How to delete pages from a PDF — precisely and safely

Every document collection has one file that is mostly filler: the 60-page report of which twelve pages matter, the scan with a blank back for every sheet, the contract draft carrying three superseded riders. Deleting pages sounds destructive, which is exactly why the right tool is built backwards — you point at what goes, the original never moves, and a counter tells you what survives. This guide covers that method, then the decisions around it: delete or extract, what happens to file size, and where trimming sits in a longer workflow.

Delete pages in your browser, step by step

The tool for this is Delete PDF pages on this site. It runs entirely in your browser on a background worker — financial statements and medical letters stay on your device.

1. Open the page remover and drop the document anywhere, click to browse, or paste it from the clipboard. It opens locally and every page appears as a thumbnail.
2. Check the pages you want gone — they turn red as they are marked, and a counter shows exactly how many pages will be removed and how many survive.
3. Tap any page for a full-size render. Blank scan backs, duplicated invoices, the wrong attachment: verify them at full size, not postage-stamp size.
4. Click Delete pages. A new PDF is built on your device containing every unmarked page, in order. Your original file is untouched — deletion here means “give me a copy without these”.

Deleting or extracting — pick the smaller side

Deletion and extraction are mirror images: extract keeps the checked pages and discards the rest; delete discards the checked pages and keeps the rest. Same grid, same counter, opposite result — so the choice is pure arithmetic. Count the pages you care about, then work with the smaller group.

Keeping five pages out of sixty? Extract them — five clicks. Removing five pages from sixty? Delete them — five clicks too. The rule matters more as documents grow: marking forty correct pages invites the one mis-click that counts, while marking the twenty wrong ones does not.

Marking in reverse — and other grid tricks

The grid rewards a bit of strategy. Deleting many scattered pages goes faster in reverse: hit “Select all”, then uncheck the few pages you want to keep — the tool was built for exactly this inversion, and the counter confirms the flip as you click.

Blank backs deserve their own tactic. If the file came from a sheet-fed scanner, the blank sides cluster at the end of each sheet; preview one to be sure, then mark them in a sweep. And before you go hunting: deleting the last page of many scanned documents also removes the trailing blank most scanners add.

One guardrail is built in: the run button stays disabled until at least one page survives. A zero-page PDF is not a document anyone wants to download.

A last convenience while you are here: the kept pages follow the on-screen order, so drag them into a better sequence and trim in the same pass.

What actually happens to file size

Expect the file to shrink roughly in proportion to the content removed — image-heavy pages buy back the most space, text-only pages the least. Two expectations to temper: deleting two blank pages will not rescue a 40 MB scan, because blank pages carry almost nothing to begin with. And PDF structure overhead makes the arithmetic approximate, not exact.

For maximum reduction, follow a big deletion with Compress PDF: fewer pages mean fewer duplicated resources, and the compressor has less to carry.

Trim first, then chain

Deletion is cheap, so spend it early. A smaller file is faster in every tool that follows — merging, rotating, converting — and easier on a phone’s memory. A sensible order for a full cleanup: unlock if the file is protected (unlock first — the browser cannot copy pages it cannot read), delete what does not belong, rotate what came out sideways, then compress last before sending.

When a desktop app makes more sense

Scripted repetition is the honest case for installed software. If the same weekly report needs the same eleven pages dropped, a command-line tool in a scheduled job beats opening a browser every time. Very large documents on old hardware can also favour a desktop application with direct memory access.

For the occasional trim — which is the overwhelmingly common case — the browser tool wins on the axis that matters most here: the document never leaves the device it lives on.