Smart City Expo World Congress is the world’s biggest and most influential event on urban innovation. As OpenZeka, we will be exhibiting for the third time alongside NVIDIA at our stand, where we will showcase our most innovative solutions for smart city applications.

At the Smart City Expo World Congress 2025 in Barcelona, OpenZeka is showcasing NVIDIA’s AI-powered technologies that are redefining the future of urban living. Central to the presentation is the transformative role of Smart Agents, powered by Large Language Models (LLMs) and Vision-Language Models (VLMs), in reshaping everyday city experiences. The exhibit features a diverse range of demonstrations — from AI-driven video search, summarization, and analytics agents to Omniverse Digital Twin and the Cordatus AI Platform — all underscoring how NVIDIA’s cutting-edge platforms and frameworks are driving innovation across industries and accelerating the evolution of smart city infrastructure.

Cordatus Smart City AI Agent

Modern cities face complex challenges — from traffic accidents and floods to fires and illegal dumping — all monitored by hundreds of cameras across urban environments. Managing this overwhelming data manually is nearly impossible. The Cordatus SmartCity AI Agent transforms city surveillance into intelligent, automated insight. Powered by advanced video analytics, it continuously monitors all camera streams, detects incidents in real time, and triggers context-aware alarms for events such as fires, floods, or traffic accidents. Instead of manually watching endless feeds, city operators can chat directly with the AI agent to query live or historical data, review incidents, and receive instant reports. The SmartCity AI Agent sees, understands, and communicates — ensuring cities stay safer, smarter, and more responsive.

Video Search and Summarization Agent

The NVIDIA AI Blueprint for Video Search and Summarization (VSS) empowers organizations to build intelligent, customizable video analytics agents with ease. Leveraging the power of generative AI, vision-language models (VLMs), large language models (LLMs), and NVIDIA NIM™ microservices, VSS transforms how users interact with video. It enables natural language conversations with visual data — allowing anyone to search, summarize, and ask questions about video content instantly.

With real-time and batch video analysis, AI-driven search and summarization, interactive visual Q&A, and intelligent alert generation, VSS provides a complete framework for understanding video at scale. Integrated tools for object tracking and event verification add precision and reliability to every insight.

By combining advanced vision and language intelligence, VSS accelerates insight generation — summarizing hours of footage up to 100× faster than manual review. It delivers deep scene comprehension through AI-powered understanding and turns vast amounts of video data into actionable intelligence, enabling smarter, faster decisions across industries — from smart cities to enterprise operations.

Cordatus AI Platform

Cordatus is a simple, user-friendly AI inference platform that requires no coding. It allows individuals and organizations to build their own personalized AI computing cloud using their existing devices, making distributed AI infrastructure accessible to everyone.
Cordatus empowers users to run advanced AI workloads seamlessly, including computer vision (CV) models, vision-language models (VLMs), and large language models (LLMs). Users can also deploy and operate customized AI agents designed for their unique business or research needs.
The Cordatus infrastructure natively supports the NVIDIA DeepStream 8.0 SDK, enabling Thor users to efficiently implement and optimize their computer vision pipelines. Additionally, it provides seamless integration with NVIDIA Video Storage and Streaming (VSS), ensuring high-performance, scalable, and reliable video analytics capabilities.