Merge PDF

How a PDF merge actually works

A PDF is a container of numbered objects: pages, fonts, images and streams of instructions, tied together by a page tree and a cross-reference table. Merging does not re-render anything. It reads the page tree of each input, copies the page objects and the resources they point at into a new document, and writes a new cross-reference table for the result.

That is why a merge is fast and lossless: the text stays text, so the merged file is still searchable, the vector content stays sharp at any zoom level, and page size and orientation are copied per page rather than forced into a single size. It is also why the merged file can be larger than the inputs added together - each input brings its own fonts and images along, and nothing is deduplicated between them.

The copy happens in your browser tab. A 20-page merge is usually a fraction of a second; the limiting factor is reading the files off disk, not the page copying.

What survives a merge, and what does not

Everything that lives inside a page survives: the text, the images, the vector drawings, links drawn as annotations on the page, and the page geometry. That covers the vast majority of what people merge.

Features that live at document level do not come along, because they sit in the catalogue of the original file rather than on its pages: bookmarks and outlines, AcroForm fields, signature fields, layer visibility settings, and the structure tags that screen readers follow. A merged file is therefore untagged, and a signed document that is merged stops being a verifiable signed document - the visible signature is copied as page content, but the cryptographic proof is not.

Password-protected PDFs cannot be merged here. The library refuses an encrypted document rather than guessing at it, and since the file never reaches a server there is nowhere to type a password. Open the file once in a reader that has the password, save an unprotected copy, and merge that.

Order, naming and the limits worth knowing

The order of pages in the output follows the order of the files you added, so the act of dropping them in is the act of arranging them. If you would rather not worry about the order you drop them in, switch the merge order to alphabetical by file name, which sorts natural numbers sensibly - appendix 2 comes before appendix 10 rather than after it.

The output is saved to your device as the name you type with .pdf added. Rename it before you start if the file is heading into a shared folder or an upload form with a filename rule.

This page accepts up to 20 files per job, 25 MB per file and 80 MB in total, because everything is held in memory in the tab. Documents beyond that size are better merged one pair at a time, which also keeps each download small enough to check as you go.

The workflow: build the parts first, merge last

A merge takes whole documents in the order they are listed; it cannot leave pages out or put them in a different order. The pieces therefore have to be right before they arrive - an excerpt comes out of the split tool, a mis-ordered file passes through the page-order box first, and only then do the prepared files join. Merging is the last step in the chain, not a repair shop on the way in.

Because document-level features do not travel with the pages, forms and signatures should be finished in each source while their fields are still live, before the join. And since every input keeps its own fonts and images rather than sharing them with the others, the result can be a little larger than the parts added together - so if size matters, compress after the merge, not before it.