How it works

Nothing leaves the tab

Most online PDF tools are upload tools with a friendly face: your file travels to somebody's server, waits in a queue, and gets processed far away from you.

pdfviz works the other way round. The page loads a PDF reader and a PDF writer as ordinary code, then runs them on the file you picked - from compression to merging.

Your file is passed to code already running in your browser. There is no upload, no queue, and no server that ever sees it.

You do not need an account, because there is nobody to recognise you. Turn off your network after the page loads and the tool still works - the simplest proof that nothing is being sent anywhere.

Every result is checked

Processing a file can quietly wreck it. Compression can drop characters and merging can scramble page order - and a tool that never looks at its own output hands you the damage with a smile.

So pdfviz does not trust its own output.

It opens each result again with the same reader a viewer would use, counts the pages, and counts the characters it can select. A result that does not match the source is thrown away, not shipped.

Your original is never at risk. The tools only ever read it, and every edit happens on a copy.

If anything fails, your file is exactly where you left it.

Built on open source you can inspect

The two halves of every pdfviz operation come from public open-source projects, not from a black box.

The reading half uses pdf.js, the reader Mozilla ships with Firefox. The writing half uses pdf-lib, a widely used library for building PDF files. Both arrive with the page as plain code anyone can inspect.

That split is deliberate. A tool that grades its own homework proves nothing, so the checker is never the writer.

It gets the finished bytes, opens them the way a viewer would, and reports what it finds. The reader confirms what actually reached the page - that is what makes the result worth anything.

Safe by architecture, not by promise

Privacy policies promise. Architecture guarantees. pdfviz is safe for sensitive documents - contracts, medical records, ID scans, financial statements.

There is no path your file could take to leave the device.

No upload endpoint receives it. No temporary storage holds it. No log records its name. Our privacy notes describe the full picture.

This also means the site works wherever you are: on office networks that block uploads, on slow connections, on a laptop in offline mode. The tools are already in the tab.

What we cannot promise

A size target is not always reachable. If the smallest surviving file is still over the target you asked for, the tool says so and shows you the number it did reach. It does not ship a broken file to make a number look good.

Very large files can also hit the limits of the tab itself. Browsers cap how much memory one page may use, and a thousand-page scan with full-page images may exceed it. When that happens you get a clear error, not a half-written download.

Questions people ask

Are my files uploaded anywhere?

No. The file you pick is opened by code running in your browser tab and never leaves your device. There is no upload step, no queue, and no server copy to delete afterwards.

Can I use pdfviz offline?

Yes, after the page loads. The reader and writer libraries arrive with the page as JavaScript and WebAssembly. Once they are in the tab, the network is idle, so disconnecting changes nothing.

Why do two different libraries handle one job?

So the result gets an independent second opinion. pdf.js reads, pdf-lib writes, and pdf.js reads again. A single library checking its own work would confirm its own intent instead of what actually reached the page.

What happens when a result fails the check?

You get an honest message instead of a download button. The failed file is discarded and your original is untouched, because the tools only ever read it.

Is it safe for sensitive documents?

Yes. Contracts, medical records, ID scans, and financial statements never leave your machine, since there is no upload path at all. Safety comes from the architecture, not from a policy promise. You can redact or remove metadata on top, still fully offline.

Do I need an account?

No. There is nobody to recognise and nothing stored about you, so there is nothing to sign into.