurrency 8

Library Usage

Basic Example

import { convert } from 'pdf2square';

// Convert a PDF file to square images with extracted text
const results = await convert('./path/to/your/document.pdf');

results.forEach((page) => {
  console.log(`Page ${page.pageNumber}:`);
  console.log(`- Image: ${page.base64EncodedImage.substring(0, 50)}...`);
  console.log(`- Text: ${page.extractedText.substring(0, 100)}...`);
});

API Reference

convert(pathToPdf, options?)

Converts PDF pages to base64 encoded square images with extracted text.

Parameters:

• pathToPdf (string): Path to the input PDF file
• options (object, optional): Conversion options

Returns: Promise<ConvertedPDFPage[]>

ConvertedPDFPage

interface ConvertedPDFPage {
  pageNumber: number; // Page number (1-based)
  originalPath: string; // Path to the original PDF file
  base64EncodedImage: string; // Base64 encoded image with data URL prefix
  extractedText: string; // Extracted text from the page
}

ConvertOptions

interface ConvertOptions {
  maxPages?: number; // Maximum pages to convert (default: 10)
  size?: number; // Target square size in pixels (default: 896)
  dpi?: number; // Render DPI (default: 700)
  first?: number; // First page to convert (default: 1)
  format?: 'png' | 'jpg'; // Output format (default: 'png')
  bg?: string; // Background color (default: '#ffffffff')
  concurrency?: number; // Max parallel processes (default: 4)
}

Library Examples

Custom Options

import { convert } from 'pdf2square';

const results = await convert('./path/to/document.pdf', {
  maxPages: 5,
  size: 512,
  dpi: 300,
  format: 'jpg',
  bg: '#ffffff',
  concurrency: 2,
});

Save Images to Files

import { convert } from 'pdf2square';
import fs from 'node:fs/promises';

const results = await convert('./path/to/document.pdf');

for (const page of results) {
  // Extract base64 data (remove data URL prefix)
  const base64Data = page.base64EncodedImage.replace(
    /^data:image\/\w+;base64,/,
    '',
  );
  const buffer = Buffer.from(base64Data, 'base64');

  // Save image and text files
  await fs.writeFile(`page-${page.pageNumber}.png`, buffer);
  await fs.writeFile(`page-${page.pageNumber}.txt`, page.extractedText);
}

Web Service Integration

import { convert } from 'pdf2square';

// Express.js route handler example
export async function processPDF(req, res) {
  try {
    // req.file.path comes from multer file upload middleware
    const results = await convert(req.file.path, {
      maxPages: 10,
      size: 896,
      format: 'png',
    });

    res.json({
      success: true,
      totalPages: results.length,
      pages: results.map((page) => ({
        pageNumber: page.pageNumber,
        image: page.base64EncodedImage,
        text: page.extractedText,
        textWordCount: page.extractedText.split(/\s+/).length,
      })),
    });
  } catch (error) {
    res.status(500).json({
      success: false,
      error: error.message,
    });
  }
}

Batch Processing

import { convert } from 'pdf2square';

async function processBatch(pdfPaths) {
  const results = [];

  for (const pdfPath of pdfPaths) {
    console.log(`Processing ${pdfPath}...`);

    try {
      const pages = await convert(pdfPath, {
        maxPages: 5,
        concurrency: 2, // Lower concurrency when processing multiple files
      });

      results.push({
        pdfPath,
        success: true,
        pageCount: pages.length,
        data: pages,
      });
    } catch (error) {
      console.error(`Failed to process ${pdfPath}:`, error.message);
      results.push({
        pdfPath,
        success: false,
        error: error.message,
      });
    }
  }

  return results;
}

// Example usage:
// const results = await processBatch([
//   './documents/file1.pdf',
//   './documents/file2.pdf',
//   './documents/file3.pdf'
// ]);

Output Format

CLI Output Files

When using the CLI, files are saved with the following naming convention:

input.pdf → input-001.png, input-001.txt
          → input-002.png, input-002.txt
          → ...

Library Output

The library returns base64 encoded images with data URL prefixes:

{
  pageNumber: 1,
  originalPath: "/path/to/input.pdf",
  base64EncodedImage: "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAA...",
  extractedText: "This is the text content from page 1..."
}

Background Colors

Supported background color formats:

• Hex colors: #RRGGBB or #RRGGBBAA
• Transparent: transparent (PNG only)
• Examples: #ffffff, #ff0000aa, transparent

Note: JPEG format cannot be transparent and will fallback to white background.

Performance Tips

1. Concurrency: Adjust --concurrency based on your CPU cores and available memory
2. DPI: Higher DPI produces better text quality but increases processing time and memory usage
3. Format: JPEG is faster and produces smaller files than PNG but doesn't support transparency
4. Page Range: Use --first and --max-pages to process only the pages you need
5. Memory: Large PDFs with high DPI settings may require significant memory

Error Handling

Common errors and solutions:

• "Could not determine page count": PDF may be corrupted, password-protected, or invalid
• "No pages to convert": Check --first and --max-pages parameters
• "Invalid PDF structure": The file is not a valid PDF document
• "Format must be 'png' or 'jpg'": Use a supported output format
• "Invalid background color": Use hex format (#RRGGBB or #RRGGBBAA) or 'transparent'
• Memory errors: Reduce DPI, page count, or concurrency settings

TypeScript Support

The library includes TypeScript declarations for better development experience:

import { convert, ConvertedPDFPage, ConvertOptions } from 'pdf2square';

const options: ConvertOptions = {
  maxPages: 5,
  size: 512,
  format: 'png',
  dpi: 300,
  bg: '#ffffff',
};

const results: ConvertedPDFPage[] = await convert(
  './path/to/document.pdf',
  options,
);

// Process results with full type safety
results.forEach((page: ConvertedPDFPage) => {
  console.log(
    `Page ${page.pageNumber}: ${page.extractedText.length} characters`,
  );
});

License

MIT License - see LICENSE file for details.