The paper introduces Embodied.cpp, a portable C++ inference runtime for deploying vision-language-action (VLA) and world-action models (WAMs) across heterogeneous robots, edge devices, and simulators. Its five-layer architecture covers input adapters, sequence builders, backbone execution, head plugins, and deployment adapters. The runtime targets embodied control requirements such as multi-rate execution, latency-first batch-1 inference, and extensible interfaces beyond token I/O. The reported closed-loop deployments achieve 100.0% success for HY-VLA and 91.0% for pi0.5, while a preliminary WAM benchmark reduces block memory from 312.2 MiB to 88.1 MiB.
No heat snapshots are available in the last 24 hours.