Confidential IP Watermarking
A content-sealing layer that lets you prove the origin of an image long after it has left your hands.
The problem
/ 01Once a proprietary image, chart, or render leaves your environment, you lose sight of it. Copies get forwarded, screenshotted, and re-encoded — and there is often no way to prove which artifact a given file actually came from.
What it does
/ 03Embeds a persistent payload within image pixels.
Extracts the embedded payload to verify provenance.
Optional key binds extraction to a known secret.
Who it’s for
/ 03- IP owners distributing renders, product shots, or brand assets
- Compliance teams that need provenance evidence for controlled artifacts
- Anyone shipping images outside a trusted boundary who wants a recoverable fingerprint
In context
/ 01Pairs naturally with Media Provenance (attribution strips for generated content) and the Air-Gap Courier (payloads sealed into image carriers).
Spec & limits
/ limits| Interface | Input limits | Output artifact | Failure mode | Non-claim |
|---|---|---|---|---|
Studio ephemeral /demo; pilot durable /v1 on your infrastructure. | TBD — verified in pilot acceptance | TBD — verified in pilot acceptance | TBD — verified in pilot acceptance | TBD — verified in pilot acceptance |
Related guide
/ guideSee the forensic watermarking guide for a deeper problem walkthrough and verification notes — evaluation notes for the live demo and for a private pilot. Also: try the live experiment and read the product overview.
Frequently asked questions
/ 0501 How is forensic watermarking different from a visible confidential stamp?
A visible stamp is cropped, covered, or upscaled away. Lorca embeds a classic 1-bit LSB mark, evaluation-only — not a forensic mark. Pair it with a hash-anchored manifest so the claim survives even when the mark does not.
02 Do invisible watermarks survive cropping or re-encoding?
No. The module is classic 1-bit LSB, evaluation-only — re-encoding or cropping can destroy the mark by design. Verify provenance against the hash-anchored manifest, not the pixels alone.
03 Can the watermark be removed without the key?
Removing a key-bound mark without detection breaks the carrier; verification will then fail or report tampering. The key is required to recover and validate the mark.
04 Where is the watermark verified — in a browser, or on a server?
Verification runs on your infrastructure — both the studio demo host and private pilot provide pixel-level verification with no third-party dependency.
05 Is this a C2PA certification / standard-compliant embedding?
No. It is an embedded provenance strip verified on your infrastructure, not a certified badge or platform claim.