How Safe Is a Password-Protected PDF? It Depends

You’re sending a document with something sensitive in it, so you set a password. The file now asks for one before it opens, and that feels like the job is done.

Mostly it is. Password-protecting a PDF is a sensible thing to do and I’d encourage it. But it’s worth understanding what you’ve actually built, because the protection has two weak points and neither is where people expect.

Two passwords, two very different jobs

First, a distinction that trips up almost everyone. PDF supports two kinds of password, and they are not equally serious.

The owner password sets permissions — no printing, no copying text, no editing. The document opens for anyone, and the restrictions are enforced only because reader software chooses to respect them. It is a request, not a lock. Free tools strip these in seconds, and that isn’t a hack; there’s nothing to break.

The user password is the real one. Without it, the file’s contents are genuinely encrypted and the document cannot be opened at all.

If you set a password and were shown options about printing and copying, check which one you set. People regularly believe a document is protected when all they’ve done is politely ask readers not to print it.

What brute force actually looks like

Assume you set a proper user password. An attacker with the file can try passwords one after another until something works — that’s brute force, and the only question is how long it takes.

The answer depends on two things: how long and varied your password is, and which version of PDF encryption was used.

Modern PDFs use AES-256, which is strong. Older files use RC4 with much shorter keys, and RC4 has been considered obsolete for years. If your PDF software still produces older-format encryption, that’s worth changing in its settings.

But encryption strength usually isn’t the limiting factor. The password is. A cracking tool doesn’t grind through every possible combination — it starts with dictionaries of common passwords, known leaked passwords, names, dates, and predictable substitutions. Invoice2026! looks complicated and falls in the first wave, because the pattern of a capitalised word plus a year plus punctuation is precisely what those tools try first.

Length beats cleverness, reliably. Four unrelated words are far harder to attack than eight characters of punctuation soup, and considerably easier to type over the phone.

The attack that skips the password entirely

Here’s the part most people have never heard, and it reframes the whole topic.

In 2019, researchers from Ruhr University Bochum and Münster University of Applied Sciences published PDFex, presented at Black Hat Europe. They showed that an encrypted PDF’s contents could be extracted without knowing the password at all.

The reason is structural: PDF encryption doesn’t encrypt the whole file. Parts of the structure stay readable so that software can work out how to handle the document. That leaves room for an attacker to modify the file — adding, for instance, an instruction to send decrypted content to a remote server the moment the legitimate recipient opens it and types the correct password.

The victim does everything right. They have the password, they open their own document, and the content leaves through a channel they never saw.

The researchers tested 27 PDF viewers. 23 were vulnerable to the direct exfiltration variant, and all of them to a second technique exploiting how the encryption mode handles modified data.

Vendors patched this, and current software is in much better shape. But the lesson survives the fixes: encryption protects the content of a file, not the integrity of the file. A stranger who intercepts your encrypted PDF can’t read it — but they may be able to alter it in ways that matter.

Brute force attacks against password-protected PDF files

What to actually do

Use a long passphrase, not a clever password. Four or five unrelated words. Length is what defeats brute force, and nothing else comes close.

Send the password separately. Emailing a protected PDF and its password in the same thread is the single most common mistake in this whole area. Anyone reading the mailbox gets both. Send the file by email and the password by phone or a messaging app.

Check you’re using modern encryption. If your software offers a compatibility choice, pick the option based on AES-256 rather than anything labelled RC4 or 40/128-bit.

Don’t confuse permissions with protection. If the requirement is “nobody who shouldn’t see this can open it”, you need a user password. Print and copy restrictions are for cooperative software, not adversaries.

Keep your reader updated. The PDFex fixes arrived as updates, and updates are how the next round will arrive too.

Think about whether encryption is the right tool. For a document going to one person once, a password is fine. For something several people need over months, a system with proper access control will serve you better than a password that gets forwarded, written down and eventually pasted into a group chat.

How long would it actually take?

People want a number here, and the honest answer is that the number depends almost entirely on you rather than on the encryption.

What determines it is how many guesses an attacker has to make. A password drawn from a predictable pattern — a dictionary word, a name, a year, a keyboard run — sits in a search space small enough that modern hardware exhausts it quickly. Add a genuinely unpredictable word and the space multiplies. Add another and it multiplies again. This is why length wins so decisively: each extra element doesn’t add to the difficulty, it multiplies it.

Two practical consequences follow.

First, predictability costs you more than length. A twelve-character password built from a familiar pattern can be weaker than four random words, despite looking more complicated. Cracking tools are built around exactly the patterns humans reach for when told to include a capital, a number and a symbol.

Second, reuse undoes everything. If the same passphrase protects every document you send, one leak — a forwarded email, a colleague’s compromised mailbox — exposes the whole set. The attacker doesn’t need to guess anything.

The practical target isn’t “uncrackable”. It’s “not worth the effort compared with the alternatives”, and four unrelated words clears that bar comfortably for the documents most of us send.

A note on where you encrypt

There’s an irony worth naming. Search for a way to password-protect a PDF and the top results are websites that will do it for you — which means uploading the document you consider sensitive enough to encrypt to a stranger’s server, unencrypted, so they can encrypt it for you.

The service then holds the plaintext, and quite possibly the password too. If the document mattered enough to protect, it mattered enough not to hand over first.

This is one of the operations PDF Manipulator handles locally: the file is encrypted on your machine and never uploaded. It doesn’t make a weak password strong, and it can’t help if you email the password alongside the file — but it does mean the sensitive version of the document never existed anywhere except your own computer.

The short version

A password-protected PDF is worth doing and easy to get subtly wrong. Make sure you set the password that actually encrypts rather than the one that just asks nicely, use a long passphrase instead of a clever short one, and never send the password down the same channel as the file. And remember what encryption promises: it protects what’s inside the document, not the document itself.

Encrypt your PDFs without uploading them first — PDF Manipulator is free →