How to edit a PDF

The two kinds of edit

A PDF is a description of how to draw pages. That makes some changes trivial and others painful, and every editing question comes down to which side of that line you are on.

• Page work is trivial. Reorder, delete, rotate, crop, stamp, split. The page is copied intact and nothing is redrawn.
• Content work is painful. Reaching inside a page to change what it says means moving text that is positioned character by character, with no paragraphs holding it together.

Knowing which side of that line your job falls on saves you an afternoon.

Page work, which is easy

Organize pages shows every page as a thumbnail, so you drag them into order or drop the ones you do not want. Rotate and crop turns sideways scans upright and trims a margin without touching page content.

Merge joins documents. Each of these reopens the output and compares it against the input page by page, checking the page count, the exact text of every page and every page size, before anything is offered for download.

Adding text to a page

The stamp tool writes real selectable text onto pages you choose. It runs in four modes: page numbers, a header and footer, a watermark, and Bates numbering for legal work. You pick one of nine positions, one of three fonts, a size, a colour and an opacity.

Page scope covers all pages, odd, even, the first, the last, or a range you type. That handles a surprising share of what people mean by editing a PDF: a case number in the corner, a DRAFT mark across the page, a footer naming the version.

Removing content permanently

Redaction is the opposite job and the one people get wrong most often. Drawing a black rectangle in a viewer leaves the text underneath, where anyone can select it. Real redaction has to destroy the content.

The tool here rebuilds every affected page from a fresh render, so the marked content is not copied into the output at all. What it keeps is an invisible text layer holding only the characters outside the marks, which is how the rest of the page stays searchable.

Text work, which is not easy

Changing a word in a paragraph means finding the exact glyphs, deleting them, inserting new ones in a font that is probably embedded as a partial subset, and then repositioning everything after them. Full editors do this and it still goes wrong.

pdfpix does not attempt it. The text editing guide explains what happens when you try, names the free software that comes closest, and covers the case where the text is not text at all.

When the page is a scan

If you cannot select a word with your cursor, there is no text on that page. It is a photograph of a document, and every character you can see is a pattern of pixels. Nothing edits that without redrawing the picture.

The route is recognition software that reads the image and writes a text layer, then editing whatever it produced. Expect mistakes on anything handwritten, faint or crooked, and check numbers by eye rather than trusting the output.

Pick the order of operations

Do the page work first, then the text work, then compress, then sign. Running them the other way round means redoing steps, because signing seals the file and compressing rewrites it. Form filling belongs before all of it.

If the finished file has to be emailed, the compress tool takes a size target and reads the result back before handing it over. The editing hub lists the rest of this cluster.