Novel Tech
Three verification methods that enable status.health's 00 architecture.
Machine Vision
Primary method
User logs into health portal. Vision runs on device. Math happens locally.
What happens:
- Signed binary or browser extension deployed
- OCR extracts health data from portal screen
- All processing on user device
- No screenshots or data leave device
- Proof generated in seconds
- Ephemeral - results wiped after use
What is different:
Others upload documents. We don't. No data to breach.
Interactive Demo:
Simulate extracting health data from a portal screen
Date:
Recent Lab Results
API Integration
Secondary method
Works with Apple Health, Epic, Quest, others. User authorizes. Proof generated. Data deleted.
Process:
- Direct from source
- User controls access
- Data used once
- Then deleted
Result:
API accuracy. Zero storage.
Interactive Demo:
Simulate API integration with health provider
AI Extraction
Fallback method
Upload PDF or image. AI reads it locally. No cloud.
Features:
- Works with any PDF or image
- State-of-the-art AI models
- Extraction and proofing locally
- AI runs on user device
- Document never leaves device
- Instant results
Different from:
Others send to cloud. We don't.
Interactive Demo:
This is a simulation - No actual files are processed
Technical Details
How it works
zkVM frameworks (Risc0, SP1). TEEs (Apple Secure Enclave, Intel SGX). Patent pending.
Facts:
- Runs in trusted execution environments
- Uses existing crypto
- Rotating attestation ledger
- Anyone can verify
- Patent: 19/200,691
- Open to audit
Not new math. Just better use of it.
Interactive Demo:
Explore the cryptographic components
Zero-Knowledge Virtual Machine
Why device-only
Math on user device = no data to steal. Privacy by physics, not policy.