Scanned pages arrive with a black border where the lid did not close. Academic papers arrive with margins wide enough to write an essay in, which is exactly what you cannot do on a tablet screen. Slides exported to PDF arrive with a strip of nothing down one side. All three are the same job: move the page boundary inwards. Here is what that does inside the file, how to do it with Crop PDF, and the one thing about cropping that catches people out.
What cropping really changes
A PDF page is a set of drawing instructions plus two rectangles saying where the paper is. The MediaBox is the sheet, measured in points of 1/72 inch - A4 is 595 x 842 points. The CropBox says how much of it a reader displays. Most files leave the two equal; scanned and press-ready documents already crop in.
Cropping moves those rectangles and touches no drawing instruction. That is why a cropped PDF keeps perfectly sharp text - nothing is re-drawn or turned into a picture - why it is instant on a 300-page document, and why the file size barely moves.
Both rectangles get written, not just the CropBox. On paper the CropBox alone is what the specification calls a crop, but plenty of software - print pipelines especially - lays the page out from the MediaBox and would ignore a CropBox-only change, so the trim would appear to have silently failed on the one machine you needed it on.
Crop a PDF in four steps
- Open Crop PDF and drop your file on. It is read in your browser tab, not uploaded.
- Drag the box on the preview over the part you want to keep - inside it to move, a corner or edge to resize.
- Check the other pages with Previous and Next, or click Trim white margins to have it measured for you.
- Choose whether to crop every page or only some, then click Crop and download.
Four boxes under the preview show the trim on each edge in millimetres, so you can type an exact figure instead of dragging - useful when a printer has asked for 10 mm off every side.
Letting the tool find the margins
Trim white margins renders the pages, finds where the ink stops on each, and sets the box to the smallest crop that keeps every page whole - the fastest route for a scanned book or a paper to read on a phone.
Two details make it work on real scans, not just clean digital files. A single speck of scanner dust would otherwise hold the whole margin open, so a mark only counts once it is connected to something: joined-up strokes and hairlines are kept, isolated specks are not. And blank pages are skipped rather than treated as full-bleed, which would drag the crop back out to the whole sheet.
On a long document it measures an evenly spaced sample and tells you how many pages it checked - a cap you cannot see reads as "I checked everything" when it did not. Drag the result tighter if you want.
A worked example: an A4 scan with a black edge
Take an A4 page, 595.28 x 841.89 points, whose text block starts 60 points in from the left and right, 70 below the top and 80 above the bottom. Keeping 5 points of breathing room, the trim on each edge is:
- Left and right: 60 - 5 = 55 points, which is 55 / 2.8346 = 19.4 mm
- Top: 70 - 5 = 65 points, or 22.9 mm
- Bottom: 80 - 5 = 75 points, or 26.5 mm
The page that comes out is 595.28 - 110 = 485.28 points wide and 841.89 - 140 = 701.89 tall, about 171 x 248 mm. The area kept is (485.28 x 701.89) / (595.28 x 841.89), or 68% of the sheet. On a tablet that is the difference between text you can read and text you have to pinch at.
Rotated pages and mixed page sizes
A PDF can store a page one way round and display it another: a sideways scan is usually stored portrait with a flag telling the reader to turn it 90 degrees. That matters here, because on such a page the edge you see at the top is the edge stored on the left. Crop PDF works from what you see, and leaves the rotation flag alone.
Documents assembled from several sources are often not uniform. The crop is held as a proportion of each page rather than a fixed number of millimetres, so the same trim lands sensibly on an A4 page and on an A3 chart and can never ask for more margin than a small page has. Where sizes differ, the tool says so.
If only some pages need it - a cover that should stay full-bleed, or scanned pages inside a digital report - switch Crop to Chosen pages and enter something like 2, 5-7. The rest keep their full size.
A crop is not a redaction
This is the part worth reading twice. Cropping hides content; it does not delete it. The words outside your box are still in the file, in the same untouched drawing instructions. Anyone who sets the page rectangles back gets them all returned, in seconds.
What makes this genuinely dangerous is that it often looks like it worked. Many readers and text extractors - including any built on the pdf.js engine, which is most browser-based ones - drop text falling outside the page box, so you can crop a signature block off a page, extract the text, and see it come back clean. It is not clean. The next tool will still find it.
So crop for margins, borders and framing. If something must actually go, delete it in whatever produced the document and export a fresh PDF, or drop the whole page with Remove Pages from PDF. Never crop a bank statement to hide an account number.
What cropping will not do
It will not shrink the file. Every piece of the original is still present, so a 4 MB PDF cropped to two thirds of its area is still close to 4 MB. For a smaller file run Compress PDF afterwards, or Compress PDF to Size when a form caps the upload.
It will not give you a standard paper size either - cropping produces whatever rectangle you drew, almost never exactly A4 or Letter. If you need a specific size, crop first, then run Resize PDF Pages to fit the result onto that sheet.
And it will not flatten anything. Pages are cropped in place rather than re-drawn, so text stays selectable and searchable, and links, bookmarks, form fields and comments stay exactly where they were.
Nothing leaves your browser
The preview is rendered and the crop applied entirely on your own machine, in the tab you have open. Nothing is sent anywhere - which matters, since the files that most need trimming are scanned contracts, statements and identity documents. Check it bluntly: load Crop PDF, disconnect from the internet, and crop a file. It still works.
Frequently asked questions
- Can someone recover the part I cropped off?
- Yes. A crop moves the page boundary; it does not delete anything, and the hidden content is still sitting in the file's drawing instructions. Anyone who sets the page rectangles back gets it all returned. It is worse than it sounds, because many readers and text extractors hide that content too, so a crop can look like a successful redaction when it is not. To genuinely remove something, delete it in the program that made the document and export a new PDF, or delete the whole page.
- Does cropping a PDF reduce the file size?
- Barely. Cropping changes the two rectangles that describe the page, not the content inside it, so every piece of text and every image is still there. Expect the file to come out within a percent or so of where it started. If you need it smaller, use Compress PDF or Compress PDF to Size on the cropped file afterwards - the two steps compose fine.
- Will cropping blur my text or images?
- No. Nothing is re-drawn or rasterised, so text stays vector-sharp and selectable and images keep their original pixels at their original resolution. That is the advantage of cropping in place over tools that redraw each page as a flattened graphic, which also lose links, bookmarks and form fields along the way.