How Forensic Watermarking Works.
A two-layer system that embeds invisible, traceable forensic marks into content — with no measurable impact on performance.
Two Layers. Complete Traceability.
Image Angel uses a two-layer watermarking system. The first layer identifies the platform. The second layer identifies the individual viewer. Together, they create an unbroken forensic chain from the moment content is accessed to the moment a leak is attributed.
Layer 1: Platform Watermark
Applied once at upload. Identifies the source platform. Cached for 30 days. Generated server-side with no impact on content delivery.
Layer 2: User Watermark
Applied at the moment of access. Unique to every viewing session. Cached for 24 hours per user per dimension. This is the layer that makes individual attribution possible.
Extraction
When a leaked file is submitted, Image Angel analyses it and returns the platform ID, session ID, and timestamp. The process works even after screenshots, compression, cropping, and messaging app re-encoding. Attribution confidence: 90%+.
No Measurable Impact on Performance.
The watermark overlay runs asynchronously, downstream of your CDN. Time-to-first-byte remains unchanged. There's no client-side JavaScript and no blocking requests. Your performance metrics stay where they are. If the API is unreachable, content delivers normally — the worst case is a temporary gap in traceability.
Privacy by Design.
No Personal Data Embedded
The watermark contains opaque session identifiers only. No names, no emails, no IP addresses, no device IDs. The identifiers are meaningful only to the platform that generated them.
No User Surveillance
Image Angel does not track behaviour, monitor browsing, or profile users. It marks content at the moment of access. That's it.
No Cross-Platform Tracking
Watermarks are platform-local. A mark embedded by one platform cannot be read by another. There is no central database of user identities.
What the Watermark Survives.
Screenshots. Screen recordings. JPEG compression. Cropping. Resizing. Messaging app re-encoding. The forensic watermark is embedded in the visual structure of the image — not in metadata or EXIF data. If enough of the original image remains visible to a person, the watermark is recoverable by our system.
What It Supports.
Developer's Corner.
REST API. JSON responses. Standard authentication. Server-side integration into your existing media pipeline. No proprietary frameworks. No lock-in. This is infrastructure, not a product bolted on.
POST /Watermark/Platform
Request platform overlay. Identifies the source platform. Cache: 30 days.
POST /Watermark
Request user payload and convert to overlay. Unique per session. Cache: 24 hours.
POST /Extract
Submit leaked file for forensic analysis. Returns platform ID, session ID, timestamp.
Get Credentials
API key and client ID issued. Sandbox environment provisioned.
Call API
Server-side call to the watermarking endpoint from your media pipeline.
Apply Overlay
Composite the returned overlay onto the source image before delivery.
Cache & Serve
Cache the watermarked output. Serve through your existing CDN. Done.
Image Angel sits downstream of your CDN, between your origin server and content delivery. It processes the overlay after your existing pipeline has finished. No new infrastructure. No new dependencies.
Requirements.
API key and client ID — provided by Image Angel on signing.
Image processing capability — ImageMagick, Sharp, GD, or Pillow.
Your existing content storage — no changes required.
Reference Implementation.
An open-source WordPress plugin exists as a working reference. Install, activate, connect your API key. Full forensic protection in five minutes. No server configuration required. Source available for inspection.
VIEW PLUGIN >>>Request API Access.
Full documentation, sandbox credentials, and a named engineer on our side. Typically within 48 hours.
REQUEST API ACCESS