Safe handling of confidential PDF files
Local processing, hidden content, why cropping is not redaction and passwords done properly
How to keep a confidential PDF private from the moment you open it to the moment you delete it: local processing, hidden content and metadata, why cropping and watermarks are not protection, and passwords done properly.
- By HelpMeJPG editorial team
- Published
- 9 min read
On this page
Quick answer: Keep a confidential PDF on your own device while you work on it, find out what the file actually contains before it goes anywhere – metadata, comments, hidden layers, content you cropped away – and understand that cropping and watermarks remove and protect nothing. If the content must stay private in transit, encrypt it with a strong open password and send the password by a different route; then get rid of the copies you no longer need.
Where the file goes when you "process" it
Most online PDF tools work by uploading the document to a server, processing it there and sending a result back. For a contract, a medical report, a bank statement or an identity document that means a copy of the file has left your control: it sits on someone else's disk for as long as their retention policy says, and you have only their word about what happens next.
The tools on this site take the other route. The file is read into memory inside your browser tab, the work is done by code running on your own machine, and the result is offered to you as a download. There is no upload endpoint. You can confirm this in the browser's developer tools: the network panel shows no request carrying the document. Passwords typed into the protect and unlock tools are used in memory and discarded.
Local processing removes the biggest risk, not every risk: the browser still writes the result to your downloads folder, a shared computer still has other users, and the file still contains whatever it contained before you started.
What a PDF carries that you cannot see
A PDF is far more than its visible pages. Before sending one, assume it includes several of the following and check:
- Document information. Title, author, subject, keywords, dates, and the creating and producing applications. The author is frequently a login name or someone other than the apparent sender.
- XMP metadata. An XML packet holding the same fields again, plus document identifiers and sometimes an editing history. The two often disagree, and a tool that cleans one may leave the other.
- Annotations. Comments, sticky notes, highlights, stamps and review threads, some of them collapsed or hidden by default.
- Form fields. Values persist even when a field is hidden or off the page, and merged forms can carry values from another copy.
- Attachments. Embedded files – spreadsheets, earlier versions, correspondence – that travel invisibly with the document.
- Optional content (layers). Content that is switched off for display but fully present in the file.
- Earlier versions. Many applications save edits as incremental updates appended to the file, leaving superseded objects in place; a document edited to remove a paragraph may still contain it.
- JavaScript. Scripts attached to the document or to form actions.
- Content outside the visible area. Anything beyond the crop box, discussed below.
The PDF metadata viewer shows the document information and XMP side by side and flags encryption, forms, annotations, structure tags and embedded JavaScript – which it detects but never runs. The PDF text extractor returns the text layer page by page; if text you believed was blacked out, cropped off or deleted appears in the extraction, it is still in the file. The PDF page counter confirms the page count, which catches appended pages and a second copy of a form merged by mistake. Attachments and layers show up in a full viewer's panels.
Cropping is not redaction
Cropping a PDF page sets its crop box: a rectangle that tells viewers and printers which part of the page to show. The content outside that rectangle is not deleted. It is still in the content stream, still in the images, still in the text layer, and anyone with a PDF editor can enlarge the crop box and see it all. The Crop PDF tool says so in its own notes; even its option to trim the media box changes only the page's boundaries, not the content drawn on it.
The same is true of the other visual "redactions" people reach for. A black rectangle drawn over a name covers the name the way a sticky note covers a line in a letter: lift it and the text is there, and a text extractor returns the name without even lifting it. Changing text to white, shrinking it to nothing or hiding it behind an image leaves it equally intact.
Real redaction removes the underlying content – the text operators, the pixels of the image region, the annotations that reference it – and then writes a fresh file so that no earlier version survives as an incremental update. That is a job for a dedicated redaction feature; no tool on this site performs it. Two operations here do remove content rather than hide it. Delete PDF Pages and the other page tools copy the pages you keep into a fresh document, so content that belonged only to a removed page is not carried over. And the compressor's option to remove metadata deletes the document information and XMP packet rather than blanking them. Neither is a substitute for redacting part of a page.
Watermarks are not security
A watermark is a visible mark – "CONFIDENTIAL", "DRAFT", a recipient's name – drawn on top of the page content. It has real uses: it tells a reader how to treat the document, discourages casual forwarding, and a watermark naming each recipient makes a leaked copy traceable. The Add PDF Watermark tool draws text or an image over the pages without rasterising anything beneath.
