Universal Robot MCP Server
A Model Context Protocol (MCP) server that provides AI assistants and LLM applications with secure, controlled access to Universal Robots functionality. This server enables real-time robot control, status monitoring, and motion planning through a standardized MCP interface.
Features
Robot Connection Management - Connect/disconnect from UR robots safely
Real-time Status Monitoring - Get joint positions, poses, and robot health
Joint Motion Control - Precise angular movement with safety limits
Linear Motion Control - Cartesian path planning and execution
Simulation Mode - Test and develop without physical hardware
Safety First - Built-in collision detection and movement validation
Installation
Quick Start with uvx (Recommended)
Install via pip
Development Installation
Usage
AI Assistant Integration
Claude Desktop
Cursor / Other MCP Clients
VS Code with MCP
Direct Usage
Available Tools
Tool | Description | Parameters |
| Establish connection to UR robot |
(default: 192.168.1.100) |
| Safely disconnect from robot | None |
| Get current joint positions and pose | None |
| Move to specific joint angles |
,
,
|
| Linear movement in Cartesian space |
,
,
|
Example Conversations
Once configured with your AI assistant:
"Connect to the robot and show me its current status"
"Move the robot to home position safely"
"Execute a pick and place motion from coordinates X to Y"
"What are the current joint angles?"
"Move the robot 10cm up in the Z direction"
Robot Configuration
Network Setup
Default robot IP:
192.168.1.100Ensure robot is connected to your network
Verify robot is in Remote Control mode
Safety Features
Automatic TCP and payload configuration
Movement speed and acceleration limits
Connection timeout handling
Emergency stop capabilities
Supported Platforms
AI Assistants: Claude Desktop, Cursor, Roo Code, Cline
IDEs: VS Code, JetBrains IDEs (with MCP plugins)
Platforms: macOS, Linux, Windows
Python: 3.8, 3.9, 3.10, 3.11
License
GPL-3.0