Vista: Edge AI Gateway for Real-Time Retail Intelligence
Vista is a high-performance, on-premises Edge AI gateway and VMS, purpose-built for large-scale deep learning and computer vision processing. It ingests video from hundreds of in-store cameras or multiple VMS platforms, executing AI models in parallel to provide real-time insights on shopper demographics, identity, behavior, loss prevention, vehicle fingerprinting and movement analytics. Vista is designed for stores where the existing VMS host lacks sufficient processing, memory, or networking capacity to support simultaneous, enterprise-scale AI workloads.
Vista empowers retailers to run AI at the edge with low latency, high reliability, and enterprise-scale performance, feeding actionable intelligence into the MindGlyph platform for analytics, shopper engagement, and loss prevention.
Vista Edge AI Gateway Overview
Platform Capabilities
High-Performance Edge Processing
Up to 6 TOPS of compute power via ARM CPUs, NPUs, and GPUs on-device.
Optimized AI Model Execution
Proprietary toolkits tune deep learning models for efficient, low-power inference.
Scalable Architecture
Vertical and horizontal scaling for enterprise-grade distributed edge AI clusters.
ONVIF-Compliant VMS Integration
Full video capture, analysis, recording, and monitoring with forensic-quality evidence.
Flexible Data Delivery
Proprietary P2P networking with NAT traversal for secure, configuration-free cloud delivery.
High-Performance Edge Processing
Combines multi-core ARM CPUs, dedicated Neural Processing Units (NPUs), and high-performance GPUs to deliver up to 6 TOPS (Trillions of Operations Per Second) of processing power per board, enabling real-time AI tasks such as face recognition, behavior analysis, and shoplifting detection directly on-device with reduced latency, enhanced privacy, and lowered power consumption compared to consumer-grade computers that commonly host the VMS.
Optimized AI Model Execution
Proprietary toolkits convert and tune deep learning models for Vista hardware, ensuring efficient, low-power inference while maintaining high accuracy across multiple AI workloads.
Scalable Architecture
Supports vertical scaling with upgraded chips and horizontal scaling by networking multiple devices into a distributed edge AI cluster, enabling enterprise-scale deployments.
ONVIF-Compliant VMS Integration
Preloaded with our feature-rich, ONVIF-compliant VMS, Vista supports the full spectrum of video capture, analysis, recording, and monitoring capabilities, delivering high-performance, low-latency edge processing tightly integrated with MindGlyph AI to provide real-time insights and forensic-quality evidence generation capabilities.
Comprehensive Video Capture
Supports ONVIF-compliant IP, USB, and analog cameras, offering flexible recording modes including continuous, motion-triggered, and event-based monitoring without recording.
Advanced Camera Controls
Offers customizable resolution, frame rate, codec, and bandwidth settings to optimize performance across devices and network conditions, with configurable zones to filter out irrelevant motion or activity.
Virtual Tripwire
Enables configuration of uni- or bidirectional tripwires within defined video zones to trigger specific actions when a shopper crosses the boundary.
Blended Event Summaries
Automatically generates synchronized, multi-camera video timelines suitable for law enforcement or legal evidence, preserving timestamps and metadata for full admissibility.
Remote Accessibility
Provides full web and mobile app access for Android and iOS, enabling managers and security personnel to monitor, review, and control streams from anywhere, with bandwidth-aware streaming that automatically adjusts quality based on network conditions.
Event and Storage Management
Offers detailed filter options, event prioritization, and efficient storage management, making it easy to search, retrieve, and archive critical footage.
Flexible Data Delivery
Uses proprietary P2P networking with NAT traversal to securely send processed and aggregated data to the cloud without requiring changes to the store’s network configuration; optional mobile data cards enable independent connectivity.