Crop PDF Online Trim the margins off a scan, and see exactly what survives

How to Crop a PDF Online for Free

Two minutes, nothing to install, and no account needed for your first run. This PDF cropping tool executes inside the page you are reading, so all four steps happen on your own machine.

1. 1

   Drop your PDF into the emerald zone

   Click the drop zone above or drag the file anywhere onto this page. Every page appears as a tile, and each one will show the crop frame as soon as you set it.

2. 2

   Open a page and drag the frame

   Click any tile and the page fills the screen with a crop frame on it. Drag the frame, drag its handles, or type the trim for each edge in millimeters. The area outside the frame is grayed out, so what you are about to lose is never a guess.

3. 3

   Or let the tool find the margins

   Press Detect the margins and every page is examined for where the ink actually starts and stops, then trimmed to it with a hair of padding. On a scan with wide white borders this is usually the only step you need.

4. 4

   Choose one crop or one per page, then run it

   Leave One crop for the whole document on and the frame applies everywhere, which is what a uniform scan wants. Switch it off to crop pages individually. Then press Crop PDF and the file comes back immediately.

The frame carries a rule of thirds grid and a live millimeter readout, and the handles are sized for a fingertip instead of a mouse pointer, because a large share of the people cropping a document are doing it on a phone with the scan somebody just emailed them.

Cropping a PDF Does Not Delete Anything

This is the single most important thing on this page and almost nobody says it. Cropping a PDF changes which part of the page is displayed. It does not delete the content outside the frame. The text, the images and the marks you trimmed away are all still sitting in the file exactly as they were, and anybody who opens the document in different software, or pulls it apart programmatically, can still read every one of them. They are outside the region a reader was told to paint. Nothing more.

That is not a limitation of this tool. It is what cropping means in the PDF format, in every application that implements it, including the expensive ones. A page is drawn in an infinite coordinate space, and a small set of rectangles in the page dictionary decides which part of that space a viewer actually paints onto your screen. Cropping moves the rectangles.

Four consequences follow, and all of them matter before you rely on it:

• Anybody can restore the original view. Setting the boxes back to the original values brings the trimmed area straight back, and there are free tools that do it in one command.
• Text extraction still finds the cropped-away text. Copying all the text out of a cropped page returns the words in the margins as well. So does any search index that reads the file.
• The file does not get smaller. Nothing was removed, so nothing was saved. A cropped 40 MB scan is still a 40 MB scan.
• Printing honors it correctly. A printer prints what the crop box says, so a cropped page prints cropped, which is the case people actually care about.

Cropping is the right tool for making a document readable, printable and pleasant. It is the wrong tool for making something disappear, and being clear about which job you are doing takes ten seconds.

The Engineering Behind a PDF Crop: Five Boxes, Not One

A PDF page can carry five rectangles, and the difference between a crop that works everywhere and one that works in your reader only is which of them a tool bothers to set.

Why setting only CropBox is not enough

Most online croppers set CropBox and stop. In your browser the page looks perfectly cropped, which is why the bug survives. Then the file is sent to a print shop, or opened in a layout application, or run through an imposition tool, and the original margins come back, because those tools read MediaBox and TrimBox and the file still says the sheet is A4.

This tool sets all five to the same rectangle. That is what "cropped" has to mean if the word is going to survive contact with a printer:

• MediaBox so the file honestly reports its new sheet size to every print path.
• CropBox so every reader displays it correctly.
• BleedBox, TrimBox and ArtBox so prepress and layout software agree with the other two rather than reverting to the original geometry.

The content is not rewritten

Nothing inside the page is touched. The content streams, the fonts, the images and the annotations are all byte identical to the ones you dropped in, and only the box entries in the page dictionary change. That is what makes the operation instant on a two hundred page document and what keeps the text layer, the links and the bookmarks working afterwards. It is also why the file does not get smaller: cropping deletes nothing, it only changes what is displayed.

Rotated pages, handled properly

A page stored sideways with a rotation flag is displayed upright, so the frame you drag is not aligned with the coordinate system the boxes are written in. Both corners of your frame are mapped from the page as you see it into the file coordinate system and then normalized, so a rotated page crops where you pointed and not ninety degrees away from it. Other tools fail here constantly. You usually find out after sending the file.

How Detect the Margins Works, and When to Trust It

Pressing Detect the margins renders each page to a small bitmap, walks every pixel looking for anything that is not paper, and trims to the bounding box of what it finds, plus about one percent of the page as breathing room.

Three design choices in that sentence are worth explaining, because they decide when it works:

• The page is rendered small, around 700 pixels on the long edge. Detection needs the bounding box of the ink, not the detail of it. At that size a one millimeter rule is still several pixels wide and reliably found, and the whole pass takes milliseconds per page instead of seconds.
• The tolerance from pure white is generous. Scanned paper is never actually white. It is a light grey with a colour cast that depends on the lamp, and a strict test would decide the entire sheet is content and trim nothing at all.
• A blank page returns nothing rather than a crop. A page with no ink on it has no bounding box, and cropping it to a point would be worse than leaving it. It is left alone.

When it works well

• Scans of a smaller document on larger paper, which is the classic case: an A5 booklet or a receipt scanned on an A4 flatbed.
• Book scans with wide white borders around the type block.
• Anything exported with generous default margins that nobody wanted.

