Merging joins two or more PDFs end to end into one document, in the order you arrange them. Every page keeps its own size and orientation, so a portrait report followed by a landscape appendix stays exactly that rather than being squashed to a common shape. The work happens in your browser through pdf-lib: the files are read into memory, the pages are copied into a new document, and the result downloads straight from the tab — nothing is uploaded, there is no size cap and no queue. Text stays selectable and searchable because the pages are copied rather than re-rendered, and the same is true of links and annotations within each file. What does not survive is anything that pointed at the document as a whole: bookmarks, the outline tree and form field names that now collide with a second copy of themselves.
What survives the merge and what does not
Page content comes across untouched. Because the pages are copied rather than re-rendered, text remains selectable, vector graphics stay sharp at any zoom, and embedded fonts travel with the pages that use them. Internal links keep working as long as both ends of the link are inside the same source file.
Document-level structures are the casualties. Bookmarks and the outline tree describe a document that no longer exists, so they are dropped. Form fields are the subtler problem: field names have to be unique, and merging two copies of the same form produces two fields called the same thing. If the inputs are filled forms, flatten them first so the answers become page content.
Order is the thing people get wrong
Files are merged in the order shown, not the order you selected them, and browsers hand over multi-file selections in an order that is not always alphabetical. Dragging the thumbnails into place takes a moment and removes the most common reason a merged document has to be redone.
Where documents have their own numbering — a contract and its schedules, chapters of a report — the merged file will have several page 1s. Adding page numbers afterwards over the top of the merged document is the fix, and it is a separate step precisely because the numbering you want depends on what you merged.
File size after merging
The result is roughly the sum of the inputs, sometimes slightly less where the same font is embedded in more than one source and can be shared. It is never dramatically smaller: merging is not compression, and a 20 MB scan joined to a 5 MB report produces about 25 MB.
If the total lands over an email or upload limit, compress afterwards rather than expecting the merge to help. Scanned documents are where compression pays — re-encoding the page images typically takes a large scan down by most of its size, while a text-based PDF has very little to give up.