Building AURA toward a fully independent biometric stack

AURA Neural currently uses browser-native biometrics to inform adaptive sound and state-based sessions.

One of the areas I am actively developing is AURA’s own facial-geometry and signal-acquisition layer. The goal is to remove the only remaining third-party, FaceMesh/MediaPipe, dependency and make the biometric pipeline fully independent.

That means the long-term architecture is intended to control the full path:

camera signal → facial geometry → physiological sampling → state interpretation → adaptive output

This work is already in progress. The current focus is reproducing the stability and signal quality of the existing reference pipeline before removing the dependency completely.

The objective is not to replace a working component for the sake of it. It is to make AURA more self-contained, easier to evolve, and less dependent on external computer-vision infrastructure.

AURA is not a medical device, and I am deliberately keeping the physiological claims conservative while the sensing layer continues to be validated.

That’s the part of the system I am currently pushing forward.