How to combine scanned documents into one PDF

Why scanned PDFs behave differently

A scanned PDF has almost nothing in common with a generated one, and knowing the difference explains every symptom below.

• Every page is one large image. There is no text, no font, no vector data. So there is nothing to compress intelligently and nothing to select.
• Page size is whatever came off the glass. A stack fed through an automatic document feeder can produce subtly different dimensions per page, and a mix of Letter and A4 originals produces a document whose pages visibly disagree with each other.
• Automatic feeders introduce their own artifacts. Blank backs from double-sided originals, a slight skew on each page, occasional double-feeds, and the order you expected only if the originals were perfectly stacked.
• The size is dominated by resolution and colour depth. A 300 dpi colour page is roughly four times the data of a 150 dpi page and about three times the data of a greyscale one at the same resolution. Those two choices are made at the scanner, before any tool of ours sees the file.

Step 0 — get the scan right, because nothing later can repair it

You can fix order, rotation, blank pages and file size after scanning. You cannot fix a scan that was made at the wrong resolution or in the wrong colour mode. These settings are worth thirty seconds of attention.

Two more settings worth changing from the default. Scan straight to PDF rather than to JPEG — you skip a conversion step and avoid JPEG damage on text. And if your scanner offers automatic blank-page removal, leave it on; it handles the single most common cleanup job before you ever open a merge tool.

Step 1 — name the files so the order is obvious

This is the step people skip and then regret. Scanners typically produce names like SCAN_0001.pdf, which sort correctly on their own. But files assembled from several sessions — two batches, a phone photo, an emailed page — rarely do, and filenames like scan.pdf, scan (1).pdf, scan (2).pdf will sort in an order nobody expects.

Rename before merging, using zero-padded numbers so the sort is unambiguous: 01-contract-p1.pdf, 02-contract-p2.pdf, and so on. Padding matters — 1, 10, 2 is the classic wrong order, while 01, 02, 10 is not. It costs a minute now and saves re-doing the merge.

You can also fix the order after merging by dragging page thumbnails, so this is not a hard dependency. It is simply much faster to get it right before the pages are interleaved.

Step 2 — merge them

Add all the files at once, drag the list into the order you want, and merge. Because the work happens in your browser there is no page cap, no daily quota and no file-size limit imposed by a server — the only bound is your device memory. Adding thirty scans at once is a normal thing to do, and the merge produces a single PDF with the pages in exactly the order shown in the list.

If you are combining scans with a PDF that was not scanned — a digital invoice alongside paper receipts, say — include it in the same run. The pages are copied as complete objects, so a digital page keeps its text layer while the scanned pages around it remain images.

Step 3 — deal with the size, in this order

Twenty pages of colour 300 dpi scans can easily exceed 60 MB, which is past the attachment limit of most mail systems and past what many portals accept. Resist the urge to jump straight to compression, because the cheap wins come first.

• Crop the scanner bed away first. A grey strip down one edge and a wide white border around the content are pure waste, and trimming them is a lossless change that also makes the pages easier to read on a phone. Use uniform margins when you know the geometry, or automatic whitespace detection when the pages are inconsistently centred.
• Then compress, if it is still too large. For a document that will be read rather than printed, 150 dpi with JPEG quality around 80 is visually indistinguishable on a screen and commonly cuts a scan stack by 60–90%. The trade-off is real and worth stating: compression rasterises the pages, so any text layer stops being selectable. Since a scan had no selectable text to begin with, you lose nothing at all — this is the one case where compression is close to free.
• Do not compress if it will be OCRed or filed as evidence. Both need the resolution. Trim the margins, leave the rest alone, and send it by link instead of by email.

The order matters because cropping and compression stack. Trimming a 12% border off each page and then compressing gives you a smaller file than compressing as-is, and the border removal costs no quality at all.

Step 4 — fix the pages you just merged

The merged file is now complete, in order, and small enough to send. It still needs a pass over the pages, and this is where a document stops looking like a scan job and starts looking like a finished file.

• Delete the blank backs. Double-sided originals fed through an automatic document feeder produce a blank page for every single-sided sheet. In a fifty-page merge that is a dozen wasted pages, and they also inflate the file.
• Turn the sideways ones upright. A landscape insert, a fold-out table, a page that went through the feeder rotated. Rotation is applied per page, so one page can be corrected without touching the rest.
• Re-check the order now that it is one file. Thumbnails make an interleaved section obvious in a way a file list never does.
• Normalise the page sizes last. A feeder that pulled a Letter original through with A4 scans leaves pages that visibly disagree with their neighbours. Resize PDF Pages gives every page the same target size without re-scanning, and because the pages are rescaled rather than re-rendered, the scans themselves are untouched.
• Add page numbers if the document will be referenced. A merged exhibit bundle, a filed appendix or a printed report needs numbering, and adding it to the finished file means the numbers match the final page count.

If you have images rather than PDFs

Phone photos of documents are the other common starting point, and the route is different. Scanning a page with a phone produces a JPEG, and a JPEG has no page size at all — so it needs a decision about the page it should live on.

Build the PDF from the images first, choosing a page size and orientation, and then merge the result with your existing PDFs if you need to. Fitting each image to its own page preserves the original proportions and is the safer default for photos taken at an angle or with irregular framing. Forcing everything onto A4 or Letter gives uniform pages, which prints better but will letterbox images whose aspect ratio does not match.

Common problems and their fixes

Nearly every complaint about a combined scan is one of these six.