AI company Eluviant, formerly IntelexVision, has launched Aurora Flow, a frontier ‘video understanding’ model purpose-built for live, enterprise-scale surveillance. Already deployed in live environments, and capable of running fully air-gapped, across multiple cameras and in near real-time, Aurora Flow extends a platform trusted by some of the largest organisations in the world, including Airbus, DP World, Prosegur and Vodafone, overseeing over 250 deployments on more than 50,000 camera feeds across five continents.
Eluviant says that Aurora Flow represents a significant technical milestone by solving one of the most challenging problems facing scale commercial deployment of video intelligence: analysing not just what is visible in individual frames of camera footage, but what is happening across a sequence of movement over time. It extends Eluviant’s existing platform, proven in production for years: an unsupervised self-learning engine that flags genuinely unforeseen events, and a vision language model (Aurora) that has sat inside the live alerting decision for the past 18 months.
There are now more than a billion cameras installed worldwide, far more than we can reliably watch, which means most incidents are only picked up long after they have happened, if at all. By recognising movement patterns and contextualising behavioural sequences as they unfold, Aurora Flow unlocks use cases that were previously out of reach for organisations operating in the world’s most secure and sensitive environments.
Callum Wilson, founder and Co-CEO of Eluviant, said: “We believe Aurora Flow is a frontier AI model in surveillance and a step change in what video intelligence can deliver, moving beyond detection and into genuine understanding of behaviours and actions in complex real-world environments. It addresses a challenge that traditional video analytics has struggled to solve efficiently: the ability to understand what is happening when a single still frame is not enough. Things like fighting, climbing and theft have typically required human eyes to detect them accurately - now we can help operators gain a clearer picture of what does and doesn’t need their urgent attention.”