ReCA was proposed in March 2025 by a research team from Georgia Institute of Technology, the University of Minnesota, and Harvard University. The related research findings were published in the paper "..."ReCA: Integrated Acceleration for Real-Time and Efficient Cooperative Embodied Autonomous Agents".

ReCA is a featureization and co-design framework for accelerating collaborative embodied agent systems, aiming to improve task efficiency and system scalability. At the algorithmic level, ReCA supports efficient local model processing to alleviate the significant model cost. At the system level, ReCA proposes a dual-memory architecture integrating long-term and short-term memory, a hierarchical co-planning scheme combining centralized and distributed collaboration, and plan-guided multi-step execution to achieve efficient and scalable collaborative embodied agent computation. At the hardware level, ReCA employs a heterogeneous hardware system, including a high-level planning GPU subsystem and a low-level planning accelerator subsystem, to ensure efficient and robust task execution.