N-Up multiple pages of a PDF in a single pass means placing either two or four source pages from one local PDF onto every page of a freshly built output document, sized to the first source page, with no upload and no server-side rendering. The output sheet box is taken from the first page's width and height, and that box is divided into a 2-by-1 grid for two-up or a 2-by-2 grid for four-up, with an 18-point outer margin and a 12-point gap between cells. Reading order runs left to right and then top to bottom, every embedded page is scaled by the smaller of its available width and height fits so it is never stretched, and mixed source sizes are handled per page. The work runs in the browser through the pdf-lib library, embedding each source page as a reusable form and drawing it into its computed cell, so vector shapes and text generally survive as PDF content rather than being rasterized into screenshots. Source files are limited to 25 MiB and 500 pages to keep the in-browser job bounded, and a slower earlier file read cannot overwrite a newer selection while the operation is in progress.
What "multiple files at once" actually means here
Search phrasing often blends two different needs: combining several separate PDFs into one imposition job, and imposing a single long, mixed-size source in a single operation. PDF N-Up: Multiple Pages Per Sheet handles the second case directly — you choose one local PDF, and every page of that file is placed on the output grid in one run. If the inputs you actually have are several separate PDFs, the practical browser workflow is to merge them locally first with a tool like Merge PDF, then run the N-Up tool on the combined document. That keeps the entire chain on your device, with nothing uploaded, while still letting you impose a multi-file source as if it were one file.
This distinction matters because the embedding step is per page, not per file. The tool walks the page list of one loaded PDF, computes a placement for each page against the cell dimensions of the grid it derives, and draws them onto sequentially numbered output sheets. Whether the source originally came from one file or from a previously merged set, the imposition behavior, the margins, the gaps, and the reading order are identical.
How the grid, margins, and fit are decided
The output sheet box is taken from the first source page: its width and height become the width and height of every sheet in the new document. There is no hidden paper-size preset. If the first page of your source is a US Letter portrait page, every output sheet is a US Letter portrait page; if the first page is an A4 landscape page, every output sheet inherits that exact A4 landscape box.
From that sheet box the tool derives the grid. Two-up uses two columns and one row; four-up uses two columns and two rows. Each grid reserves an 18-point outer margin, so embedded pages never sit on the cut edge of the sheet, and a 12-point gap between cells. The remaining usable rectangle is split into equal-width columns and equal-height rows, and each cell becomes the target for one embedded source page.
Each source page is scaled proportionally against its assigned cell. The tool computes an available-width scale (cell width divided by source page width) and an available-height scale (cell height divided by source page height), picks the smaller of the two, and applies that scale. The scaled page is then centered horizontally and vertically inside its cell. A tall source page therefore keeps its aspect ratio with space at the sides, while a wide source page can leave space above and below. Source pages are never stretched to fill their cell, even when that would visually appear to fill the cell better.
| Property | 2-up | 4-up |
|---|---|---|
| Grid layout | 2 columns x 1 row | 2 columns x 2 rows |
| Source pages per output sheet | 2 | 4 |
| Outer margin | 18 pt | 18 pt |
| Gap between cells | 12 pt | 12 pt |
| Reading order | Left to right | Left to right, then top to bottom |
| Fit behavior | Smaller of width and height scales, centered | Smaller of width and height scales, centered |
| Output sheet dimensions | Width and height of first source page | Width and height of first source page |
How to N-Up multiple pages of a PDF at once
- Open PDF N-Up: Multiple Pages Per Sheet in a current desktop browser and choose a non-empty PDF from your device that is no larger than 25 MiB. Encrypted, damaged, empty, or mislabeled files are rejected with a visible error rather than partially processed.
- Select the arrangement: two source pages per sheet for a 2-by-1 output, or four source pages per sheet for a 2-by-2 output. There is no separate grid picker; the two options map directly to the two supported grids.
- Confirm the source-to-sheet count the tool reports, for example that a 9-page source in 4-up produces three output sheets with the last sheet holding one embedded page, so you know what you are about to download.
- Select Create N-up PDF. While the operation runs, the control stays disabled so the same file is not read twice and a slower earlier read cannot overwrite a newer selection.
- When the run finishes, the tool displays the source page count, the output sheet count, and an approximate encoded size. Download the new PDF; the filename reflects the chosen arrangement, such as report-4-up.pdf.
- If you change either the file or the arrangement, the old result is invalidated and its temporary download URL is revoked. Select Create N-up PDF again to produce a fresh file.
Reading order, mixed sizes, and partial final sheets
Pages are assigned to cells in the same order they appear in the PDF, going across first and then down. In a 4-up layout that means cell 1 (top-left), cell 2 (top-right), cell 3 (bottom-left), cell 4 (bottom-right), then the next sheet. This matches normal reading flow for a left-to-right script and keeps the logical page sequence intact for handouts and review packs.
When the final sheet is not full, the remaining cells are left empty rather than redistributed. A 5-page source in 4-up produces two output sheets: the first sheet holds four pages in all four cells, and the second sheet holds one page in the top-left cell with the other three cells blank. There is no automatic rotation, no rearrangement, and no attempt to spread leftover pages across more output sheets than the reading-order path naturally produces.
Mixed source page sizes are supported. Because the fit calculation is applied per page rather than once for the whole document, a landscape diagram page and a portrait text page can sit on the same output sheet, each centered in its own cell and scaled by its own smaller fit factor. This is the practical reason the tool walks the document page by page instead of treating the source as one homogeneous shape.
Limits, errors, and what the tool does not do
The tool is bounded on two axes: a 25 MiB maximum source size and a 500-page maximum source length. Anything outside that window is rejected with a visible error. Empty files, mislabeled files (a JPG renamed to .pdf, for example), encrypted documents, damaged PDFs, and unsupported structures also fail with a clear error rather than being silently rewritten. For the practical workarounds that fit inside the cap, the N-Up a Large PDF guide explains the options in more detail; for what to verify before you send the output to a printer, the PDF N-Up for Printing walkthrough covers the print-readiness checks.
The focused N-Up tool deliberately does not impose duplex booklet signatures, add crop marks, rotate pages, reorder source pages, remove annotations, flatten form fields, preserve interactive links as repositioned clickable regions, print directly, or guarantee a specific commercial printer workflow. It also does not alter the original file. If your real need is to combine several separate PDFs first and then impose the result, run them through Merge PDF in your browser and feed the merged file into the N-Up tool.
Local processing and the quality of the output
Loading the source, embedding every page through the embedPage API, drawing each page into its computed cell, serializing the resulting document, and creating the download all happen locally in the browser. The source PDF is not uploaded. The output keeps vector shapes and text as PDF content rather than rasterizing pages into screenshots, which keeps the file small relative to a per-page image export and keeps the text selectable in the new document. After the run, if you want to shrink the result further you can pass it through Compress PDF in the browser, and if you want the output to carry page numbers for distribution, Add Page Numbers to PDF stamps them on without leaving the browser.