WiseVision/mcp_server_ros_2

Overview

A Python-based MCP server that connects AI agents to ROS 2 robotics systems. It enables agents to interact with robotic hardware and software through standardized MCP interfaces. The server handles communication between the AI agent and ROS 2 topics, services, and actions.

Best for

Best for
Developers connecting MCP-compatible AI agents to real ROS 2 robots

Use cases

• Controlling robot navigation via natural language commands
• Feeding sensor data from ROS 2 topics into AI decision models
• Triggering robot actions through MCP tool calls

How to use

Install

npx @modelcontextprotocol/inspector uv --directory /path/to/ros2_mcp run mcp_ros_2_server

Tools exposed

• ros2_topic_list
• ros2_topic_subscribe
• ros2_get_messages
• ros2_get_message_fields
• ros2_topic_publish
• ros2_service_list
• ros2_service_call
• ros2_list_actions
• ros2_send_action_goal
• ros2_cancel_action_goal
• ros2_action_request_result
• ros2_action_subscribe_feedback

Tested with

Claude Desktop, Visual Studio Code Copilot, Warp

Notes

A Python-based MCP server that connects AI agents to ROS 2 robotics systems. It enables agents to interact with robotic hardware and software through standardized MCP interfaces. The server handles communication between the AI agent and ROS 2 topics, services, and actions.

78 stars on GitHub. Last updated 2026-05-08. Licensed MPL-2.0.

Use cases

• Controlling robot navigation via natural language commands
• Feeding sensor data from ROS 2 topics into AI decision models
• Triggering robot actions through MCP tool calls

Pros

• Uses the MCP standard for seamless integration with compatible agents
• Python codebase with 78 stars indicating community traction
• Directly bridges AI agents to ROS 2 without proprietary middleware

Cons

• Relatively new project with limited adoption (78 stars)
• Requires proficiency in both ROS 2 and the Model Context Protocol
• No documented support for ROS 1 or simulation environments

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.