How to Extract Pages from a PDF: In-Browser Page Selection & Xref Reconstruction

A standard Portable Document Format binary is not an image stream or sequential text markup. Under ISO 32000-1, a PDF file is an arbitrary acyclic directed graph of indirect dictionaries, content byte-streams, and reference tables linked by absolute physical byte positions. Decomposing pages cleanly requires navigating through three foundational layers:

Below is the industrial client-side extraction pattern implemented in high-reliability WebAssembly environments using JavaScript ArrayBuffers.

const fileInput = document.getElementById('pdf-dropzone');
const rawBuffer = await fileInput.files[0].arrayBuffer();

import { PDFDocument } from 'pdf-lib';

const sourceDoc = await PDFDocument.load(rawBuffer, {
  ignoreEncryption: false,
  updateMetadata: false
});
const extractedDoc = await PDFDocument.create();

const targetZeroIndices = [0, 2, 3, 4]; // Represents pages 1, 3-5
const copiedPageRefs = await extractedDoc.copyPages(sourceDoc, targetZeroIndices);

for (const page of copiedPageRefs) {
  extractedDoc.addPage(page);
}

const outputBytes = await extractedDoc.save({ useObjectStreams: true });
const blob = new Blob([outputBytes], { type: 'application/pdf' });

const downloadLink = document.createElement('a');
downloadLink.href = URL.createObjectURL(blob);
downloadLink.download = 'extracted-pages.pdf';
downloadLink.click();
URL.revokeObjectURL(downloadLink.href);

Enter sample page queries into the sandbox parser below to see how string queries translate into zero-indexed ArrayBuffer ranges with boundary safety checks:

Traditional online PDF converters represent one of the most widespread document leak vectors in modern operations. When legal covenants, salary sheets, or clinical case files are processed via remote APIs, document payloads are exposed to intermediate transit sniffing, non-compliant server storage, and training scrapers.