Introduction: Core Value of the node-red-contrib-mcp Project

The node-red-contrib-mcp project integrates Anthropic's MCP (Model Context Protocol) into the Node-RED low-code platform, allowing industrial automation and IoT developers to connect AI agents with external tools via drag-and-drop, realizing code-free intelligent workflow orchestration. This project links the Node-RED ecosystem (over 4 million installations) with MCP servers (over 10,000 available), bringing new possibilities for low-code AI automation in Industry 4.0, smart manufacturing, and IoT applications.

Project Background and Overview

With the rapid development of Large Language Model (LLM) capabilities, how to seamlessly integrate AI into industrial automation and IoT systems has become a problem for engineers. Node-RED, as a globally popular low-code IoT development platform, has over 4 million installations and is widely used in industrial automation, smart home, and other fields. MCP is an open standard launched by Anthropic for connecting AI with external tools/data sources, with over 10,000 MCP servers available. The node-red-contrib-mcp project combines the two, enabling developers to build AI Agent workflows in the Node-RED visual interface without code, empowering scenarios like Industry 4.0.

Introduction to the MCP Protocol

Model Context Protocol (MCP) is an open standard protocol that defines the communication method between AI models and external tools/data sources. Through MCP, AI agents can dynamically discover and call external functions (such as querying databases, reading files), and the standardized interface avoids writing separate integration code for each tool. Traditional AI tool calls require hard-coding details, while MCP abstracts the process through a unified protocol, improving interoperability and development efficiency. node-red-contrib-mcp brings this capability into the Node-RED ecosystem.

Core Features and Node Introduction

The project provides a series of Node-RED nodes covering server configuration to AI Agent orchestration:

Server and Configuration Nodes

• mcp server: Configure MCP server connections, supporting Streamable HTTP/SSE protocols and authentication;
• llm config: Configure LLM providers (OpenAI, Anthropic, Ollama, and other services compatible with the OpenAI API).

Tool Call Nodes

• mcp tool: Call any MCP tool, with parameters passed via msg.payload;
• mcp tools: List available tools on the MCP server;
• mcp resource: Read resources exposed by the MCP server.

LLM Call and AI Agent Nodes

• llm call: Call OpenAI-compatible LLMs, supporting system prompts, JSON mode, and multi-turn conversations;
• ai agent: Core node that implements the agent reasoning loop (analyze problem → call tool → get result → iterate → generate answer).

Typical Application Scenarios

Manufacturing OEE Monitoring

Build an intelligent agent to automatically answer the reasons for OEE decline: call the get_oee tool to obtain data, call the get_downtime_events tool to get downtime events, and conduct comprehensive analysis to draw conclusions (e.g., OEE decline caused by bearing failure, tool change delay, etc.).

IIoT Intelligent Alarm Handling

Build an MQTT → AI Agent → MQTT pipeline: When a machine alarm occurs, the agent investigates the cause via MCP tools and recommends measures, then sends suggestions via MQTT to achieve unattended intervention.

Smart Building and Energy Management

Combine with BACnet/Modbus protocol nodes, the AI agent analyzes energy usage patterns, automatically adjusts HVAC settings, optimizing energy consumption while maintaining comfort.

Technical Features and Ecosystem Compatibility

Technical Features

• Openness and flexibility: No lock-in to LLM providers or MCP servers; free to choose;
• Open source and localization: Apache-2.0 license, no cloud dependency, supports local operation, meeting industrial data privacy requirements;
• Production-ready: Complete error handling, status indicators, timeout mechanisms, natively integrated with the Node-RED system.

Compatibility

• MCP servers: Supports servers with Streamable HTTP/SSE protocols (e.g., OpenShopFloor, Anthropic official servers, custom servers);
• LLM support: Compatible with all OpenAI API services (OpenAI, Ollama, Azure OpenAI, etc.), Anthropic Claude can be called via LiteLLM proxy.

Quick Start Guide

Installation Methods

• Via npm: Execute npm install node-red-contrib-mcp in the Node-RED user directory;
• Via Palette Manager: Search for "node-red-contrib-mcp" in the Node-RED interface to install.

Quick Experience

After installation, import the JSON examples provided by the project to quickly experience MCP tool calls and AI Agent functions.

Documentation and Examples

The project provides detailed documentation and multiple examples (simple tool calls, complete AI Agent, MQTT-integrated IIoT, etc.) to help developers build custom workflows.

Project Summary and Value

node-red-contrib-mcp brings AI agent capabilities to the industrial automation and IoT fields, allowing engineers to build intelligent applications in the Node-RED environment without deep knowledge of Python or AI frameworks. It connects 4 million Node-RED users with over 10,000 MCP servers, opening a new chapter in low-code AI automation, and is an extremely valuable tool for exploring AI applications in industrial scenarios.