There is a sharp divide between PDF operations that work reliably and ones that are always a bit of a gamble. Merging, splitting and reordering are on the reliable side, and it is worth knowing why.
Why these three always work
As the PDF format guide explains, a PDF stores each page as a self-contained object: its content stream, its resources, its dimensions. The document is essentially a list of those objects plus a table saying what order they go in.
So merging is concatenating two lists. Splitting is taking a slice. Reordering is permuting. None of these operations touch page content at all. They rearrange whole objects whose internals are never opened.
That is why they are lossless and why they never mangle a layout. Text editing, by contrast, requires reconstructing a document model that the file may not contain, which is why PDF text editors are inconsistent.
If you can express what you need as a page-level operation, you will get a predictable result.
Merging
Merge combines several PDFs into one, in the order you arrange them.
The common cases:
- Assembling an application: cover letter, CV, transcripts, references and identity documents, since most submission portals want one file.
- Combining scanned batches: a document scanner that produces one PDF per batch, reassembled into the whole.
- Adding a signed page back: print page 4, sign it, scan it, then split the original, drop the old page 4, and merge the signed one in.
- Compiling a report from separately authored sections.
Two things to know. Page sizes are preserved per page, so merging a Letter document with an A4 one gives you a PDF whose pages are different sizes, which prints inconsistently. If that matters, normalise the sizes before merging. And bookmarks and form fields from the source documents may not survive a simple merge, so check if your document relies on them.
Splitting
Split extracts a page range into a new PDF.
- Meeting a size limit: a portal that accepts 10 MB and a 30 MB scan.
- Sending one section: a single chapter, or one form from a packet.
- Removing pages you do not want to share, such as a cover sheet, an internal note, or a page with information the recipient does not need.
- Separating a batch scan into individual documents.
On the third point, a real caution: splitting out a page removes it from the new file entirely, which is genuinely different from visually covering something. Content hidden by a black rectangle drawn over it is still present in the file and trivially recoverable. If you are removing sensitive information, remove the page, or use proper redaction software that deletes the underlying content. A drawn box is no redaction.
Organizing
Organize reorders, rotates and deletes pages within one PDF.
- Fixing scan order: a document fed into a scanner in the wrong sequence.
- Rotating sideways pages: a landscape page scanned in portrait, or a batch where every other page is upside down because of a duplex feeder.
- Removing blank pages, which duplex scanning generates in quantity.
- Moving an appendix to where it belongs.
Rotation here is a page-object property, so it is lossless. The page is not re-rendered, just marked as displaying at a different angle.
Assembling a document from scratch
The general pattern, when you have a pile of things that need to become one PDF:
- Convert everything to PDF first. Images go through JPG to PDF and documents through Word to PDF. Normalise the page size at this stage by picking Letter or A4 and using it throughout.
- Name files so they sort correctly, with zero-padded numbers:
01-cover.pdf,02-cv.pdf. Without padding,10sorts before2. - Merge in order.
- Organize to fix anything out of place or rotated.
- Split if the result exceeds a size limit.
- Open the finished file and page through it before sending. This step catches nearly every mistake and takes a minute.
What page operations cannot do
The boundary, stated plainly:
- Editing text. Change the source document and re-export.
- Combining two pages onto one sheet (n-up). That is a printing setting and no file operation, so set it in your print dialogue.
- Compressing a large scan. Reducing file size means re-encoding the images, which is a different operation. Splitting is the workaround when there is a hard limit.
- OCR. Making a scan searchable requires text recognition, and page manipulation cannot supply it.
- Unlocking a password-protected PDF. These must be unlocked with the password before any tool can restructure them.
Doing it without uploading anything
This matters more for PDFs than for most files because the documents people merge and split are contracts, medical records, tax paperwork and identity documents. On most PDF sites, splitting a page out of a sensitive scan means first uploading the entire sensitive scan to a stranger’s server.
Every tool on this site runs in your browser using pdf-lib loaded into the page. The file is read from your disk, restructured in memory on your machine, and handed back as a download. There is no server-side component to receive it, and you can watch the network tab or work offline entirely.