Anthropic releases Model Hardware Standard to let AI agents run lab devices

On August 29, Anthropic released a research preview of Model Hardware Standard (MHS), a hardware and communication protocol standard for physical lab devices that lets Claude and other AI agents safely operate microscopes, liquid-handler robots, mechanical arms, and other lab equipment.

MHS compresses device integration time from months to hours, removing the need for integrators to write dedicated control interfaces for each instrument. Anthropic positions the standard in the announcement as a key piece of "AI controlling the physical world", complementing Computer Use and the Skills API to form a complete Agent capability stack from the lab to the real world.

MHS is released by Anthropic Labs, with open-source reference implementations and SDKs to facilitate hardware vendor integration. Several biotech companies and robotics vendors have begun integration, including cultivated-meat labs and automated synthetic biology platforms.