Stack PDF
A PDF merger that runs entirely in the browser. Files are never uploaded —
there is no backend to upload them to. The site builds to a static export, so
serving it costs nothing per merge and scales with CDN traffic rather than CPU.
app/ routes (/, /privacy, /terms, robots.txt, sitemap.xml)
components/ merger.tsx (tool state), sheet-row.tsx (one file), ad-slot.tsx
lib/merge.ts PDF inspection + merging, pdf-lib loaded on first use
lib/site.ts site name, canonical URL, ad publisher id
cli/pdfMerger.py the original local script; unrelated to the site
Built with Next.js (App Router, static export), Tailwind, pdf-lib for the PDF
work and dnd-kit for reordering.
Develop
npm install
npm run dev # http://localhost:3000
npm run build # static export into out/
npm run build writes out/, which any static host will serve.
NEXT_PUBLIC_SITE_URL sets the canonical URL, Open Graph tags, robots.txt
and sitemap.xml. Copy .env.example to .env.local to override it locally.
How the merge works
Each file is checked by its leading bytes rather than its extension, then
opened to count pages. Merging copies every page of every document into a new
one in list order, and the result is handed back as a blob the browser
downloads. All of it happens on the main thread in the user's tab.
Worth knowing:
- Page content, text, images and vector graphics are copied exactly. Form
fields and bookmarks are not carried over; the output is flat. Any signature
on a source document is invalidated by merging, as it is in every tool,
because the signed bytes change.
- Encrypted PDFs are refused with a message rather than silently skipped.
pdf-lib is ~400 kB and loads on first use rather than first paint, which
keeps it out of the initial bundle.- Ad slots reserve their space and load no script unless an ad publisher id is
configured, so a blocked or absent ad never shifts the layout.
- There is no file size limit beyond available memory. If very large files ever
feel janky, moving the merge to a Web Worker is the next step.
