Adding to a PDF, and the one mistake that leaks data
This tool adds content on top of a page: a line of text, a signature image, a solid box. It does not change the text already in the document — that is a genuinely different and much harder problem, covered further down. For the vast majority of what people need from "editing a PDF" — filling a form that was never made fillable, signing, dating, adding a reference number — drawing on top is exactly the right operation, and it leaves the original content untouched and still searchable.
A white box is not redaction
This is the part worth reading twice. When you cover a bank account number with a box, you are drawing a rectangle over it. The number is still in the file, underneath, exactly where it was. Anyone can select the text through the box and copy it, extract it with a script, or simply open the document in a tool that lists its text content. It has happened to law firms, governments and newspapers, with real consequences.
A cover box is fine for tidying a page, hiding a stray mark, or blanking something on a document that will only ever be printed. It is not a way to remove sensitive data from a file you are going to send. If the goal is that nobody can recover what was underneath, the reliable route is to flatten the page into an image: convert the document with PDF to image, cover what you need in an image editor, and rebuild the PDF with image to PDF. You lose the searchable text layer, which is precisely the point.
Where your text actually lands
A browser measures from the top-left corner downward. A PDF measures from the bottom-left corner upward, in points rather than pixels, and a page can additionally carry a rotation that viewers apply when displaying it. Placing an element where you dropped it therefore means translating between two coordinate systems that disagree about which way is up.
Rather than reimplement that arithmetic, this editor asks PDF.js — the engine that rendered the page you are looking at — to convert each point for it, so the position you see on screen is the position written into the document. Pages with rotation are the case most likely to drift, so if you are working on a scan that came in sideways, check the saved file before sending it.
Text, boxes and signatures
Text is drawn with Helvetica, one of the fonts every PDF reader is required to have, which keeps the file small and avoids embedding anything. The trade-off is its character set: accented Latin letters are fine, but symbols outside it — other alphabets, some typographic marks — cannot be written and are dropped with a warning.
Images accept PNG and JPG. For a signature, photograph or scan it on white paper and run it through background removal first: a transparent PNG sits on the signature line instead of covering it with a white rectangle.
What this cannot do
It cannot rewrite the text that is already there. A PDF does not store paragraphs — it stores instructions to draw specific glyphs at specific coordinates, with fonts embedded as subsets containing only the characters actually used. Changing one word means re-extracting that font, re-measuring the line and reflowing the paragraph around it. Desktop software does this imperfectly; in a browser, honestly, it is not on the table. Anything advertising it usually turns your page into a picture and draws over it, which quietly destroys the searchable text.
It also does not open password-protected files, and adding content invalidates any existing digital signature on the document — that is the signature working correctly, since it certifies one exact file. For page-level work such as reordering, rotating or splitting, use the page organizer.