Extract pages from a PDF into a new document

The common case is a certificate buried inside a larger scan: page nine of a forty-page bundle, needed alone for one upload slot. Extraction turns an unusable file into exactly the document the form asked for.

Extraction is also the honest way to meet a page limit. Where a portal allows two pages and your proof runs to six, sending the two relevant pages is better than compressing six and hoping — and it shares less of your information than the alternative.

Because pages are copied rather than re-encoded, an extraction keeps the text selectable and the images untouched. What you get out is precisely what was on those pages of the original file.

This tool exists for PDF documents handled in a hurry: no queue, no account, no upload, and a result you can verify on the spot. The explanation below is the same one we would give a colleague who asked how it works.

PDFs can contain text, images, fonts, forms and document properties. A structural rewrite may reduce a file, but the saving depends on its contents. Keep the original and inspect the output: a rewrite is not guaranteed to remove all identifying metadata, preserve interactive features or make a file safe.

A habit worth keeping: before you upload anything produced here, open the result and check three things — the page count, that the first and last pages are the ones you expect, and that the smallest text on the densest page is still readable at 100%. That takes twenty seconds and catches the two failure modes that actually happen: a page quietly lost during a rebuild, and a target that was reached by making a scan unreadable.

Accepted inputs are PDF files up to 50 MB. SVG, HTML and other script-bearing formats are refused outright rather than sanitised, because a file that can execute code has no business being re-encoded — and an upload field that accepts one is a security hole rather than a feature. The tool works on one file at a time, which keeps the result unambiguous.

The source is parsed once, then a new document is assembled from copies of the selected pages, with each page's content streams and resource dictionaries cloned intact. Selection is order-preserving: 3,1,2 produces the pages in that sequence, which is occasionally useful when a bundle was scanned out of order. Out-of-range page numbers are reported instead of being silently clamped, and every output is written with a fresh cross-reference table so it opens cleanly in strict readers such as those used by banks. Extraction keeps the original page geometry, rotation and resource references, so an extracted page looks identical to the same page in the source document. Because the new file no longer contains the pages you dropped, it is also usually a great deal smaller — which is why extracting the two relevant pages is often a better answer to a byte limit than compressing the whole transcript.