Model Context Protocol (MCP) & AI Copilot Integration

Connect external AI assistants directly to live industrial robot telemetry and workspace diagnostics.

Overview

The Model Context Protocol (MCP) is an open standard that enables AI systems to securely interface with development tools and industrial hardware.

The KUKA KRL extension features an embedded local MCP Bridge Server running on 127.0.0.1. It connects modern AI programming environments (Cursor, Google Antigravity, Claude Code CLI, Windsurf) directly to your KUKA workspace, offline AST diagnostics, and live robot controller telemetry.

+-------------------------------------------------------------+
|                      AI Assistant Client                    |
|          (Cursor / Google Antigravity / Claude Code)        |
+-------------------------------------------------------------+
                              |
                    Model Context Protocol
                   (JSON-RPC / HTTP Bridge)
                              |
                              v
+-------------------------------------------------------------+
|             Embedded KUKA KRL MCP Server (127.0.0.1)        |
|  +---------------------------+ +--------------------------+ |
|  | Workspace Diagnostics     | | Live Hardware Telemetry  | |
|  | - AST Syntax Validation   | | - $AXIS_ACT, $POS_ACT    | |
|  | - Safe Auto-Repair Engine | | - Active Tool / Base     | |
|  | - Block Balance Guard     | | - SmartPAD Alarm Stream  | |
|  +---------------------------+ +--------------------------+ |
+-------------------------------------------------------------+
                              |
                     Industrial Network
                              v
+-------------------------------------------------------------+
|                 Physical KUKA Controller                    |
|           (KR C4 / KR C5 via KUKAVARPROXY & WCF)            |
+-------------------------------------------------------------+

Available MCP Tools

External AI assistants gain access to specialized industrial tools:

Tool Name	Scope	Capability
`krl_live_telemetry`	Hardware	Reads real-time axes, Cartesian TCP, active tool/base, operating mode, and alarms from running robot
`krl_workspace_diagnostics`	Project	Audits all `.src`, `.dat`, and `.sub` files for unclosed FOLDs, syntax bugs, and kinematic violations
`krl_safe_auto_repair`	Code	Deterministically repairs structural fold mismatches, missing BRAKE commands, and illegal STEP 0 loops
`krl_system_variable_spec`	Knowledge	Retrieves specifications, physical units, and array limits for 957+ KSS system variables
`krl_alarm_lookup`	Diagnostics	Decodes numeric KSS error codes (e.g. `KSS00203`, `KSS01442`) across 6 languages

Connecting External AI Clients

1. Automatic Cursor Setup

When running inside Cursor, the extension automatically registers the local MCP server via vscode.cursor.mcp.registerServer on launch with zero manual setup.

2. Manual JSON Configuration Export

Use command KUKA KRL: Copy MCP Configuration (Ctrl+Shift+P) to copy client-ready JSON configurations:

json

{
  "mcpServers": {
    "kuka-krl": {
      "command": "node",
      "args": ["c:/Projects/01_Robotics/kuka-krl-mcp/bin/server.js"],
      "env": {
        "KUKA_BRIDGE_PORT": "49152"
      }
    }
  }
}

3. Open-Source Ecosystem

Standalone Server: LiskinLabs/kuka-krl-mcp (MIT License)
AI Prompt Rules & Skills: LiskinLabs/kuka-krl-skills

Model Context Protocol (MCP) & AI Copilot Integration ​

Overview ​

Available MCP Tools ​

Connecting External AI Clients ​

1. Automatic Cursor Setup ​

2. Manual JSON Configuration Export ​