Convert Word and Excel to PDF in the browser (no server, no upload)

Office-to-PDF converters almost always upload your file to a server. For an invoice, a contract or a salary sheet, that's a lot of trust for a format change. But you can convert Word and Excel to PDF entirely in the browser — read the file with the right parser, rebuild it as clean HTML, and let the browser's own print engine produce a pixel-faithful PDF.

Here's the pattern for both, plus the one trick that makes the PDF output rock-solid instead of blank.

The idea: parse → HTML → native print

There's no need for a heavyweight PDF-drawing library. Browsers already have a great PDF renderer behind window.print() ("Save as PDF"). So the job is just: get the document into clean HTML, then print that HTML off-screen.

The reliable HTML→PDF helper

Render the HTML in a hidden <iframe> and call print() on it. This never produces the blank pages that html2canvas-based approaches sometimes do, and it keeps real text (selectable, searchable) instead of rasterizing:

function printHtmlAsPdf(html, { format = "a4", orientation = "portrait", margin = 12 } = {}) {
  const pageCss =
    `<style>@page{size:${format} ${orientation};margin:${margin}mm}
     html,body{margin:0;background:#fff;-webkit-print-color-adjust:exact;print-color-adjust:exact}</style>`;
  const doc = `<!DOCTYPE html><html><head><meta charset="utf-8">${pageCss}</head><body>${html}</body></html>`;

  const iframe = document.createElement("iframe");
  iframe.style.cssText = "position:fixed;left:-10000px;top:0;width:820px;height:1160px;border:0;";
  document.body.appendChild(iframe);

  iframe.onload = () => {
    setTimeout(() => {
      iframe.contentWindow.focus();
      iframe.contentWindow.print();
      setTimeout(() => iframe.remove(), 120000);
    }, 300);
  };
  const d = iframe.contentWindow.document;
  d.open(); d.write(doc); d.close();
}

The user picks "Save as PDF" in the print dialog. That's the only UX trade-off, and you get perfect pagination and crisp text for free.

Word (.docx) → PDF with mammoth.js

mammoth.js converts .docx into semantic HTML — headings, lists, tables, bold/italic, embedded images. It deliberately ignores fiddly Word styling, which is exactly what you want for a clean PDF.

<script src="https://cdnjs.cloudflare.com/ajax/libs/mammoth/1.6.0/mammoth.browser.min.js"></script>

const buf = await file.arrayBuffer();                // .docx read locally
const { value: html } = await mammoth.convertToHtml({ arrayBuffer: buf });

const styled = `
  <style>
    body{font-family:Georgia,serif;font-size:12pt;line-height:1.5;color:#111}
    h1,h2,h3{font-family:Arial,sans-serif;line-height:1.25}
    table{border-collapse:collapse;width:100%}
    td,th{border:1px solid #999;padding:6px 8px}
    img{max-width:100%;height:auto}
  </style>${html}`;

printHtmlAsPdf(styled, { format: "a4", margin: 16 });

Gotcha: .docx only

mammoth handles the modern .docx (Open XML) format, not the legacy binary .doc. Detect it and tell the user to re-save:

if (!/\.docx$/i.test(file.name)) {
  alert("Please use a .docx file (open old .doc in Word and Save As .docx).");
}

Excel (.xlsx/.csv) → PDF with SheetJS

SheetJS reads .xlsx, .xls and .csv, and can emit an HTML table per sheet:

<script src="https://cdnjs.cloudflare.com/ajax/libs/xlsx/0.18.5/xlsx.full.min.js"></script>

const bytes = new Uint8Array(await file.arrayBuffer());
const wb = XLSX.read(bytes, { type: "array" });

let sections = "";
for (const name of wb.SheetNames) {
  const fullHtml = XLSX.utils.sheet_to_html(wb.Sheets[name]);
  // sheet_to_html returns a whole document — pull out just the <table>
  const table = (/<table[\s\S]*<\/table>/i.exec(fullHtml) || [fullHtml])[0];
  sections += `<h2>${name}</h2>${table}`;
}

const styled = `
  <style>
    body{font-family:Arial,sans-serif;font-size:10pt}
    table{border-collapse:collapse;width:100%}
    td,th{border:1px solid #b3b3b3;padding:4px 7px;white-space:nowrap}
    h2 + table{page-break-inside:auto}
  </style>${sections}`;

printHtmlAsPdf(styled, { format: "a4", orientation: "landscape", margin: 12 });

Gotcha: sheet_to_html returns a full document

XLSX.utils.sheet_to_html() gives you a complete <html> page, not a fragment. If you concatenate several of those you get nested documents. Extract just the <table> (regex above) before stitching sheets together. Also: default to landscape — spreadsheets are wide and clip badly in portrait.

What carries over

Source	Preserved	Dropped
Word `.docx`	headings, lists, tables, images, basic styling	text boxes, footnotes, complex columns
Excel `.xlsx`	every sheet's values + table structure	charts, conditional formatting, cell colors

The tables and text are rebuilt from content, so the result is clean and readable rather than a pixel copy — and for sharing a finished document, that's the point.

Privacy: files are parsed and rendered locally; nothing is uploaded.
Cost: pure static hosting.
Text stays text: real, selectable PDF text — not a screenshot.

I built both of these as free tools — Word to PDF and Excel to PDF — running fully in the browser with no upload. The whole free PDF toolkit is here. Happy to talk through the mammoth/SheetJS edge cases in the comments.