How to split a PDF: extracting pages, ranges and single sheets

The word covers several operations, and the tool you pick should match the one you actually want.

• Extract a range — keep pages 9 to 12 as a new document. This is the most common need by a wide margin, usually because one certificate lives inside a larger scan.
• Extract scattered pages — keep pages 1, 4 and 7 together. Useful for pulling a form, a photograph page and a receipt out of a bundle in one pass.
• Remove pages — the inverse: keep everything except pages 3 and 4. Typical case is dropping a cover sheet or a bank footer page.
• Split every page — one file per page. Mostly used for ticket-style documents and for feeding a scanner-style workflow where each page becomes its own attachment.

A range is written as two numbers with a dash: 3-7 keeps pages three, four, five, six and seven — both ends included. A comma separates separate selections: 1,4,9-11 keeps pages one, four, nine, ten and eleven. Ranges can be combined with single pages in any order.

The trap is that the number you see in your PDF reader is not always the number the file uses internally. Documents with front matter often carry roman numerals for a preface, so the "page 9" printed at the bottom of the sheet may be the twelfth physical page of the file. Always split by the physical position, which is what every splitter counts, then check the output. If in doubt, extract a single page first as a test.

If the upload limit is the problem, the instinct is to compress. Sometimes splitting is the better answer, because compression always costs something and splitting costs nothing.

• The portal wants one file per document. Splitting is not optional here; compression is irrelevant.
• The limit is on pages, not megabytes. Many government portals accept a fixed number of pages per upload. Compressing a 30-page scan never makes it 5 pages.
• Only part of the document is relevant. Sending twelve pages when two are needed is a privacy problem before it is a size problem. Extract the two.
• A single page is unreadable at the size needed. The opposite case: if one page is a dense scan that must stay legible, split it out and compress it at a higher quality target rather than degrading the whole bundle to meet a global limit.

Scanned bundles arrive with all the useful structure stripped out. A typical bank statement bundle is one file per month, but the scan might be three months stapled together by the clerk; an education bundle may contain degree, marksheets and migration certificate in a single 18-page scan.

The efficient approach is to open the file once, note the physical page number where each document starts and ends on a piece of paper — or in a note on your phone — and then run a single split operation for each range. Doing it in one pass beats ten passes of hunting.

Watch for pages that were scanned upside down or out of order. Rotating a page is a separate, also lossless, operation, and it is worth doing before you split, because fixing six files afterwards is more work than fixing the bundle once.

A PDF encrypted with a user password cannot be opened, let alone split, without that password — this is a property of the format, not a limitation of any particular tool. Documents protected only by an owner password, which forbids editing but not reading, are usually re-saveable in any mainstream reader; once saved without restrictions they split normally.

Digitally signed documents are a different matter. A signature covers the exact bytes of the file it was applied to. Splitting produces a new file, so the original signature no longer applies to it, and any reader will report the signature as invalid on the split output. If an authority requires a signed declaration, submit the signed file itself and split only the unsigned pages around it.

Split files are new files; the original is untouched on your disk. That is deliberate — but it also means the usual failure is not data loss, it is uploading the wrong extraction to a portal that will not let you replace it.

Two habits prevent it. First, name outputs after their content, not their page numbers: "degree-certificate.pdf" survives a week of confusion, "pages-9-12.pdf" does not. Second, open the split file, scroll to the last page and confirm where it ends. The off-by-one error is invisible until a reviewer notices that your marksheet ends mid-table.

Some portals limit pages rather than bytes. When that happens, splitting is the only honest fix, because compression never turns a ten-page document into a five-page one.

There is one legitimate workaround and it is worth understanding what it costs. Printing two or four pages of a document onto a single sheet and then producing a PDF from that layout halves the page count — at the price of halving the size of everything on the page. For a form set in large type that is acceptable; for a dense statement or a statutory document it renders the content unreadable at normal zoom, and the human reviewing the upload will notice.

Our splitter deliberately does not do page imposition, because the operation looks tidy in a page counter while quietly changing whether the document can be read. Extracting only the pages that matter is almost always the better answer to a page limit.