Skip to main content

MCP (Model Context Protocol)

The MCP package provides comprehensive integration with the Model Context Protocol, enabling seamless communication between AI applications and external tools, resources, and prompts. This package allows you to build both MCP servers and clients within your Robomotion automation flows.

What is MCP?​

The Model Context Protocol (MCP) is an open standard that enables AI applications to securely connect to data sources and tools. It provides a unified way for AI systems to:

  • Access dynamic data through Resources
  • Execute operations via Tools
  • Generate structured outputs using Prompts

Package Capabilities​

The MCP package is divided into two main components:

Server Nodes​

Create MCP servers that expose your automation capabilities to AI clients:

  • Listen - Start an MCP server that listens for incoming connections
  • Prompt In/Out - Define and return prompt templates with parameters
  • Resource In/Out - Expose resources (files, data, documents) that clients can read
  • Tool In/Out - Define and execute tools that clients can call

Client Nodes​

Connect to and interact with MCP servers:

  • Connect - Establish connections to MCP servers
  • Call Tool - Execute tools on connected MCP servers
  • Get Prompt - Retrieve and render prompts from servers
  • List Prompts - Discover available prompts on a server
  • List Resources - Discover available resources on a server
  • List Tools - Discover available tools on a server
  • Read Resource - Read content from server resources

Use Cases​

Building MCP Servers​

  • Expose automation workflows as AI-callable tools
  • Provide dynamic data sources for AI applications
  • Create reusable prompt templates for AI assistants
  • Integrate legacy systems with modern AI platforms

Consuming MCP Services​

  • Connect AI workflows to external MCP servers
  • Access data from MCP-enabled applications
  • Leverage third-party MCP tools in automations
  • Build AI agents that interact with multiple MCP servers

Getting Started​

Creating an MCP Server​

  1. Add a Listen node to start your MCP server
  2. Connect Tool In/Out, Prompt In/Out, or Resource In/Out nodes to define capabilities
  3. Configure the SSE server settings (IP, port, endpoint)
  4. Your MCP server is now ready to accept client connections

Connecting to an MCP Server​

  1. Add a Connect node with your server configuration
  2. Use the returned Client ID with List Tools, Call Tool, Get Prompt, or Read Resource nodes
  3. Process the results in your automation flow

Architecture​

The MCP package uses a server-sent events (SSE) transport layer for client-server communication. Servers expose three types of capabilities:

  1. Tools - Executable functions with JSON Schema-defined parameters
  2. Prompts - Template-based text generation with parameter support
  3. Resources - Static or dynamic content accessible via URI

Requirements​

  • Robomotion RPA platform
  • Network connectivity for SSE communication (server mode)
  • Valid MCP server configuration (client mode)