About Opcua MCP Server
Opcua MCP Server is an MCP server in the Commerce category: mCP server that connects AI agents ( Claude, ChatGPT, Gemini, Cursor etc.) to OPC UA systems - read, browse, history, subscriptions, events and alarms.
The repository has 28 stars and 0 forks, with the last commit yesterday. Recent commit activity is the strongest signal that reported issues will get attention.
Install
claude mcp add opcua-mcp-server -- npx -y opcua-mcp-servernpx -y opcua-mcp-serverThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about Opcua MCP Server
Powered by Claude · Grounded in docs
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
OPC UA MCP Server
A Python application that connects to an OPC UA server and exposes it via MCP (Model Context Protocol) using FastMCP.
Features
- Connect to any OPC UA server with optional authentication
- Browse the complete node tree from a configurable starting node
- Expose the OPC UA node tree as MCP resources
- Read values from OPC UA nodes
- Search nodes by display name or browse name with wildcard support
- Support for both stdio and SSE (Server-Sent Events) transport methods
Installation
pip install -r requirements.txt
Usage
Basic usage with HTTP transport (default):
python opcua_mcp_server.py opc.tcp://localhost:4840
With authentication:
python opcua_mcp_server.py opc.tcp://localhost:4840 --username myuser --password mypass
Using stdio transport:
python opcua_mcp_server.py opc.tcp://localhost:4840 --transport stdio
With custom starting node and HTTP port:
python opcua_mcp_server.py opc.tcp://localhost:4840 --start-node "i=84" --http-port 8080
Command Line Arguments
opcua_url(required): OPC UA server URL (e.g., opc.tcp://localhost:4840)--start-node: Starting node ID for browsing (default: i=85)--username: Username for OPC UA authentication--password: Password for OPC UA authentication--transport: MCP transport method - 'stdio' or 'http' (default: http)--http-port: Port for HTTP transport (default: 3000)
MCP Tools
The server provides the following tools:
read-value
Read the current value of one or more OPC UA nodes.
Parameters:
node_ids: List of node IDs to read
Example:
{
"node_ids": ["ns=2;s=MyVariable", "ns=2;s=Temperature"]
}
find-by-display-name
Find nodes by display name pattern (supports wildcards).
Parameters:
pattern: Display name pattern with wildcards (e.g., Temperature)
Example:
{
"pattern": "*Temperature*"
}
find-by-browse-name
Find nodes by browse name pattern (supports wildcards).
Parameters:
pattern: Browse name pattern with wildcards (e.g., Sensor)
Example:
{
"pattern": "*Sensor*"
}
Node Information
Each node in the tree contains:
node_id: The OPC UA node identifierbrowse_name: The browse name of the nodedisplay_name: The human-readable display namenode_class: The type of node (e.g., Object, Variable, Method, etc.)children: List of child nodes (only for non-leaf nodes)
Example
- Start an OPC UA server (e.g., using
opcua-asyncioexample server) - Run the MCP server:
python opcua_mcp_server.py opc.tcp://localhost:4840 - The MCP server will connect to the OPC UA server, browse the node tree, and expose it via MCP
- You can now use any MCP client to interact with the OPC UA data
Error Handling
- The server will exit if it cannot connect to the OPC UA server
- Individual node read errors are returned in the response
- The server properly disconnects from the OPC UA server on shutdown
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.