Unitree Go2 MCP Server

Example Use Cases

• 1

  Convert natural language intent into ROS2 commands to control a Unitree Go2 robot (movement, behaviors, or demos).

• 2

  Integrate LLM-driven contextual reasoning with robot control, e.g., interpreting high-level user requests into timed velocity commands or safety actions.

• 3

  A local MCP server to interface LLMs with the unitree_ros2 ecosystem and optionally rosbridge for networked ROS2 access.

Description

This repository provides a Model Context Protocol (MCP) server that lets agents send natural language commands which are interpreted by a large language model and translated into ROS2 actions for the Unitree Go2 robot. It integrates with the unitree_ros2 stack and optionally rosbridge, and includes configuration guidance and demo tasks (movement, obstacle avoidance, contextual interactions). The server is intended to run on a PC connected to the robot and helps bridge LLM-based agents with real-world robot control.