B&R Automation Studio MCP Server
An MCP (Model Context Protocol) server that enables AI assistants like Claude to interact with B&R Automation Studio projects. Build projects, run simulators, and read/write OPC UA variables - all through natural language.
Features
- Build Projects - Compile Automation Studio projects for simulation
- Run Simulator - Launch ARsim and connect automatically
- OPC UA Integration - Read and write PLC variables in real-time
Prerequisites
- Windows 10/11 with B&R Automation Studio 4.x installed
- Python 3.10+ on Windows (with pip)
- Claude Code (native Windows or WSL2)
Installation
1. Clone this repository
git clone https://github.com/AndrewMusser/br-automation-mcp.git
cd br-automation-mcp
2. Install Python dependencies
pip install -r requirements.txt
3. Configure server settings
Edit server.py to match your environment:
3a. B&R Tool Paths
Update these paths to match your Automation Studio installation:
AS_BUILD_PATH = "C:\\BrAutomation\\AS412\\bin-en\\BR.AS.Build.exe"
PVI_TRANSFER_PATH = "C:\\BrAutomation\\PVI\\V4.12\\PVI\\Tools\\PVITransfer\\PVITransfer.exe"
Common installation locations:
- AS 4.12:
C:\BrAutomation\AS412\bin-en\BR.AS.Build.exe - AS 4.9:
C:\BrAutomation\AS49\bin-en\BR.AS.Build.exe - PVI: Typically under
C:\BrAutomation\PVI\V4.x\PVI\Tools\PVITransfer\
3b. OPC UA Parameters
Configure the OPC UA connection settings for communicating with ARsim:
OPCUA_HOST = "localhost"
OPCUA_PORT = 4840
OPCUA_USERNAME = "Admin"
OPCUA_PASSWORD = "password"
Important: These credentials must match your Automation Studio project's OPC UA configuration:
- User/role settings:
Physical/<Config>/<CPU>/AccessAndSecurity/UserRoleSystem/ - OPC UA config:
Physical/<Config>/<CPU>/Connectivity/OpcUA/
4. Configure Claude Code
Choose the setup that matches how you run Claude Code:
Option A: Native Windows (Recommended)
Claude Code now runs natively on Windows via Git Bash, PowerShell, or CMD. This is the simplest setup since both Claude Code and Automation Studio run on the same machine.
Install Claude Code on Windows:
winget install Anthropic.ClaudeCode
Configuration file: %APPDATA%\Claude\settings.json
Add to settings:
{
"mcpServers": [
{
"name": "br-automation-mcp",
"type": "stdio",
"command": "python",
"args": ["D:\\path\\to\\br-automation-mcp\\server.py"],
"enabled": true
}
]
}
Architecture:
┌─────────────────────────────────────────────────────────────┐
│ Windows │
│ │
│ ┌─────────────┐ ┌─────────────────────────────────────┐ │
│ │ Claude Code │───▶│ server.py (MCP STDIO Server) │ │
│ │ (Git Bash) │ └───────────────┬─────────────────────┘ │
│ └─────────────┘ │ │
│ │ │
│ ┌───────────────────────────┼───────────────┐ │
│ ▼ ▼ ▼ │
│ ┌─────────────┐ ┌──────────────┐ ┌─────────┐ │
│ │ BR.AS.Build │ │ PVITransfer │ │ OPC UA │ │
│ │ (Build) │ │ (Simulate) │ │ (Test) │ │
│ └─────────────┘ └──────────────┘ └─────────┘ │
└─────────────────────────────────────────────────────────────┘
Option B: WSL2 (Cross-Platform)
If you run Claude Code in WSL2, it can invoke the Windows Python server via PowerShell interop.
Configuration file: ~/.claude/settings.json
Add to settings:
{
"mcpServers": [
{
"name": "br-automation-mcp",
"type": "stdio",
"command": "powershell.exe",
"args": [
"-ExecutionPolicy",
"Bypass",
"-File",
"D:\\path\\to\\br-automation-mcp\\run-server.ps1"
],
"enabled": true
}
]
}
Architecture:
┌─────────────────────────────────────────────────────────────┐
│ WSL2 (Linux) │
│ ┌─────────────────┐ │
│ │ Claude Code │ │
│ │ │──── powershell.exe ────┐ │
│ └─────────────────┘ │ │
└─────────────────────────────────────────────│───────────────┘
│
┌─────────────────────────────────────────────│───────────────┐
│ Windows ▼ │
│ ┌─────────────────┐ ┌─────────────────────────────────┐ │
│ │ run-server.ps1 │───▶│ server.py (MCP STDIO Server) │ │
│ └─────────────────┘ └───────────────┬─────────────────┘ │
│ │ │
│ ┌───────────────────────────────┼───────────────┐ │
│ ▼ ▼ ▼ │
│ ┌─────────────┐ ┌──────────────┐ ┌─────────┐ │
│ │ BR.AS.Build │ │ PVITransfer │ │ ARsim │ │
│ │ (Build) │ │ (Simulate) │ │ (OPC UA)│ │
│ └─────────────┘ └──────────────┘ └─────────┘ │
└─────────────────────────────────────────────────────────────┘
5. Restart Claude Code
After updating the configuration, restart Claude Code to load the MCP server.
Available Tools
build_automation_studio_project
Compiles an Automation Studio project for simulation.
Parameters:
project_file- Path to the .apj file (Windows path)configuration- Name of the configuration to build
Example prompt: "Build my project at D:\Projects\MyApp\MyApp.apj using the X20CPU configuration"
run_automation_studio_simulator
Creates ARsim structure and launches the simulator.
Parameters:
ruc_package- Path to RUCPackage.zip (generated by build)start_simulator- Whether to start immediately (default: true)
Example prompt: "Start the simulator using the RUC package at D:\Projects\MyApp\Binaries\X20CPU\X20CP1686X\RUCPackage\RUCPackage.zip"
opcua_read_write
Reads or writes OPC UA variables on the running simulator.
Parameters:
node_id- OPC UA node identifier (e.g.,ns=6;s=::TaskName:varName)operation- "read" or "write"value- Value to write (for write operations)value_type- Data type: Boolean, Int16, Int32, Float, Double, String
Example prompts:
- "Read the temperature variable from the TempCtrl task"
- "Set the targetTemp to 75.0 in the TempCtrl task"
OPC UA Node ID Format
For task-local variables:
ns=6;s=::<TaskName>:<VariableName>
For global variables:
ns=6;s=::AsGlobalPV:<VariablePath>
Note: Variables must be exposed in the OPC UA configuration (OpcUaMap.uad) to be accessible.
Troubleshooting
"Command not found" errors
Ensure Python is in your Windows PATH and the B&R tool paths in server.py are correct.
OPC UA connection failures
- Verify the simulator is running
- Check that variables are exposed in OpcUaMap.uad
- Default credentials are Admin/password
WSL2 can't find powershell.exe
Ensure WSL2 has Windows interop enabled (default). Check with:
which powershell.exe
License
MIT License - see LICENSE for details.
Contributing
Contributions are welcome! Please feel free to submit issues and pull requests.
Acknowledgments
- Model Context Protocol by Anthropic
- FastMCP for the Python MCP SDK
- B&R Automation for Automation Studio