How to scan a two-sided document into one PDF
If your document feeder reads both faces in one pass, a single setting does the whole job and the pages land in the file in order. If it does not, you scan the fronts, turn the stack over, scan the backs — and the backs arrive last-sheet-first, which is the part nobody warns you about. Both routes are below, with the switch names for HP, Brother, Canon, Epson and Kyocera, and the fix for a stack that came out backwards.
Merge PDF →First find out whether the machine can do it at all
Two-sided printing and two-sided scanning are separate pieces of hardware. A printer that happily prints on both faces of a sheet may still read only one face at a time, and that mismatch is behind most of the forum threads on this subject.
The feeder decides, not the glass. Duplex scanning happens in the automatic document feeder — the tray you drop a stack into. The flatbed glass reads one face by definition, because only one face is lying on it. If you are working on the glass, there is no two-sided setting hiding somewhere; there is a second pass and a bit of page shuffling, covered further down.
- Read the spec sheet wording. «Duplex ADF», «single-pass duplex» or «two-sided scanning» means the machine reads both faces as the sheet travels. «Duplex printing» on its own tells you nothing about scanning.
- Set the source before you look for the switch. On nearly every driver the two-sided option only appears once the source is the feeder, and stays invisible while «Flatbed» is selected. This is the single most common reason people conclude their machine cannot do it.
- Watch for the two-pass variant. Some models scan the whole stack, then ask you to put it back in and scan the other faces. That still counts: the firmware interleaves the pages for you. It just needs your hands halfway through.
The switch, make by make
The feature carries a different name on every panel, which is exactly why people who own the right hardware still cannot find it.
- HP Smart. Open the Scan tile, set Source to Document Feeder, and the 2-sided (duplex) toggle appears among the scan settings. On enterprise LaserJet and PageWide panels the equivalent lives under Copy or Scan → Options → 2-sided.
- Brother, control panel. Scan → your destination (to File, to Email, to PC) → 2-sided Scan → DuplexScan: LongEdge or DuplexScan: ShortEdge. Long edge for ordinary portrait paperwork; short edge for pages bound along the top, like a flip pad. The option is unavailable for the glass.
- Epson Scan 2. Document Source: ADF, then Scanning Side: Double-Sided. In the older Epson Scan the same thing is a source entry — «ADF - Double-sided» — and it shows up only in Office or Professional mode.
- Canon MF Scan Utility. The checkbox is spelled out in full: «Scan both sides when scanning from the feeder». Tick it in the scan settings for Document, Photo or Custom before you press the scan button.
- Kyocera. Functions → 2-sided/Book Original → 2 side, then set the binding direction. The neighbouring choice, Book, is for a magazine spread and splits two facing pages into two images.
- Windows Scan app. The Source list distinguishes «Feeder (Scan one side)» from «Feeder (Scan both sides)». The second entry exists only when the driver reports duplex support; the older Windows Fax and Scan usually hides it.
- macOS. Image Capture and the scan pane in Printers & Scanners both expose a Duplex choice once Use Document Feeder is ticked, and Preview shows the same thing when it scans.
Result: one PDF, pages in reading order — front, back, front, back — and nothing left to fix. If that is what you got, the rest of this page is not for you.
No duplex feeder: two passes, and why the backs come out backwards
The manual method is simple to describe and has one counterintuitive consequence worth understanding before you start, because it explains an hour of confusion later.
Turning a stack over reverses the order in which the sheets travel. The sheet that went through first now goes through last — and that holds whether your feeder pulls from the top of the pile or the bottom. So pass one hands you pages 1, 3, 5, 7 in order, and pass two hands you 8, 6, 4, 2: the backs, last sheet first.
The routine itself:
- Load the stack as you normally would and scan it to a PDF. Call it fronts.
- Take the whole stack out without shuffling it and turn it over as one block, left to right, the way you turn a page in a book.
- Put it back in and scan again, to a second PDF. Call it backs.
- Check both page counts. Each file should have exactly as many pages as there are sheets. A short count means two sheets travelled together; a long one means a sheet went twice.
Zipping the two files into one document
You now have two PDFs and one job: get the faces back into their proper alternating sequence.
