node-opcua-modeler-mcp-server
Provides AI agents with offline access to OPC UA companion specification types, dependencies, and engineering units for industrial modeling.
README
node-opcua-modeler-mcp-server
An MCP server that gives AI agents access to the OPC UA companion specification type system — 589 types across 22 industrial namespaces, plus 1,533 engineering units.
Built on node-opcua, the most widely used OPC UA stack for Node.js.
Why?
When an AI agent needs to build an OPC UA information model, it must know:
- What companion spec types exist (DI, Machinery, Robotics, Machine Tools…)
- What components, properties, and methods each type has
- What namespace dependencies are required
- What engineering unit symbols are valid (UNECE Rec. 20)
This MCP server answers all of those questions — offline, for free, in milliseconds.
Quick Start
With Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"opcua-modeler": {
"command": "npx",
"args": ["-y", "node-opcua-modeler-mcp-server"]
}
}
}
With any MCP client
npx node-opcua-modeler-mcp-server
The server communicates over stdio using the Model Context Protocol.
Tools
list_namespaces
List all 25 OPC UA companion spec namespaces with aliases, URIs, and dependencies.
→ list_namespaces()
← [
{ "alias": "di", "name": "OPC UA for Devices", "uri": "http://opcfoundation.org/UA/DI/", "dependencies": [] },
{ "alias": "robotics", "name": "OPC UA for Robotics", "uri": "http://opcfoundation.org/UA/Robotics/", "dependencies": ["di", "ia"] },
...
]
resolve_dependencies
Resolve the full dependency chain for companion spec aliases. Returns a topologically sorted list for the YAML namespaces: block.
→ resolve_dependencies({ aliases: ["machineTool"] })
← ["di", "ia", "machinery", "isa95JobControl", "machineryJobs", "machineTool"]
list_types
List all ObjectTypes and VariableTypes defined in a companion spec namespace.
→ list_types({ alias: "robotics" })
← [
{ "browseName": "MotionDeviceType", "kind": "ObjectType", "subtypeOf": "di:ComponentType", ... },
{ "browseName": "AxisType", "kind": "ObjectType", "subtypeOf": "di:ComponentType", ... },
...
] // 25 types
get_type_details
Get the full structure of a type — components, properties, methods, interfaces, including inherited members.
→ get_type_details({ alias: "robotics", browseName: "MotionDeviceType" })
← {
"browseName": "MotionDeviceType",
"kind": "ObjectType",
"subtypeOf": "di:ComponentType",
"interfaces": ["di:IVendorNameplateType", "di:ITagNameplateType"],
"components": [
{ "browseName": "robotics:Axes", "nodeClass": "Object", "typeDefinition": "FolderType", "modellingRule": "Mandatory" },
{ "browseName": "robotics:PowerTrains", "nodeClass": "Object", "typeDefinition": "FolderType", "modellingRule": "Mandatory" },
...
],
"properties": [
{ "browseName": "di:Manufacturer", "dataType": "LocalizedText", "modellingRule": "Mandatory" },
{ "browseName": "robotics:MotionDeviceCategory", "dataType": "MotionDeviceCategoryEnumeration", "modellingRule": "Mandatory" },
...
]
}
search_types
Search for types across all companion specs by keyword.
→ search_types({ query: "temperature" })
← [
{ "alias": "glass", "browseName": "MotorTemperatureTooHighEventType", ... },
{ "alias": "padim", "browseName": "TemperatureMeasurementVariableType", ... },
{ "alias": "amb", "browseName": "OverTemperatureConditionClassType", ... }
]
find_engineering_unit
Find the official UNECE Rec. 20 engineering unit symbol. Supports fuzzy matching and natural language aliases.
→ find_engineering_unit({ query: "celsius" })
← { "symbol": "°C", "matchType": "alias", "confidence": 1 }
→ find_engineering_unit({ query: "revolutions per minute" })
← { "symbol": "rpm", "matchType": "alias", "confidence": 1 }
→ find_engineering_unit({ query: "bar" })
← { "symbol": "bar", "matchType": "exact", "confidence": 1 }
Coverage
Companion Specifications (25)
| Alias | Specification | Types |
|---|---|---|
padim |
OPC UA for PA-DIM | 101 |
ijtBase |
OPC UA for IJT Base | 65 |
machineTool |
OPC UA for Machine Tools | 63 |
di |
OPC UA for Devices | 44 |
glass |
OPC UA for Glass Manufacturing | 36 |
machineVision |
OPC UA for Machine Vision | 36 |
adi |
OPC UA for Analyzer Devices | 35 |
commercialKitchenEquipment |
OPC UA for Commercial Kitchen Equipment | 35 |
robotics |
OPC UA for Robotics | 25 |
ia |
OPC UA for Industrial Automation | 20 |
amb |
OPC UA for AMB | 18 |
autoId |
OPC UA for AutoID | 18 |
metalForming |
OPC UA for Metal Forming | 16 |
machinery |
OPC UA for Machinery | 15 |
gds |
OPC UA GDS | 14 |
woodworking |
OPC UA for Woodworking | 13 |
cnc |
OPC UA for CNC Systems | 12 |
| …and 5 more | ||
| Total | 22 namespaces | 589 types |
Engineering Units
1,533 official UNECE Rec. 20 symbols plus 134 natural language aliases (e.g., "celsius" → °C, "revolutions per minute" → rpm).
How It Works
The server ships with a pre-generated catalog.json containing all type information extracted from OPC Foundation's official NodeSet2.xml files via node-opcua. All queries are answered from this static catalog — no network required, no API key needed.
┌─────────────────────────────────────────┐
│ node-opcua-modeler-mcp-server │
│ │
│ ┌─────────────────────────────────┐ │
│ │ catalog.json (1.7 MB) │ │
│ │ • 25 companion spec registries │ │
│ │ • 589 type summaries + details │ │
│ │ • 1,533 engineering units │ │
│ └─────────────────────────────────┘ │
│ │
│ 6 MCP tools → query the catalog │
│ stdio transport (JSON-RPC) │
└─────────────────────────────────────────┘
Use Cases
- AI-assisted OPC UA modeling — agents can discover types, resolve dependencies, and validate unit symbols before generating YAML/XML models
- Copilot integration — add OPC UA awareness to coding assistants
- Industrial digital twin design — explore companion spec type hierarchies interactively
- Learning OPC UA — ask an AI to explain types and their relationships
Requirements
- Node.js ≥ 18
Related
- node-opcua — Full OPC UA stack for Node.js
- OPC UA Modeler — Create, validate, and generate OPC UA information models
- Model Context Protocol — Open protocol for AI tool integration
License
MIT © Sterfive
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。