How PDFsansar works

The short version

When you open a tool, your browser downloads a page and a few hundred kilobytes of JavaScript from us. After that, the network is done. Everything else runs on your device:

• You pick a file. The browser hands the page a reference to it. Nothing has moved yet.
• The page reads the bytes into memory, on your machine.
• It does the work there: rendering pages, reordering them, rotating, writing a new PDF.
• It builds the result as a file in memory and asks your browser to save it.

At no point does a copy go over the network, because no part of that list involves us.

What does the work

Three open-source libraries, all served from this domain rather than from someone else's CDN.

pdf.js

Mozilla's PDF engine, the same one built into Firefox. It reads your PDF and draws pages onto a canvas, which is what makes thumbnails and image export possible.

pdf-lib

Handles the document surgery: merging, splitting, reordering, rotating. It edits the PDF structure directly rather than re-printing pages, so a rotated page keeps its original text and quality instead of turning into a picture of itself.

JSZip

Bundles multi-file results, so converting a 40-page PDF gives you one zip instead of 40 separate downloads.

Why there is no upload

This site is static. It is a folder of HTML, JavaScript and images sitting on a content network. There is no application server, no database, no storage bucket, and no code of ours running anywhere except inside your tab.

That is a design decision, not a limitation we are apologising for. A server that never receives your file cannot lose it in a breach, cannot keep it after you leave, and cannot be asked to produce it. The full story is on the about page.

Verify it yourself

Do not take this on faith. Either of these takes seconds:

• The offline test: open a tool, wait for it to load, then switch off your wi-fi or turn on airplane mode. Drop in a PDF. It works exactly as before.
• The network test: press F12, open the Network tab, then use a tool. You will see the page and its scripts load. You will not see a request carrying your file, because there is none.

The offline test is the one we would trust if we were you. Software cannot send your file anywhere if there is no connection, and these tools do not care that there isn’t one. We have since made that official; see below.

What this costs you

Doing the work locally is not free of trade-offs, and we would rather say so than let you find out on a deadline.

• Speed depends on your device. A 300-page scan converts quickly on a laptop and slowly on a five-year-old phone. A server farm would be faster.
• Very large files can strain memory. The page holds the whole document while it works. If a tab runs out of room, close the others and try again.
• Page thumbnails and the redaction pass still render on the main thread, so a big job there can make the page feel stuck for a moment. Converting to JPG or PNG already runs on a background thread; the rest is being moved across.
• Some jobs genuinely need a server. Converting Office documents is the obvious one. We would rather not offer it than quietly start uploading your files. Reading text out of scans turned out not to: OCR PDF runs in your browser like everything else, with a few megabytes to download the first time.

What does leave your browser

Being precise matters here, so here is the complete list of network traffic this site can cause.

• The page itself: HTML, JavaScript, and the PDF engine, all from pdfsansar.com.
• Page-view analytics, if you agreed to it. It records which page was opened, never anything about a file.
• The contact form, but only if you fill it in and press send. That sends your message and reply address, and nothing else.
• Signing in, if you choose to. Pressing the sign-in button sends you to Google and back; nothing on this site contacts Google otherwise. After that, your settings and starred tools sync, never anything about a document.

That is the lot. Your documents are not on it, and never will be. The privacy policy spells out each item, and the security page covers what an attacker could and could not reach.

Installing it, and using it offline

Because the tools never needed a server, working offline was never a feature that had to be built. It was already true, and the offline test above is how people kept discovering it by accident. So it is now official: PDFsansar installs to your home screen and keeps working with no connection.

There is one honest limit. A tool has to have been opened once, with a connection, for its code to be saved on your device; after that it opens offline like anything installed. A tool you have never used will tell you it needs a connection rather than failing strangely. If you would rather not wait to find out, one tap in Account Settings gets every tool ready for offline use, a few megabytes, and they all work without internet from then on.

Installing changes nothing about how the tools work or what happens to your files. It is the same site in a window without an address bar, and the button is on Account Settings. On an iPhone, Safari does the installing rather than the page: Share, then Add to Home Screen.

An installed PDF editor that never uploads anything and works on a plane is a genuinely odd thing to be able to give away for nothing. It only works because the hard decision (no server) was made first.

Accounts, and what happens if we ever add a server

An optional account exists now. It syncs your starred tools, your language and your appearance choice between your devices, and that is the whole of it. No tool has been put behind a sign-in and none will be, there is no free-tier page limit, and you can use this site forever without an account. Signed out, the dashboard saves your favourites in your own browser exactly as it always did.

Signing in does not change where your documents are processed. The sign-in code is a few small endpoints that handle account records; they never see a file, a filename or even a page count, and no tool talks to them at all. The privacy policy lists in full what an account holds.

The harder thing is still ahead. A few things people ask for cannot be done in a browser at all. Word and Excel conversion is the obvious one. If we ever build them, they will be separate, clearly marked tools that say on their own page, before you pick a file, that the file is sent to our server and what happens to it there. They may require an account. What they will not do is change any tool that exists today: everything on the current list runs on your device, and will keep running on your device. Moving one of them onto a server would break the only promise this site has, and we are not going to do it quietly or at all.