How to Merge, Split and Convert Images to PDF Without Uploading
Merging, splitting and turning images into PDFs are all structural rewrites. They move page objects around without understanding the content, so your browser can do the whole job and the files never leave your device. The hard parts are mismatched page sizes, page-range syntax and image resolution.
- Merging never normalises page size, because every page keeps its own MediaBox and an A4 and Letter mix produces a document with pages of different dimensions
- PDF measures in points at 72 per inch, so A4 is 595.28 × 841.89 pt and Letter is 612 × 792 pt, making A4 16.7 pt narrower and 49.9 pt taller
- Page ranges accept comma-separated values such as 1-3,5,7-9; reversed ranges like 9-7 are corrected automatically and a repeated page keeps only its first position
- Image to PDF puts one image per page: A4 mode scales and centres the image and leaves a 24 pt (about 8.5 mm) margin on all four sides
- Original size mode converts pixels to points at 96 DPI, so a 4000 px wide photo becomes a 3000 pt wide page
- Merging, splitting and image conversion all run in the browser with pdf-lib, so nothing is uploaded and the tools keep working offline
#Why these operations can run in your browser
Merging, splitting and converting images to PDF only rewrite document structure, so they run entirely in the browser and upload nothing. Each one is small bookkeeping: copy page objects into a new document, pick pages by index, or create a page and draw an image on it.
| Kind of operation | Examples | Can it run locally? |
|---|---|---|
| Structural rewrite | Merge, split, reorder, embed images | Yes, existing objects are copied as-is |
| Content understanding | OCR, text reflow, format conversion | Usually needs a server-side model |
No upload means no copy on someone else's disk, and it means the tools work offline. For contracts, invoices or ID scans, that difference is the whole point. To verify it, open the browser network panel, process a PDF and look for requests carrying your file — there are none.
#What happens when page sizes differ
Merging does not normalise page size. Every page keeps its own MediaBox, so an A4 and Letter mix produces a document whose pages differ in size. Merge PDF copies pages one at a time; nothing is rescaled or reflowed.
| Paper | Millimetres | Points (72 per inch) | Ratio |
|---|---|---|---|
| A4 | 210 × 297 | 595.28 × 841.89 | 1 : 1.414 |
| Letter | 215.9 × 279.4 | 612 × 792 | 1 : 1.294 |
| A5 | 148 × 210 | 419.53 × 595.28 | 1 : 1.414 |
A4 is 16.7 pt narrower and 49.9 pt taller than Letter. Print with "fit to page" and the reader scales each page to the paper separately, so the font size visibly changes from one page to the next.
Before merging, check the page count and size shown on each row so no encrypted file slipped into the list. Encrypted PDFs may fail to copy, and the status bar reports how many were skipped. Do not ignore that number.
There are only two fixes: normalise page sizes in the source files before merging, or accept the mix and print at "actual size" so nothing gets rescaled a second time.
#Splitting works by range, not by equal parts
Splitting only works by page range. There is no "divide into N equal parts" mode. To split by count, work out each range yourself and extract them one at a time.
Split PDF reads the total page count and fills in every page by default. The ways to fill the field cover different jobs:
| Job | Range syntax | Notes |
|---|---|---|
| Take pages 1 to 3 | 1-3 | A single range |
| Take odd pages (fronts of a duplex scan) | Click Odd pages | Fills 1,3,5,… |
| Take even pages (the backs) | Click Even pages | Fills 2,4,6,… |
| Assemble a few pages into one file | 1-3,5,7-9 | Comma-separated, mixed freely |
Output order follows what you typed, not the order in the original file. Typing 9,1-3 gives you page 9 first and pages 1 to 3 after it. When a page appears twice, only its first position is kept.
1-3,5,7-9 -> pages 1,2,3,5,7,8,9, 7 pages in total
9-7 -> corrected to 7-9, giving pages 7,8,9
1-3,2-4 -> pages 1,2,3,4, the repeat of 2 and 3 is droppedA page number beyond the document, or an empty field, raises an error rather than producing an empty PDF.