What a watermark does not do is prevent anything. It is an ordinary graphics object that a PDF editor can delete, the text under it remains selectable and extractable, and the pages can be printed, photographed and copied exactly as before. Treat it as labelling, never as access control. If a document must not be read by the wrong person, that is a job for encryption; if it must not be altered without detection, that is a job for a digital signature.
Passwords and encryption, done properly
Encrypting a PDF with an open password is the one measure in this list that actually prevents access: the content is encrypted with AES-256 under a key derived from the password, and without the password it cannot be read. The Protect PDF tool applies the PDF standard security handler in the browser and encrypts both the streams and the strings – page content, images, metadata, comments, field values – so that nothing readable is left in the clear.
Getting value from it depends on a few habits:
- Set an open password, not just an owner password. An owner-only file opens for anyone; the permission restrictions it carries are honoured by compliant readers and ignored by the rest. How PDF passwords and encryption work explains the difference.
- Use a passphrase. Several unrelated words or a long random string; the cipher is never the weak point, the password always is.
- Send the password by a different channel. A phone call, a text message or a messaging app – never the same email as the attachment.
- Do not reuse passwords across documents or recipients. One disclosed password should open one file.
- Keep a record you can find. A strong password nobody remembers destroys the document as surely as deleting it; there is no recovery.
Encryption covers the file at rest and in transit; once the recipient has opened it, they can do anything a person with a printout could do. Decide whether you trust them with it before you send it, not after.
Before you send: a checklist
- Open the metadata viewer and read the title, author, subject and dates in both metadata sources; note any flags for forms, annotations or JavaScript you did not expect.
- Run the text extractor and skim the output for names, figures or passages that should not be there – including anything you covered, cropped or thought you had deleted.
- Check the page count and flick through every page, looking for a stray appendix or a second copy of a form.
- If metadata should not travel, run the file through Compress PDF on the Lossless profile with the remove-metadata option; the pages are untouched and both metadata sources are gone.
- If part of a page must be removed, use a real redaction tool and save as a new file; then run the extractor again to prove the content has gone.
- Look at the file name: one called after a client, a case or a diagnosis says a great deal before it is opened.
- Protect with an open password and share it separately, or send a link to storage with access control instead of an attachment.
- Open the final file one more time, exactly as the recipient will.
Preparing PDFs for email covers the size and formatting side of the same send.
Disposing of copies
A confidential document multiplies as you work on it: the original, the protected version, the browser's automatic download, the attachment in an email draft, the copy a cloud folder synchronised. When the work is done, remove the ones that no longer need to exist and empty the recycle bin afterwards.
Be realistic about what deletion achieves. On modern solid-state storage, "securely overwriting" a single file is not reliable, because the drive decides where data physically lives; the dependable protection is full-disk encryption, which makes everything on the drive unreadable without the login. On a shared computer, assume the next user can find anything you downloaded unless you delete it and clear the browser's download history. Keep one canonical copy where access is controlled, and let that be the only one.
Handling confidential files you receive
A PDF from an unknown sender may carry JavaScript or links intended to do harm; open such files in a viewer that does not run scripts, or inspect them with the metadata viewer first, which flags scripts without executing them. A protected file you have been given the password for can be unlocked with Unlock PDF, but the unlocked copy is now an unprotected document on your disk – protect it again if you keep it.
Frequently asked questions
Is it safe to use online PDF tools with confidential files?
It depends entirely on where the processing happens. A tool that uploads the file hands a copy to a server you do not control; a tool that runs in the browser never sends the file anywhere, which you can verify in the network panel of the browser's developer tools. The tools on this site work the second way, and the remaining risks – metadata, hidden content, copies in your downloads folder – are yours to manage.
Does drawing a black box over text remove it?
No. The box is a graphic drawn on top; the text beneath remains in the file, is returned by a text extractor and can be exposed by anyone who moves or deletes the box. Cropping behaves the same way. Only a dedicated redaction tool that deletes the underlying content and re-saves the file as a fresh document removes it.
Should I send the password in the same email as the PDF?
No. Anyone who can read that email – an intermediary, a forwarded recipient, someone with access to the mailbox later – has the file and its key together. Send the password by a different channel such as a phone call or a message, use a passphrase rather than a word, and do not reuse it for other documents.
Sources and further reading
About the author
HelpMeJPG editorial teamEditorial team
The people who build and maintain the HelpMeJPG tools. Guides are written from the tools' own behaviour and the standards they implement, and are revised whenever a tool changes.