The rapid growth of detectors, predicted to double in five years, is challenging conventional computing with large data, high energy use, and delays. In-sensor computing offers a game-changing solution by processing data directly in the sensor, inspired by the brain. However, current in-sensor computing detectors face issues with complexity, cost, and scalability. Building on our promising preliminary results, we propose the first perovskite-based in-sensor computing detector for light processing, with potential applications in robotics, monitoring, and autonomous systems. Using perovskite’s ability to convert light into energy, we aim to develop self-powered sensors, unlocking groundbreaking applications in surveillance, wearables, and IoT.