A new smart security camera developed with the QVAC SDK offers a privacy-centric alternative by performing all its processing on-device, ensuring no footage is uploaded to the cloud. This innovation addresses growing concerns regarding cloud-based cameras that could expose sensitive information. The camera is designed to recognize people and vehicles, assess the risk level of a scene, and sound an alarm, all without transmitting any data externally.

QVAC: QVAC is Tether’s decentralized local AI platform and SDK that enables running AI models privately and entirely on user devices across platforms like Linux, macOS, Windows, Android, and iOS. It provides a unified API for capabilities including model loading, inference, embeddings, and specialized tasks like vision processing without any cloud dependency or data transmission. In this news, QVAC powers a demo security camera that performs on-device person and vehicle detection, risk assessment, and alarm triggering while keeping all footage local.

Privacy Focus: On-device AI processing addresses widespread concerns about cloud-based cameras uploading sensitive footage to third-party servers.
SDK Development: Recent updates to the QVAC SDK have expanded support for vision models and cross-platform local AI applications suitable for edge devices.