When it will disappoint you

• Scans with a dark edge. A flatbed with the lid open produces a black border, and a photograph of a page includes the desk. Both are ink as far as the detector is concerned, so it trims nothing. Drag the frame by hand instead.
• Pages with a header, a footer or a page number you wanted gone. Those are ink and they will be kept. Crop by hand if the point was to remove them.
• Speckled or dusty scans. A single dark speck in the corner is the bounding box. If the result looks like nothing was trimmed, that is usually why, and cleaning the glass is a more permanent fix than any software.
• Documents where the pages differ. With one crop for the whole document, the first page is measured and its frame is applied everywhere. Switch that off and every page is measured on its own, which takes longer and is correct for a mixed document.

When Cropping Is the Right Answer, and When It Is Not

Cropping solves a specific set of problems well and is a poor substitute for several other operations. Knowing which is which saves a lot of time.

Genuinely good reasons to crop

• A small document scanned on big paper. A receipt or an A5 leaflet on an A4 flatbed wastes half the sheet and makes the content tiny on a phone screen. Cropping to the content makes it readable.
• Reading on a tablet or an e-reader. This is the strongest case there is. A journal PDF with 30 mm margins shows the type at half the size it could, and cropping the margins away is the difference between a readable article and a squint.
• Double page book scans. Split each spread into two pages, or crop to one side at a time, so the result is a page per page rather than a spread per page.
• Removing scanner artifacts. The black strip down one edge, the shadow of the spine, the fragment of the next page.
• Presenting one region. A drawing set where only the title block matters, or a spreadsheet export where only the summary table is needed.

Problems that look like cropping and are not

One Crop for Everything, or One Per Page

The switch marked One crop for the whole document decides whether the frame you set applies everywhere or only to the page you set it on. Both are correct for different documents and the default is uniform, because most documents that need cropping are scans where every sheet went through the same feeder in the same position.

The frame is stored as a proportion of each page instead of as an absolute rectangle, which is what makes the uniform mode meaningful. Trim eight percent off the left edge and every page loses eight percent of its own width, so a document with an A3 foldout in the middle gets a proportionally identical crop rather than one that eats half the foldout.

When to switch it off

• A document assembled from several sources, where the margins genuinely differ page to page.
• A book scan where left hand and right hand pages have the binding on opposite sides, so the crop needs to mirror.
• Anything with a mix of portrait and landscape pages where the wasted space is in a different place on each.

With it off, opening a page and setting a frame affects only that page, and Detect the margins measures every page individually. On a long document that takes noticeably longer, because every page has to be rendered, and it is the correct trade for a document that genuinely varies.

The switch appears in two places, in the panel and inside the full size editor, and they are wired to each other instead of kept separately. Changing it in one place changes it in the other immediately, because two controls for one setting is how a tool ends up cropping a single page while the switch on screen says all of them.

Zero-Trust: Your Document Never Reaches Our Server

The usual architecture for a free crop PDF online service is an upload form, a queue, a command line utility on a server you do not control, and a temporary download link. Your document crossed the internet. It sat on a disk you cannot inspect for however long a privacy page claims. Look at what people actually crop: scans of things that were on paper for a reason, which means contracts, deeds, medical letters, bank statements and identity documents.

There is no upload form here, and no queue behind it. Your file opens through the browser File API. A PDF library running as JavaScript on your own processor rewrites the page boxes, and the cropped document arrives as a Blob your browser downloads from itself. The file exists on your disk and in your RAM, and nowhere else.

Verify it in thirty seconds

1. Open developer tools, switch to the Network panel, and clear it.
2. Crop a document, including a run of Detect the margins. Watch the requests.
3. You will see the engine scripts load once and one small JSON call carrying the tool name. No request contains your PDF, and none is remotely the size of one.
4. The definitive version: load this page, disconnect from the network completely, then crop anyway. It works, because there is no server-side step for it to depend on.

Detect the margins is worth calling out here, because it is the most computationally expensive thing on the page and it is entirely local: every page is rasterized and every pixel examined on your own processor. That is why a four hundred page document takes a moment on a laptop, and it is the same reason nothing about your document is knowable to us.

When a Crop Goes Wrong, and What It Means

Before cropping a long document

1. Set the frame on page one and open two or three other pages at full size to check it.
2. Look specifically at the page with the most content, not the emptiest one. The tight page is the one that gets clipped.
3. Run it, then open the result and check the same pages again.
4. Keep the original. Cropping is reversible in principle and awkward in practice, and the original costs you nothing to keep.

What People Usually Do Straight After Cropping

Cropping is normally one step in cleaning up a scan. These are the ones that come with it:

• The pages are sideways. Rotate PDF turns them and saves the rotation permanently, and it is worth doing before cropping so the frame is drawn on an upright page.
• The pages are out of order, or some are blank. Organize PDF and Remove Pages deal with both.
• Each spread should be two pages. Crop to one side, then to the other, and Merge PDF to interleave the results.
• It needs numbering now that it is a proper document. Add Page Numbers.
• It should be images rather than a document. PDF to JPG and PDF to PNG export the cropped region, since they render what the crop box says.

Frequently Asked Questions

The rest of the PDF suite

All of these run the same way: your file is read by JavaScript on your own machine, the work happens on your processor, and the result is handed back to you as a download that never existed anywhere else.