On-device human sensing runtime giving edge AI agents real-time physical awareness without cloud dependencies.
Fully open-source codebase without license restrictions for local edge and microcontroller deployments.
EDNA addresses the pervasive security and privacy vulnerabilities inherent in smart home devices and cameras that constantly stream intimate video feeds to cloud servers for inference. It delivers an ultra-lightweight human perception runtime designed to run directly on microcontrollers and edge compute modules. By fusing low-power radar signals, thermal arrays, and micro-sensors, EDNA detects human presence, micro-respiration, approaching gait, and spatial gestures within milliseconds, entirely without capturing high-resolution visual frames, establishing a privacy-first foundation for ambient computing.