Skip to main content
Orilami

PDF Leak Tracer: before and after

An illustration with sample files; your own files show their own results.

Before3 copies found online

  • Copy A: who leaked it?

  • Copy B: who leaked it?

  • Copy C: who leaked it?

AfterEach one traced to a recipient

  • Traced to Ana Costa

  • Traced to Ben Okafor

  • Traced to Chen Wei

PDF Leak Tracer: Watermark Each Recipient’s Copy to Trace a Leak

PDF Leak Tracer fingerprints a document uniquely for each recipient, so the source of a shared or leaked copy can be identified. Every layer is written and read in your browser; the file never leaves your device.

Layers written per document

  • Catalog: /Orilami /Payload — the 81-byte RS(81,49) codeword.
  • XMP: <orilami:payload> — the same codeword, base64.
  • Info dict: an Author hex tripwire plus zero-width bits (Hamming(15,11) SECDED) hidden in Subject.

How it works

The payload is 4 B magic, 1 B version, a 16 B HMAC-SHA256 fingerprint, a 24 B recipient id and a 4 B timestamp, protected by shortened Reed-Solomon(255,223). Up to 16 corrupt bytes per surviving layer are recoverable, so the recipient can still be identified after cropping, re-compression or a metadata purge that leaves any one layer intact.

Frequently asked questions

How do you identify who leaked a document?
Each recipient's copy is marked differently, so the copy itself names its source. When a file turns up where it should not be, you compare its marks with the batch record and the originating recipient is identified.
What does layered steganography mean here?
The identifier is carried in several independent places in the file. A leaker who defeats one layer leaves the others intact, so attribution survives ordinary attempts to clean the document.
Will the marks survive printing or a screenshot?
Survival depends on the layer. Marks in visible page geometry tend to survive print and capture; marks in the file's structure do not survive a screenshot. Using several layers together keeps attribution possible across different kinds of copying.
Is this the same as DRM?
No. DRM tries to stop access, usually with a dedicated reader, licence server and account. This marks each copy instead: anyone can open the PDF normally, but every copy is traceable. It is attribution, not prevention.
Does the recipient list leave my browser?
No. The names and the master PDF are processed locally and the marked copies are built in your browser.
Do I need the same secret to decode?
Yes. The HMAC secret verifies authenticity and must match the one used to encode the document.
How much tampering can it survive?
Up to 16 corrupt bytes per surviving layer are recoverable (shortened Reed-Solomon(255,223)), so cropping and re-compression don't necessarily destroy the fingerprint.
More questions about Orilami

PDF Leak Tracer at a glance

Key facts about PDF Leak Tracer: where it runs, what it costs, what it accepts and produces, and its limits.
Where it runsEntirely in your browser. Files are never uploaded.
Cost250 tokens per batch (about $2.50 at the base rate)
AcceptsA master PDF and a recipient list
ProducesOne uniquely marked PDF per recipient, as a ZIP.
BatchOne at a time.
Account neededOnly to spend tokens. Browsing and previewing are free.
What it does not doThe mark identifies which copy leaked. It does not prevent the file from being opened, copied or shared.