Merge PDF files into one
Drop in as many PDFs as you like, put them in the order you want, and take whole files or just a few pages from each. The pages are copied across untouched — nothing is re-compressed and nothing is uploaded.
Leave a file's Pages box empty to take the whole thing.
Your pages are copied, not redrawn
There are two ways to build one PDF out of several, and they are not equally good. The cheap way is to render every page to an image and staple the images together: the file balloons, the text stops being text, and you can no longer search it or copy a line out of it. The other way is to lift each page's object graph out of its original file and write it into a new one, leaving the page description untouched.
This does the second. A page's content stream — the sequence of drawing operators that says where each glyph and each line goes — is copied byte for byte, with whatever compression it already had. Embedded fonts, ICC colour profiles, JPEG and JBIG2 images and vector artwork all travel across unchanged. Nothing is decoded and re-encoded, so a scan does not get softer and a 300 dpi drawing does not lose a hair of resolution. Selectable text stays selectable.
That is also why merging is fast and why the result is roughly the sum of the inputs rather than several times larger. The only bytes that change are the structural ones: a new catalogue, a new page tree, a new cross-reference table.
Order, and taking only part of a file
Files merge in the order they appear in the list, and the arrows move them. Within each file, the Pages box takes a range: leave it empty for the whole document, or write something like 1-3, 2,5,9, 4- for “page four to the end”, or -2 for “up to page two”. Ranges may be listed in whatever order you want them to come out in, so 3,1,2 is a valid instruction and does what it says. The same file can be added twice with different ranges if you need to interleave.
Page numbers here are positions in the document, counting from one. They are not whatever number is printed on the page: a report whose first two pages are an unnumbered cover and contents page will have its “page 1” at position 3.
Mixed page sizes stay mixed
If you merge an A4 contract with a Letter invoice and a landscape spreadsheet, the result has A4 pages, Letter pages and landscape pages, each at its original size. Nothing is scaled to a common sheet. This is how the PDF format is meant to work — page size is a property of the page, not the document — but a number of tools normalise everything to A4 and quietly shrink or crop the odd ones out.
The same goes for rotation. A page saved with a 90-degree rotation flag keeps that flag, so it opens the same way round it did before. The list above tells you each file's page size and flags when they are not all the same, so you can see what you are about to get.
Password-protected files
Roughly one PDF in eight carries encryption, and most of the time it is not there to keep you out. Banks, payroll systems and tax portals routinely set an owner password, which marks the file as no-printing or no-copying while leaving the user password empty — that is why the file opens with a double-click and no prompt. Those open here normally, and the merged copy comes out without the restrictions.
A file that genuinely asks for a password when you open it is a different thing, and this tool will not try to get into it. You will see a message saying it needs a password. Open it in your usual reader, type the password, save an unlocked copy, and merge that. There is no password cracking here and no attempt at one.
One consequence worth saying plainly: the merged file is not protected. Whatever encryption or permission flags the inputs had, the output is an ordinary unencrypted PDF. If the document needs to stay restricted, re-apply that in a reader afterwards.
What does not come across
A PDF holds more than pages, and some of it has no sensible meaning once documents are combined. Being specific about what is dropped is more useful than a vague promise that everything is preserved:
Bookmarks and the outline tree are not carried over. They are document-level and point at page objects that no longer exist in their old form; merging several outlines into one correct tree is a different job from merging pages, and a half-done version would be worse than none.
Digital signatures stop being valid. This is not a limitation that can be engineered away: a signature certifies a specific sequence of bytes, and the merged file is by definition a different sequence. Any tool that claims to merge signed PDFs and keep the signatures intact is either wrong or is embedding the old file whole.
Structure tags (the invisible reading-order tree that screen readers use), document-level JavaScript, file attachments and named destinations are also left behind. Links between pages survive when both pages come along and are dropped when the target does not; external web links always survive.
Form fields do come across, with their field tree and their appearance, so a fillable form stays fillable. The one thing to watch is merging two different forms that happen to use the same field names: in PDF, fields with identical names share a value, so typing in one would fill the other. Rename the fields first if that matters.
Damaged files
PDFs get truncated by interrupted downloads and mangled by bad exports more often than you would think. When the cross-reference table at the end of the file is missing or points at the wrong places, this falls back to scanning the whole file for object headers and rebuilding the index from what is actually there. It is not magic — pages whose content was in the part that never arrived stay blank — but it frequently recovers a document that a reader declares unreadable. The file list says so when a file had to be rebuilt this way.
Common questions
Is there a limit on how many files or pages I can merge?
No. There is no queue, no daily cap and no file-size ceiling, because there is no server involved — the work happens in this page, using your own machine's memory. The practical limit is how much memory your browser will give a tab, which on a desktop is usually enough for several hundred megabytes of input. Very large merges take a few seconds of visible effort rather than failing.
Does merging reduce the quality of scans or photos?
No. Image data is copied across in whatever compression it already had — a JPEG stays that exact JPEG, byte for byte. Nothing is decoded and re-encoded, so there is no second round of lossy compression. This is the main practical difference between merging pages and the “print to PDF” route, which rasterises everything and typically both degrades the image and loses the text layer.
Why is the merged file not much smaller than the sum of the parts?
Because nothing is re-compressed, and because resources cannot always be shared. If two input files both embed the same font, the merged file ends up with two copies of it: they are separate objects with separate subsets and no safe way to prove they are interchangeable. Expect the output to be about the sum of the inputs, sometimes a little less where an input carried structural bloat that is not copied.
Can I merge a file that asks for a password?
Not directly, and deliberately so. Files with an owner password only — the common case, where the file opens without a prompt but is marked no-printing or no-copying — are read normally. A file that prompts you for a password when you open it is refused with a message. Unlock it in a reader, save a copy, and merge that copy.
Do bookmarks and links survive?
Links do, with a caveat; bookmarks do not. External links to web addresses always work in the merged file. Internal links between pages keep working when both the link and its target page are included in the merge, and are dropped when the target was left out, so you never get a link that jumps somewhere arbitrary. The bookmark tree is document-level and is not rebuilt.
Can I take just a few pages from each file?
Yes — that is what the Pages box next to each file is for. Leave it empty to take everything, or write a range such as 1-3, a list such as 2,5,9, or an open-ended range such as 7-. The order you write them in is the order they come out in, and you can add the same file twice with different ranges.
What happens to a PDF form?
Form fields come across and stay fillable: the widget annotations, their field tree and the form's default resources are all rebuilt in the output. The exception worth knowing is that two fields with the same name share one value in PDF, so merging two copies of the same blank form gives you fields that fill in together. Merging a filled form with an unrelated document is fine.
Is anything uploaded, logged or kept?
No. The file never leaves the tab: it is read with the browser's own file API, parsed in JavaScript on this page, and the result is handed straight to your downloads folder. There is no server to receive it. You can load this page, disconnect from the network, and merge documents with the connection off.