Back to feed
News Story
APriority79
量子位
1 sources

Anthropic Unveils Model Hardware Standard (MHS) to Let Claude Control Physical Devices

Anthropic has released the Model Hardware Standard (MHS), a new standard that unifies hardware devices from different manufacturers into a common interface for AI agents, enabling Claude to directly interact with the physical world. Originating from a collaboration with HHMI Janelia, MHS has already allowed Claude to control a SO-ARM101 robotic arm and laboratory instruments, paving the way for distributed embodiment.

SynthePulse Insight · AI deep readingMembers

Claude's 'Superbody' Route: How the MHS Standard Lets AI Take Over the Physical World

Version 1 · 1 source

Anthropic releases the Model Hardware Standard (MHS), enabling Claude to call any hardware without a fixed body, from robotic arms to microscopes, ushering in a new paradigm of distributed embodiment.

  • Anthropic releases the Model Hardware Standard (MHS), aiming to unify hardware interfaces so AI agents can directly control physical devices.
  • MHS originated from a collaboration with HHMI Janelia to address the issue of incompatible lab equipment interfaces.
  • Claude can already control the SO-ARM101 robotic arm via MHS without pretraining or teleoperation.
Open section navigationWhat is MHS: The MCP of the Physical World

What is MHS: The MCP of the Physical World

On August 28, 2026, Anthropic released the Model Hardware Standard (MHS), a set of hardware-oriented driver and device description standards, often likened to 'the MCP of the physical world.' While MCP unifies software, data sources, and tools into standard interfaces that agents can call, MHS translates the control methods of various hardware devices into consistent interfaces that agents can discover, read, and invoke.

MHS originated from a collaboration between Anthropic and the HHMI Janelia Research Campus, aiming to solve the problem of incompatible interfaces across lab instruments from different vendors, which often cannot communicate and are difficult for AI to control uniformly. Traditionally, integrating these devices required engineers to write extensive 'glue code,' taking weeks or even months, and agents struggled to understand device capabilities, statuses, and risks.

Free for now

Read the full analysis

4 more sections of analysis, plus the full takeaway

Loading

Credibility boundary

This article's information primarily comes from reports by QbitAI, a secondary source. Key facts such as the MHS release and capability demonstrations are cited from Anthropic's official announcements, but specific details have not been independently verified. Some descriptions, like 'Claude can control robotic arms,' are based on official demos and may carry promotional elements.

Sources for this version

  1. Claude开始接管物理世界！能用机械臂阻拦5000万美元打款了量子位

Primary report

量子位

Primary source