Merge and split PDF

What it does

Two things everybody does constantly, and that almost always end up on a site that asks you to upload the document: merging several PDFs into one, and extracting the pages you need from a big one.

When merging, the order is the order of the list: drag a row to move it, or use the arrows. When extracting, write the pages — 1-3, 7, 9-10 — or tick them one by one.

The files never leave

That is why this tool exists. Online converters upload the document to a server, process it there and send it back; that document is often a contract, an invoice or a medical record. Here the PDF is read by the browser, changed by the browser and downloaded from the browser: no network request goes out, and closing the page leaves nothing behind.

The order you write is the order you get

7, 1-3 extracts page seven, then one, two and three — in that order. Reordering it silently would remove the only way to move a page to the front by writing it.

For the same reason 5-2 comes out backwards: five, four, three, two. And a page written twice comes out twice, which in a PDF is a legitimate thing — a cover sheet to repeat, a page to duplicate.

The pages are not thumbnails, and that is a declared choice

Each page is a rectangle at the true proportion of the page, with its number and its size in millimetres. It is not a preview of the content, and it will not become one: the library that manipulates the PDF works on the structure of the document and does not draw it. A real thumbnail would need a second engine, which would weigh more than everything else on this site put together — for a picture two centimetres wide.

What you actually need when scanning a list of pages is to find the landscape one among the portraits, and the proportion is enough for that.

Rotation adds up

Rotating a page ninety degrees when it was scanned sideways — and already declares ninety in the file — gives one hundred and eighty. Writing the value instead of adding it would leave the page where it was, and whoever pressed the button would see nothing happen.

Rotation follows the page, not its position: if the same page appears twice in the extract, it comes out rotated both times.

What it does not do

Password-protected PDFs do not open. The library does have a way to carry on over encrypted content: it produces a file that saves and does not read, which is the worst defect in this category because it looks like it worked. Here it refuses and says why. A protected PDF has to be unlocked first, with the program that made it.

Large files are a memory problem. Everything happens in the browser, so the limit is the tab’s: a document has to be read into memory, copied page by page and written back out. Above fifty megabytes per file a warning appears — not a refusal, because on a recent laptop that file goes through, and saying no would be a lie about someone else’s machine.

It does not compress, convert, sign or recognise text. It merges, extracts and rotates.

About your data

Nothing leaves this page and nothing is saved.