Merge PDF files

How to use it

1. Choose two or more PDFs — the page count for each is shown so you can check you have the right file.
2. Reorder them with the arrows; the merge follows the order on screen.
3. Press Merge and the combined PDF downloads straight away.

Why a merged file can be larger than its parts

Merging copies each source document’s page objects into a new file along with everything those pages reference: font programs, colour profiles, embedded images. Nothing is re-encoded, so quality is untouched, but nothing is shared between the sources either. Twelve invoices that each embed their own subset of the same corporate typeface arrive as twelve separate font programs, and the merged file carries all of them.

The effect shows up most when combining many small, similar documents - generated invoices, certificates, exported reports - where the per-file overhead dwarfs the actual page content. Merging two large scans behaves as expected, because the bulk there is image data that was never duplicated. If the combined file needs to be small, the saving has to come from the sources: shrink oversized scans before combining, because the merge copies page streams verbatim rather than recompressing them.

What survives the merge, and what does not

Everything attached to a page travels with it - the content stream, images, link annotations, form widgets, selectable text in its original encoding. Structures that belong to the document rather than to any page do not. Bookmarks, document-level attachments, and the form dictionary that ties scattered field widgets into one fillable form all live above the page tree, and have no single sensible meaning once pages from several documents sit side by side.

In practice, expect a merged file whose pages render identically to the originals but whose bookmark sidebar is empty, and treat merged fillable forms with suspicion - check that fields still accept input before sending one on. Internal links frequently break too, because a link’s destination is a reference to a page object rather than a page number. Rebuild those in whatever produced the PDF.

Questions