Working with PDFs: merge, split, compress, rotate, convert

The one idea behind all of it: lossless vs lossy

Before the individual tools, one distinction makes everything else make sense. Most PDF operations are lossless — they rearrange, copy, or restructure pages without touching the actual content. Merge, split, and rotate all fall here: text stays selectable, images are never re-encoded, and the output is byte-equivalent to the input apart from how pages are ordered or oriented. You can run these as many times as you like with zero decay.

The exception is compression, which comes in two flavours. Lossless compression removes waste — uncompressed streams, duplicate resources — with no visible change. Lossy compression (downsampling images) throws away pixel detail to make a file much smaller, and that's the only operation here that can reduce quality. Keeping this split in mind tells you when you can be carefree and when to keep an original.

Merge: many files into one document

Merging joins several PDFs into a single, ordered document. It's the tool you reach for when a thing that should be one file arrived as several: a contract body plus its annexes and a signed page, a period of invoices a client wants as one attachment, or a job application that needs a CV, cover letter and portfolio bundled before a portal will accept it. The merge is lossless — pages, embedded fonts and vector graphics carry across untouched — so nothing is degraded.

The one thing to get right is order: the merged file follows the sequence of the thumbnails, top to bottom, so arrange them before you download. Naming source files with a numeric prefix (01-cover.pdf, 02-body.pdf) keeps them predictable. Because merging is lossless, the result weighs roughly the sum of its parts, so if it needs to fit an email limit, compress it afterwards.

Split: one document into its parts

Splitting is merging's opposite: it takes one PDF and turns each page into its own file, delivered as a single ZIP archive. Use it when you only need part of a document — one invoice out of a scanned batch, the signature page of a contract, a single chapter of a report, or the one form buried in a government packet. Each page is copied out losslessly and named with a zero-padded index (page_001.pdf, page_002.pdf) so the files always sort in the right order.

Split and merge are a powerful pair. To keep pages 5 to 8 of a document as a single file, split it, take page_005 through page_008 from the ZIP, and merge those four back together. The same routine handles anything the other tools can't do per-page — including selective rotation.

Compress: make a heavy file fit

Compression shrinks a PDF so it fits an email cap, uploads faster, or costs less to store. The lossless pass — the safe default — deflates uncompressed streams, drops duplicated resources, and rebuilds the file structure, with zero visible change. Office exports from Word or LibreOffice often shrink 30–60% this way because they ship their content uncompressed; already-optimized files barely move because there's little waste left to remove.

Scans are the exception. A scan is a full-page image, so its weight is pixel data that lossless compression can't touch. To meaningfully shrink a heavy scan you have to downsample — lower the image resolution, a lossy trade that softens the picture but can turn a 40 Mb file into a 4 Mb one. For born-digital documents, the lossless pass is almost always enough on its own.

Rotate: fix pages that arrive sideways

Rotation corrects a page's orientation — a quarter turn (90°), a half turn (180°, for upside-down pages), or a quarter turn the other way (270°). It's almost always a scan problem: a flatbed records the orientation the paper was fed in, and phone-scanner apps save the angle the phone was held at, so pages end up turned even though the content is upright. The fix updates a single display instruction in the PDF, leaving the actual content bit-for-bit identical.

The rotate tool applies one angle to the whole file, which is perfect when a bad scan run turned everything the same way. For a mixed document where only some pages are wrong, use the split-rotate-merge routine: split the file, rotate just the offending pages, and merge everything back in order.

Images to PDF: from a folder of photos to one document

Sometimes the document doesn't start as a PDF at all — it's a folder of photos or scans. JPG to PDF places each image on its own A4 page, centered and scaled to fit without distortion, in the order you arrange the thumbnails. It's the natural first step for receipts headed to an accountant, an ID copy a portal demands as PDF, or site-survey photos bundled into a report. The original JPEG bytes are embedded as-is, so there's no re-compression and the pages look exactly like your photos.

Once your images are a PDF, the rest of the toolkit applies: merge it with other documents, compress it to fit an upload limit, or rotate a page that came out sideways. If your images are PNG or WebP rather than JPG, convert them to JPG first, then bring them in.

The five operations at a glance

A real workflow: receipts to a single, emailable PDF

Here's how the tools chain together on an everyday task — turning a month of phone-photographed receipts into one tidy file to send to an accountant.

• Start with the photos. Drop your JPG receipts into JPG to PDF and arrange them in date order; you now have a single PDF, one receipt per page.

• Add anything already in PDF. If a supplier emailed a PDF invoice, use Merge PDF to append it to the document you just built.

• Fix any sideways pages. If a receipt was shot in landscape, split the file, rotate that page, and merge it back — or, if they're all turned the same way, rotate the whole file in one pass.

• Compress last. Run the finished document through Compress PDF so it slips under the mailbox limit. If it's still heavy because the photos are high-resolution, downsample the images.

The order matters: assemble and convert first, fix orientation next, and compress at the very end so the final file is as small as it can be. The same pattern — build, tidy, shrink — works for reports, application packets, and legal bundles.

A note on privacy

PDFs are personal. Contracts, invoices, ID copies and medical forms are exactly the kind of documents you don't want lingering on someone else's server. Every tool here is built to leave nothing behind: files travel over an encrypted TLS connection, are processed in memory rather than written to disk, and are auto-deleted the moment the result is returned. There's no account to create, no email to hand over, and no watermark stamped onto your document. The processing runs in a Rust service for speed, and a content-hash cache makes repeat operations on the same file instant without ever exposing it to anyone else.