Crop a photograph and the discarded pixels are gone. Crop a PDF and nothing is discarded at all — which is fine when you are trimming margins for an e-reader, and a serious problem when you thought you were removing something.
Does cropping a PDF remove the content?
No — it hides it
A crop writes a rectangle into the page and viewers show only what falls inside it. The text and images outside that rectangle are untouched and still in the file, and resetting the rectangle brings every bit of it back. That is how nearly every PDF cropper works, ours included.
The boxes a page has
A PDF page carries several rectangles, and the two that matter here are easy to keep straight.
The MediaBox is the page itself — the full sheet, the thing that existed before anyone touched it.
The CropBox is what a viewer is asked to display. It sits inside the MediaBox, and a program that respects it draws only that region.
Cropping sets the CropBox. The content stream — the actual drawing instructions for every character and every line — is not edited, not truncated, and not re-encoded. The page still says what it said. It is simply being read through a smaller window.
Two further boxes exist for printing, the TrimBox and the BleedBox, describing where a finished page is cut and how far ink runs past that edge. Consumer tools leave both alone, which is why a professionally prepared print file can behave differently from one that has merely been cropped.
Why this is fine, most of the time
The commonest reason to crop is a scan with margins wider than the text. On a phone or an e-reader that white space is half the screen, and trimming it makes the words visibly larger without touching a single glyph. Nothing is being concealed, so the fact that the margins are still in the file is of no consequence.
The same is true of removing a scanner's black edge, straightening a lopsided capture, or cutting a header that repeats on every page. In each case the crop is cosmetic and the hidden content is content nobody wanted anyway.
Cropping here takes margins in millimetres and applies them to the pages you choose, in your browser, without the file going anywhere. What it writes is a CropBox, exactly as described above.
When it is not fine
If you are cropping to remove a name, an address, a signature, a price or anything else that should not travel with the document, cropping is the wrong instrument. The recipient opens the file in a tool that ignores the CropBox — or simply resets it — and reads what you thought you had cut off.
The same trap catches the other intuitive method. Drawing a black rectangle over sensitive text hides it visually and changes nothing underneath: the characters are still in the content stream, still selectable, still copyable, and still found by search. Newspapers have published documents redacted this way and had them read within the hour.
Making it actually gone
The reliable approach removes the possibility rather than the appearance. Render the page to an image, which turns every character into pixels and leaves nothing to recover, then build a new PDF from those images.
PDF to images, then images to PDF, does exactly that. Crop the images in between if the point was to cut something off.
It costs you the searchable text of those pages, and that is not a side effect — it is the guarantee. Text that cannot be searched is text that cannot be extracted either.
While you are there, the document properties are worth a look. A PDF remembers its author, the software that made it and often the path it was saved from, none of which is visible on the page; stripping the metadata takes a moment and there is more in there than most people expect.
The short version
Crop freely for margins, e-readers and tidiness — that is what a CropBox is for, and the hidden content harms nothing. Never crop to remove something that matters. When it genuinely has to go, flatten the page to an image and rebuild, and accept the loss of searchable text as the price of being sure.