Shanghai AI Lab Launches SCP-Lab to Enable Autonomous Physical Lab Experiments

The Shanghai AI Laboratory has recently unveiled ShuSheng SCP-Lab, an experiment device access and collaboration protocol designed to bridge the gap between artificial intelligence and physical laboratory environments. Built upon the laboratory’s previously released Scientific Context Protocol, the new system addresses a critical bottleneck in automated scientific research: the heterogeneous nature of experimental equipment across different manufacturers and platforms. Existing lab systems lack standardized interfaces, making it difficult for AI agents to safely and efficiently orchestrate cross-platform automation. ShuSheng SCP-Lab resolves this by establishing a unified, layered interaction framework that decouples high-level task scheduling from low-level hardware control. The protocol operates across five distinct control granularity levels, ranging from L1, which manages complete pre-packaged experimental workflows, down to L5, which allows direct, atomic control of hardware actuators. This tiered structure enables diverse equipment to be seamlessly integrated without overriding existing local safety and execution systems. By standardizing capability discovery, task submission, and result feedback, the protocol ensures that AI-driven scientific planning can be reliably translated into physical experiments. The system has successfully undergone validation across nine distinct scientific scenarios, including chemical synthesis, material optimization, drug screening, molecular biology, and protein-directed evolution. Key academic and research institutions, including Tsinghua University, the Lingang Laboratory, and Shanghai Jiao Tong University, have already integrated the protocol to accelerate their wet-lab workflows. Researchers at Shanghai Jiao Tong University utilized the system to streamline the iterative development of lipid nanoparticle formulations, establishing a closed-loop pipeline from computational prediction to robotic experimentation. Similarly, a collaborative effort between Tsinghua University and the Shanghai AI Laboratory successfully coordinated a multi-node protein evolution project across three cities, demonstrating the protocol’s capacity to synchronize distributed research tasks while maintaining full result traceability. By providing a secure, standardized interface for cross-vendor and cross-regional laboratory equipment, ShuSheng SCP-Lab effectively powers the laboratory’s broader ShuSheng DuanYan scientific discovery platform. The protocol’s built-in state tracking and fault-tolerance mechanisms further ensure reliable operation during network interruptions or equipment conflicts. Ultimately, this advancement significantly lowers the technical threshold for AI-driven dry-wet experimental cycles, establishing a scalable foundation for autonomous scientific discovery and accelerating the integration of machine intelligence into real-world research infrastructures.