Anthropic on Thursday opened a research preview of the Model Hardware Standard (MHS), a shared specification intended to let artificial intelligence agents operate physical devices in laboratories and industrial facilities.
The standard enables AI agents to simultaneously control multiple instruments such as microscopes, liquid handlers and robotic arms, reducing hardware integration time from weeks or months to hours or minutes. It operates with any device that has a programmable interface and is model-agnostic, allowing any agent to access it via standard protocols like the Model Context Protocol.
A standardized driver translates commands between the operating system and hardware, making each device discoverable in a uniform format without custom integration programs. Anthropic developed the MHS in collaboration with HHMI Janelia Research Campus and has lined up support from Amazon Web Services, Automata, Tecan, Universal Robots, Danaher, Doosan Robotics and QIAGEN.
AWS will offer MHS support through its Strands Robots library for connecting AI agents to physical devices, while Automata plans to add support for its LINQ lab automation platform. Tecan is incorporating MHS into its Fluent liquid handling platforms, and Universal Robots intends to add support to its robotics platform. Danaher is exploring how MHS can enable smart instruments and autonomous labs to scale biomedical research, and Doosan Robotics is testing the standard with robotic arms for automated quality control. QIAGEN is experimenting with MHS on its QIAsymphony Connect nucleic acid purification platform.
During the research preview period, Anthropic and its launch partners will build safety evaluations and design safeguards for using AI with physical equipment. After the preview, the company plans to make the standard open-source and publish its findings as part of its guidelines for safe deployment. Interested parties across sectors are invited to sign up for the research preview waitlist.
