Continuo at ECCV 2026

At ECCV 2026, Action Intelligence demonstrated Continuo running live on smart glasses.

Designed for always-on perception, Continuo continuously observes the world from a first-person perspective, connects observations and events across time, and builds contextual understanding and memory from ongoing human experiences. During the live demonstration, Continuo showed how it could understand a day of conference talks, poster sessions, and conversations, extract meaningful information from these continuous experiences, generate daily summaries, and suggest plans for what comes next.

Rather than responding to isolated frames or individual commands, Continuo focuses on understanding and organizing experiences as they unfold over time. Throughout the conference, we had the pleasure of welcoming many distinguished researchers to our booth, where we shared live demonstrations of Continuo and exchanged ideas about the future of continuous perception and contextual understanding.

For Action Intelligence, Continuo is more than a smart glasses application. As the first chapter of our technical journey, it explores how continuous first-person observation can help intelligent systems develop a deeper understanding of the physical world and human behavior.

From understanding the world through continuous observation, to transforming human experience into robot training data, and ultimately enabling autonomous action and continuous learning in real-world environments, Continuo marks the starting point of our System Scaling approach.