How to Scrub, Redact, Shrink, and Lock Down a PDF Before You Send It to Someone Outside Your Company
Published October 20, 2026
A practical order of operations for preparing a sensitive PDF before it leaves your hands: clean out hidden metadata, black out anything that shouldn't be visible, compress the file, and only then add a password — because each of those steps can quietly undo the one before it if you do them out of order.
You're about to send a PDF to someone outside your organization — a vendor, a client, opposing counsel, a journalist, a regulator. The visible content of the document is fine to share. But PDFs carry more than what's on the page: there's an author field with someone's real name, a "created with" string naming the exact software and version used, sometimes GPS coordinates if the file started life as a scanned photo, and occasionally a paragraph or an image buried in there that genuinely shouldn't leave the building. Getting this file ready to send safely is not one step, it's four, and the order you do them in actually matters — get it wrong and you can undo earlier work without realizing it.
Here's the order that holds up, and the reasoning behind each position.
Step 1: Redact first, while the file is still fully editable
Start with PDF Redaction Tool if there's anything in the document that needs to be permanently removed — a client's account number buried in an exhibit, a name that needs to stay confidential, a diagram with a logo you don't have rights to distribute, a page of internal notes that got stapled into the export by mistake. Redaction here means actually deleting the underlying content in that region, not just drawing a black box over it — a black rectangle placed on top of text in a PDF viewer is cosmetic; the text is still there underneath, selectable and copyable, and plenty of leaked documents over the years have been un-redacted by someone simply selecting the "blacked out" text and pasting it elsewhere. A real redaction tool removes the content itself, so there's nothing left to recover.
Redaction goes first for a structural reason: it's the step most likely to change what's actually in the file — removing text, flattening an image, sometimes restructuring a page. Every step after this should be operating on the final content of the document, not a draft of it. If you redacted last, you'd have already compressed and possibly re-encoded content you were about to delete anyway, which is wasted work at best and at worst leaves compressed artifact traces of the original in a way that's harder to reason about than just redacting the clean original first.
Step 2: Strip the metadata next
Once the visible and hidden content of the page is final, clean out the metadata — the information about the file that isn't part of any page but travels with it anyway. Open PDF Metadata Editor and look at what's actually in there. It's often more revealing than people expect:
- Author — frequently a real employee's name or username, pulled automatically from whatever software created or last saved the file.
- Producer / Creator — the exact application and version that generated the PDF, which can expose internal tooling or outdated, known-vulnerable software versions you'd rather not advertise.
- Creation and modification timestamps — not usually sensitive on their own, but combined with other metadata they can reveal an internal timeline you didn't intend to share (e.g., a document stamped as "created" the same day as a confidential meeting).
- GPS coordinates — this one catches people off guard. If a PDF page originated as a photo (a scanned document, a photographed whiteboard, a picture of a receipt) and that image retained its EXIF location data when it was placed into the PDF, the exact latitude and longitude of wherever the photo was taken can be sitting in the file's metadata, readable by anyone who knows to look.
Clear or overwrite the fields that don't need to be there, and verify the result before moving on. This step comes after redaction for the same reason redaction came first: you want to be scrubbing the metadata of the document you're actually going to send, not a draft that's about to change.
It comes before compression, though, for a more practical reason: it's just easier to confirm metadata is clean on a normal-sized working file than to second-guess yourself after a compression pass has touched the file's internal structure. There's no hard technical requirement here, but doing it in this order means each step is checking its own box cleanly rather than overlapping with the next one.
Step 3: Compress once the content and metadata are both final
Now, and only now, run the file through PDF Compress. Compression works by re-encoding images and streamlining the internal structure of the PDF — which is exactly why it needs to happen after redaction and metadata cleanup rather than before. If you compressed first and redacted second, the redaction tool would be operating on an already-recompressed file, and you'd typically want to verify the compressed output doesn't somehow retain remnants of the pre-redaction version — it's simpler to just redact the original, then compress the final result once.
For a typical scanned or image-heavy PDF — a signed contract with a few embedded images, a report with charts and screenshots — expect the file to come down substantially from a redaction-and-metadata-pass version, often from several megabytes down to well under a megabyte or two, depending heavily on how many images are in it and how they were originally encoded. A text-only document with no images won't shrink nearly as dramatically, since there's not much redundant image data to squeeze out — compression helps most on documents that are visually heavy, not word-processor documents that are mostly plain text.
Step 4: Password-protect last, always last
Finally, if the document needs access control, add a password with PDF Add Password as the very last step in the pipeline — not because of some arbitrary convention, but because of a hard technical constraint: an encrypted PDF generally can't be read, parsed, redacted, have its metadata edited, or be compressed by most tools without first supplying the password to unlock it. If you encrypt the file and then try to run it through any of the earlier steps, you'll either hit an error or be forced to decrypt it again anyway, defeating the point of locking it in the first place. Password protection is the one step in this pipeline that's genuinely order-dependent in a strict sense: everything else happens before it, full stop.
When you set the password, think about who's receiving the file and how you'll actually get them the password — sending the PDF and the password in the same email thread defeats most of the purpose, since anyone who intercepts one typically intercepts the other. A phone call, a text message, or a separate channel entirely is a meaningfully stronger practice if the contents are genuinely sensitive.
Why not just do it in whatever order feels natural?
It's tempting to compress first because file size is usually the visible, annoying problem, and tackle redaction and metadata as an afterthought once the file is already "small." The trouble is that redaction changes the content, metadata cleanup only matters once the content is final, and encryption locks the door on all of it. Doing the steps in reverse order doesn't just risk extra re-work — it risks ending up with a password-protected file that still has the real author's name in its properties, or a compressed file that has a redacted-looking black box sitting on top of text that's still fully recoverable underneath.
A note on what redaction tools can and can't catch
Redaction tools do a thorough job on the region you select — the content in that box is genuinely gone, not hidden. What they can't do is read your mind about what else in the document might be sensitive. A redaction pass is only as good as your own review of the document beforehand: skim every page first, note anything that needs to go, and redact each instance individually. If the same sensitive string appears on five different pages, it generally needs to be redacted five separate times, not found-and-replaced in one pass — treat it like proofreading, not a global search-and-destroy operation.
The short version
- Redact anything that can't leave the building — PDF Redaction Tool.
- Scrub the metadata — author, software version, GPS data — with PDF Metadata Editor.
- Compress the now-final file with PDF Compress.
- Password-protect last, since an encrypted file can't easily be processed further — PDF Add Password.
All four of these tools run entirely in your browser — the PDF never gets uploaded anywhere, which is precisely the point when the whole exercise is about controlling who sees a sensitive document. If the file never has to leave your machine to get cleaned up, that's one fewer place it could end up somewhere you didn't intend.