Edit PDF

The reason PDF has a reputation for being uneditable is worth understanding, because it tells you what this editor can and cannot do. Inside the file there is no paragraph. There is a content stream that says: switch to font F1 at 11 points, move to these coordinates, draw these glyph codes. Line breaks were decided when the file was written and baked in. That is why no browser tool retypes an existing sentence properly; there is nothing to retype, only positioned glyphs whose spacing was calculated once. So the editing here is additive, and it works in two layers. The page you see is rendered by pdf.js onto a canvas, where you place text, drop in a PNG or JPEG, or draw freehand. When you save, pdf-lib writes each of those onto the real page: text through drawText with an embedded font, strokes converted into SVG path operators, images embedded as image objects. The coordinate system flips on the way in, because PDF measures from the bottom left corner in points, 72 to the inch, while a canvas measures from the top left. What matters about that is where the additions end up. They go into the page content, not into an annotation or a comment. No reader can toggle them off, they survive flattening and printing, and they look the same in Acrobat, in Chrome, in Preview and on paper. The trade-off is that they are not draggable afterwards either; the editing session is where you move things. For what the editor cannot do there are other doors. To make existing words disappear for good, with the underlying text objects removed rather than covered, use Redact PDF. To get an editable document instead of a document with additions, convert to Word. To put words on a scan that has none, run OCR PDF first. And the whole process happens in your browser, so a payslip or a contract can be marked up without uploading it anywhere.