The platform — SenseCryptA person-bound
root of trust.
In the PKI flow, a live biometric generates post-quantum key-pairs in memory. Retain the public keys for verification and encapsulation. The private keys exist only during face sign or face decapsulate and are never stored. Public keys, signatures and decapsulated shared secrets go directly to relying party endpoints, without the IdP seeing them.
≥256BITS
Residual-entropy construction targetThe entropy figure is a construction target, not a measurement or NIST certification. In PKI, public keys are stored and private keys exist only during face sign or face decapsulate. Enrollment artifacts, certificates and infrastructure keys have separate lifecycle requirements. Public keys, signatures and decapsulated shared secrets go directly to relying party endpoints, without the IdP seeing them.
Four disciplines — fused into one primitive
MLMachine learningA high-dimensional random projection that separates genuine and impostor margins by design.
PQCPost-quantumML-DSA (FIPS 204) supports signatures; ML-KEM (FIPS 203) supports encapsulation and decapsulation. Store the public keys for standard library use. Face sign and face decapsulate generate the private keys transiently. Algorithm support is not module validation.
ZKZero-knowledgeSTARK-based, transparent setup — no trusted ceremony to compromise. Post-quantum security depends on the full construction and parameters.
POLARSoft-decision ECCConfidence-weighted decoding that tolerates realistic face drift — with a minimum 256-bit residual-entropy construction target.
The moatCompetitors pick one discipline. We fused all four to target the residual-entropy threshold.