Start with Merge PDF: add fronts, then backs, and you have a single file — all the odd pages, then all the even ones in reverse. The merge runs in your browser, so a stack of payslips or medical notes never leaves the machine you are sitting at.
Then open Rearrange PDF and drag the pages into place. For five sheets the target sequence is 1, 10, 2, 9, 3, 8, 4, 7, 5, 6 — first front, last back, second front, second-to-last back, and so on inward. It looks strange written down and becomes automatic by the third document.
Be honest with yourself about the size of the job. Up to about a dozen sheets this is a two-minute chore. At forty sheets it is a mistake waiting to happen, and there is software that does the interleave in one action instead:
- NAPS2 (free, Windows) has Reorder → Interleave and Reorder → Alternative Interleave. The alternative one is built for precisely the reversed-backs case, and the same behaviour is available from its command line as --altinterleave.
- Epson Scan 2 owners have a shortcut that skips the second file entirely: tick «Add or edit pages after scanning» in the save settings, scan the fronts, reload the stack, scan the backs, then move and rotate pages inside the same job before saving it.
- Qoppa PDF Studio offers an Insert At: Interleave option at scan time, and VueScan carries its own two-sided workflow. Both are paid desktop programs, and both are the right answer if you do this every week.
The three things that go wrong afterwards
Every second page is upside down. You turned the stack end over end. Do not scan it again — open Rotate PDF, tick the affected pages, turn them 180° in one click and save. The rotation is written into the page itself, not into the view, so it survives being emailed to somebody else; the reason that distinction matters is unpicked in fixing an upside-down PDF.
Every second page is blank. A duplex pass over single-sided originals produces a blank back for every sheet, and forms with a printed reverse side mixed into the same stack make it worse. Delete PDF pages shows the pages as thumbnails so you can tick the empty ones and drop them in one go; on a long document, sorting the order first with merging in the right order saves a second pass over the thumbnails.
A page is missing entirely. Two sheets stuck together and travelled as one. The page count catches it: a two-sided stack should always produce an even number of pages, twice the number of sheets. Fan the paper before loading it, and check the corners of stapled documents where the holes leave a burr.
Before you send it
One last look at the file, because a document that is right and unopenable is no better than a wrong one.
Weight is the usual problem. A colour A4 page scanned at 300 dpi lands somewhere between half a megabyte and a megabyte and a half, so a twenty-sheet contract at forty pages can reach thirty megabytes — past the twenty-five megabyte ceiling most mail gateways still enforce. Run it through Compress PDF after you have checked the pages, not before, or you will redo the work when page seven turns out to be blurred.
If the file is heavier than it should be for its length, the scanner was probably set higher than the job needs; what dpi to scan documents at has the numbers for choosing once and forgetting about it. And if what you are digitising is a card rather than paperwork, both faces belong on a single sheet instead of two pages — that is a different routine, set out in copying both sides of an ID card onto one page.
Common questions
Why does my printer print on both sides but not scan both sides?
Because those are two different mechanisms. Duplex printing flips the sheet inside the print path; duplex scanning needs a feeder that reads both faces as the sheet passes, or one that pulls the stack through twice. Plenty of home models have the first and not the second, and the spec sheet is the only reliable place to check.
My scan has a blank page between every page. What happened?
The machine scanned in duplex mode over originals printed on one side only, so it recorded the empty back of each sheet. Either switch the setting back to one-sided and rescan, or drop the blanks with Delete PDF pages — for a handful of pages the second option is faster.
The back sides are upside down. Do I have to scan them again?
No. Select those pages in Rotate PDF, turn them 180° and save. One click covers all of them, because you can tick several pages and rotate them together.
Can I do a two-sided scan on the flatbed glass?
Not in one action — the glass sees one face at a time. Scan the fronts, then the backs, then merge and reorder. On the glass you at least control the sequence yourself, so the backs need not come out reversed.
How many pages can I merge and reorder in the browser?
There is no limit set by the tools. The work happens on your device, so the ceiling is memory rather than a server quota; a couple of hundred scanned pages is routine on a current laptop. A very large batch on an old phone is easier as two documents joined at the end.
Do the files get uploaded while I merge them?
No. Both tools read the PDFs from your device, do the work in the page and hand the result back as a download. Nothing is transmitted, which matters when the stack is a set of payslips or a signed contract.