Skip to main content
You will understand how media is authenticated from the moment of capture. Provenance guarantees that a piece of media is genuine and unaltered. Extra Scoop establishes this through hardware-backed cryptographic signatures and continuous forensic auditing.

Hardware Attestation

When you capture footage using the Eyewitness App, the application does not just save a file. It requests a cryptographic signature directly from the physical hardware (such as the Apple Secure Enclave). This proves the media originated from a genuine device, rather than an emulator or a spoofed API request.

Forensic Invariance

Once uploaded, the raw media undergoes a forensic invariance process.
1

Metadata Scrubbing

Edge functions strip personally identifiable information (PII) from the EXIF data to protect your privacy, while preserving the cryptographic hashes that prove authenticity.
2

Silicon Trust™ Watermarking

The media receives an invisible neural watermark. This embeds the unique identifier of the asset directly into the pixels or video frames.
3

Raw File Purge

The system permanently deletes the raw, unwatermarked source file from storage. This ensures the unwatermarked master footage cannot be leaked or misused.
The Silicon Trust™ Protocol provides enterprise-grade tamper resistance, deepfake detection, and cryptographic origin verification for all visual assets ingested into the Extra Scoop ecosystem.

2. Forensic Invariance & Permanent Raw Purge Pipeline

To guarantee that an unwatermarked raw file can never be leaked or scraped:
  1. The raw asset is sent to the Silicon Trust microservice over HMAC SHA-256 (Internal Cryptographic Signature).
  2. The sealed, watermarked asset is verified for byte integrity via HTTP HEAD.
  3. Permanent Purge: The original raw unwatermarked file is immediately deleted from Supabase Storage (Internal Storage Purge Routine).
  4. Only the protected, watermarked file is passed to Mux for video streaming and public display.

4. Fallback Engine

  • Algorithm: High-redundancy frequency-domain embedding.
  • Error Correction: Advanced cryptographic error-correction protocols.
  • Fault Tolerance: Highly resilient against aggressive adversarial degradation and compression.