privacy · client-side utility

100% Free | No Sign-Up Required

Remove PDF Metadata Offline — Zero‑Upload Scrubber

Zero uploads. Your file never touches our servers.

The PDF is read from disk into memory, rebuilt in this tab, and saved back to your device.

transport
0 bytes
storage
none
process
this browser
audit
view-source

input

local only

Drop a PDF here

…or click to browse. The file is read from disk into memory — never onto the wire.

Press Enter to open the file picker.
metadata targets status
  • /Titlepending
  • /Authorpending
  • /Subjectpending
  • /Creatorpending
  • /Producerpending
  • /CreationDatepending
  • /ModDatepending
  • custom keyspending
  • XMP packetpending
local session · pdf-scrub no upload
> pdf metadata scrubber — local session > runtime: your browser · transport: 0 bytes · storage: none > waiting for a PDF…
Fields cleared 0 7 /Info keys + custom keys + XMP
XMP packet — RDF/XML inside /Metadata
Size change — before → after
Bytes uploaded 0 always — there is no upload code

direct answer

To remove metadata from a PDF without uploading it, load the file into a tool that runs locally: your browser reads it with the File API, pdf-lib clears the /Info dictionary and XMP packet in memory, and the scrubbed document downloads as cleaned-[original-name].pdf. No server is involved at any point.

how it works

Three steps. No uploads.

It all happens between your disk and your downloads folder.

  1. Step 1

    Select your PDF

    Drop a file into the zone or click to browse. The browser reads it with the File API into an ArrayBuffer — no request is opened, so there is nothing for a server to receive.

  2. Step 2

    The scrubber analyzes it locally

    pdf-lib parses the document in memory, walks the /Info dictionary and the XMP packet, and clears Title, Author, Subject, Creator, Producer, CreationDate and ModDate.

  3. Step 3

    Download the cleaned PDF

    The file is rebuilt, verified by parsing it a second time, and saved as cleaned-[original-filename].pdf. Uploaded bytes: 0.

field notes

Why the "Delete in 1 Hour" Promise Is a Security Risk

Free online PDF cleaners all advertise the same reassurance: upload your document, and it will be deleted within the hour. That sentence describes a system in which your file is copied to somebody else's hardware, processed by software you cannot inspect, and removed only after a timer you cannot verify. The promise is the product. Everything that happens before the clock runs out is the risk.

An upload means your document exists in at least four places: the inbound buffer, the working directory the converter writes to, the cache or CDN edge that serves the result, and the backup snapshot that nightly jobs recreate. "Deleted in one hour" usually describes only the working copy. Snapshots expire on their own schedule — thirty days is common — and log lines such as 4.2 MB received at 14:07 UTC are rarely purged with the file. Where a service outsources conversion, the document travels a second time to a processor named in a terms-of-service appendix you did not read.

The deeper problem is that a deletion promise is an assertion, not a property. You cannot observe it from outside, and most companies cannot prove afterwards that a particular byte is gone. The threat models you actually care about — an exposed storage bucket, a curious contractor, a subpoena, a ransomware snapshot — all sit precisely in the window the promise does not cover. Scrubbing locally removes the window entirely: the file is read into an ArrayBuffer, rewritten in memory by JavaScript, and handed back as a download. There is no request to intercept because no request is ever made.

What Hidden Data Is Actually in a PDF

People treat a PDF as a picture of a page. It is closer to a database, and two structures in it describe the file rather than the content.

The /Info dictionary

The document trailer points to an /Info dictionary holding /Title, /Author, /Subject, /Keywords, /Creator, /Producer, /CreationDate and /ModDate. /Creator is the application that started the document — Word, LaTeX, InDesign, a scanner utility. /Producer is the library or distiller that finished it, often naming a company and an exact version number. The two dates carry time zones, so they can place a document on a specific afternoon in a specific office. Any custom keys a workflow added live here too, which is why this tool clears every entry left in the dictionary, not just the seven standard ones.

XMP metadata

Modern exporters also write an XMP packet: RDF/XML stored as a /Metadata stream attached to the document catalog. It mirrors the dictionary but can hold far more — software and version, export presets, colour profiles, resolution, camera or scanner model, edit history, sometimes "saved by". Because it is plain XML it is trivially greppable, and because the two copies can disagree, a cleaner that touches only /Info quietly leaves the second copy behind. This scrubber blanks the packet and leaves an empty one in place so the structure stays valid.

What else leaks

Subset font names, embedded attachments, annotation notes and earlier incremental revisions can all carry identifying traces. Read in this order, the fields a stranger sees first are Author, Creator and Producer — exactly the ones this tool removes first.

Who Needs This Tool

Lawyers and paralegals

Filings, discovery bundles and exhibits travel to opposing counsel daily. Metadata discloses drafting software, internal file paths and the paralegal who exported the bundle, and courts increasingly treat an inadvertent metadata disclosure as a competence problem rather than a curiosity. Scrub each exhibit before it leaves the office.

Journalists and their sources

A leaked PDF whose /Author or /CreationDate narrows the document to four people who touched it on one afternoon can endanger the chain behind it. Cleaning the file before publication protects the source, not just the text — and doing it locally means the unpublished draft never sits on a third-party server while you work.

HR, recruiting and consulting

Offer letters, performance reviews and internal memos carry the HRIS or template name in /Creator and a manager's name in /Author. Sending them out intact leaks org structure to the recipient; sending a signed contract out with a stale /ModDate invites arguments about when it was last edited. Researchers and anyone publishing downloadable forms get the same benefit: the document says what it says, and nothing about how it was made.

Related Tools and Reading

  • How to Remove Metadata from a PDF — the full field list, every manual removal method, and how to verify a file is actually clean.
  • Privacy tools category — every browser-side utility on Daily Utility Lab that keeps data on your device.
  • Privacy policy — what we log (nothing), what we store (three small local preferences, not your files), and how the zero-upload claim is enforced.
  • Guides — longer write-ups on the numbers and the fine print behind the tools.

faq

Frequently Asked Questions

What the log proves, in plain sentences.

Is my PDF uploaded to a server?

No. It runs via local JavaScript. Your browser reads the file with the File API, pdf-lib parses and rewrites it in memory, and the cleaned file comes back as a download. There is no upload endpoint in the code — open DevTools, switch to the Network tab, and watch it stay empty while a file is scrubbed.

What metadata is removed?

Author, Creation Date, Mod Date and Creator tool — plus Title, Subject, Producer, any custom keys remaining in the /Info dictionary (Keywords, Trapped, whatever a workflow added), and the XMP metadata packet that mirrors them in RDF/XML. The terminal log names each entry as it is cleared.

Does scrubbing metadata change my PDF?

No. Pages, text, images, links and form fields stay exactly as they were; only document-level metadata is cleared and the file is re-serialized locally. One exception is worth knowing: removing metadata invalidates a digital signature, so re-sign a document after scrubbing it.