Editing PDFs Without Uploading Them Anywhere

What "upload" actually means for a document

The word makes it sound like a transfer. It is a copy. Your file now exists on a server, and what happens next is governed by a privacy policy you did not read and cannot audit.

For a holiday itinerary, that is a reasonable trade for a free service. For a signed contract, a passport scan, a medical letter, a payslip or anything under NDA, it is a decision worth making deliberately — and most people do not realise they are making one.

The practical questions nobody can answer for you: how long is it retained, who can reach it, is it encrypted at rest, is it used to train anything, and what happens in a breach. "Files are deleted after an hour" is a claim, not a mechanism.

How the browser-side version works

Modern browsers can run compiled code at close to native speed through WebAssembly. The PDF libraries that used to live on a server now run inside the tab.

The sequence is: you pick a file, the browser reads it into memory, WebAssembly code manipulates the bytes, and the result is handed back as a download. At no point does the document leave the machine. The page was downloaded once; the file never travels.

This is why such tools work offline after the first load, and why there is no upload progress bar — there is nothing to upload. On this site every PDF and image tool works this way, and every library — and every webfont — is self-hosted rather than pulled from a CDN, so no tool here makes a third-party request while it works. Open the Network tab and check: apart from the analytics tag in the page head, every row is this domain, and blocking that tag changes nothing about what the tools do.

Verify it yourself in about twenty seconds

This is worth doing once on any tool you plan to trust with something sensitive, including this one. A privacy claim you can check in twenty seconds is worth more than a privacy policy you cannot.

For the same reason, you can pull the network cable and the tools keep working. That is not a feature anyone designed; it is a consequence of there being no server involved.

1. Open the tool you want to check — Merge PDF will do.
2. Press F12 to open developer tools and choose the Network tab.
3. Clear the list, then add your files and run the operation.
4. Watch what appears. You should see no request carrying your file — no POST with a large payload, nothing to an upload endpoint.

What can be done locally, and what genuinely cannot

Nearly everything people want from a PDF tool works in the browser: merging, splitting, reordering and rotating pages, deleting pages, compressing, adding a watermark, converting images to PDF and PDF pages to images, filling forms, and extracting text from a PDF that contains text.

Two things do not, and it is worth being straight about them.

Setting an open password. Password Protect PDF cannot do it, and the page says so rather than pretending. The one browser library that claimed to ships builds that throw at runtime and export nothing. Use Preview on a Mac, Word, or LibreOffice — all of which do it locally on your own machine anyway, which was the point.

Reading a scanned document. PDF to Text Extractor extracts text that is already text. A scan is a picture of text, and pulling words out of a picture needs optical character recognition. There are OCR tools here that also run in the browser, but they work on images rather than reaching inside a PDF.

The trade you are actually making

Browser-side is not free of downsides, and pretending otherwise would be the same overselling this article is complaining about.

It uses your processor, so a 300-page merge is slower on an old laptop than it would be on a server. It uses memory, so a very large file can exhaust a tab. It cannot do anything needing a service you do not have — a translation model, a huge OCR model, a font library you lack.

What you get for that is: no upload wait, no server-imposed size cap, no account, no retention question, and no queue behind other users. For most documents most of the time, that is the better side of the trade — and for anything confidential it is not really a trade at all.