#Image to PDF: DPI and page fitting
Image to PDF offers two page modes. Fit to A4 scales each image proportionally, centres it and adds a margin; Original size converts pixels to points at 96 DPI and keeps the image ratio. Multiple images become one page each, never several to a sheet.
With Image to PDF, decide the destination first, because the two modes diverge quickly:
| Mode | How the page is sized | Best for |
|---|---|---|
| Fit to A4 | Fixed 595.28 × 841.89 pt, image scaled and centred with a 24 pt margin | Printing, archiving, one paper size |
| Original size | Width and height = pixels × 0.75 pt, i.e. 96 DPI | Images that already have a known physical size |
Resolution decides how sharp the print looks, and A4 needs very different pixel counts depending on the target:
| Target DPI | Pixels needed for A4 (210 × 297 mm) | Notes |
|---|---|---|
| 72 | 595 × 842 | Screen reading only; text looks soft |
| 150 | 1240 × 1754 | Minimum for ordinary printing |
| 300 | 2480 × 3508 | Print standard, use it for photos and scans |
Prepare at least 300 DPI of real pixels for anything you plan to print, which means 2480 × 3508 for A4. A PDF page stores physical dimensions rather than pixels, so raising DPI after conversion adds no detail at all.
Images are embedded as JPEG and transparent areas are filled white. A transparent PNG logo therefore arrives on a white background, and screenshots with sharp text and flat colour can pick up visible blocking.
Photographing a document with your phone? Straighten it, check for stray fingers, crop the extra edges with Image Cropper, then convert. Doing it in the other order leaves you no way to fix the crop inside the PDF tool.
#Which tool for which job
| What you need | Tool | Notes |
|---|---|---|
| Combine several PDFs into one | Merge PDF | Reorder, remove, per-file page count and size |
| Pull pages out of a PDF | Split PDF | Ranges, odd pages, even pages, custom selection |
| Turn photos or scans into a PDF | Image to PDF | JPG/PNG/WebP, A4 fit or original size |
| Trim edges before converting | Image Cropper | 1:1 / 4:3 / 16:9 presets |
| Shrink images so the PDF is smaller | Image Compressor | Reduce pixels before embedding |
| Replace a transparent background with white | Image Converter | JPEG fills transparency with white |
#Frequently asked questions
#The merged PDF has pages of different sizes. Is it corrupted?
No, that is normal. Merging copies page objects without changing any page's dimensions, so an A4 and Letter mix stays mixed. For a uniform document, normalise the page sizes in the source files before merging. Printing with "fit to page" is a temporary visual patch, not a fix.
#Can I split a PDF into N equal parts?
Not directly, because the tool works by page range. To split 100 pages into four, run it four times with 1-25, 26-50, 51-75 and 76-100, then rename each output. The total page count is displayed as soon as the file loads.
#Why is the page order different from the original file?
Because output order follows your input. 9,1-3 puts page 9 first with pages 1 to 3 after it. A reversed range such as 9-7 is corrected to ascending order, and a repeated page keeps only its first position.
#Should I use 300 DPI or 150 DPI for image to PDF?
It depends on the destination. For printing or print production, use 300 DPI, which means 2480 × 3508 pixels for A4. For screen-only reading, 150 DPI is enough. Page size in a PDF is physical, so raising DPI after conversion adds no real detail — get the pixels right before you convert.
#Can encrypted PDFs be merged or split?
It depends on the encryption. The tools request the file with the encryption flag ignored, so many protected PDFs still open, but a document with strict content encryption fails or gets skipped when pages are copied. The status bar reports how many files were skipped. If that happens, decrypt and save a copy locally first.
#Do these tools upload my files?
No. Merging, splitting and image conversion all run in the browser using pdf-lib, so nothing is uploaded and the tools work with no connection. The browser network panel confirms it — no request carries your file contents.