MCP integration
The MCP package (Robomotion.MCP, v0.6.0) adds 19 nodes to Robomotion flows, including Listen SSE, Listen Stdio and Listen HTTP.
It runs on Windows, macOS and Linux robots. For AI agents, 18 of its nodes work as tools, and its Toolkit node exposes 6 tools at once.
What the MCP package does
MCP integration for Robomotion RPA
Model Context Protocol (MCP) client and server package for Robomotion. Provides both client-side nodes for connecting to external MCP servers and server-side nodes for exposing Robomotion flows as MCP servers. Client nodes support three transport modes: Stdio (JSON config with command/args/env for local processes), SSE (Server-Sent Events via URL), and Streamable HTTP (with optional stateless mode and continuous listening). Client operations include calling tools, listing tools/prompts/resources, getting prompts, and reading resources. Listen HTTP serves Streamable HTTP with optional stateless mode (no session id, so any instance behind a load balancer can answer), a bearer token from an API Key vault item, a call timeout for slow RPA work, CORS origins for browser clients, a proxy-host escape from the DNS rebinding guard, and streaming disable. Server nodes listen via Stdio, SSE, or HTTP transports and expose tools, prompts, and resources through paired In/Out nodes - ToolIn defines a tool with a JSON schema and ToolOut returns results (an object result also travels as structuredContent, and is_error marks the call as failed); PromptIn defines a handlebars-templated prompt and PromptOut returns rendered content; ResourceIn defines static or URI-template resources and ResourceOut returns content. Credentials are optional: an API Key item can be bound to Listen HTTP as a bearer token. Connection details are configured per-node.
- Namespace
Robomotion.MCP- Latest version
- 0.6.0 (12 versions published)
- Written in
- Go
- Categories
- Artificial Intelligence
MCP nodes (19)
These are the node IDs an agentic flow or an RPA bot uses, exactly as they appear in the Flow Designer and in flow code. A trigger node has no input port and starts the flow when its event arrives. An agent tool node can also be called by an AI agent on its own. The Toolkit node hands an agent several of the package operations at once.
| Node | Node ID | Type |
|---|---|---|
| Listen SSE | Robomotion.MCP.Server.ListenSSE | TriggerAgent tool |
| Listen Stdio | Robomotion.MCP.Server.ListenStdio | TriggerAgent tool |
| Listen HTTP | Robomotion.MCP.Server.ListenHTTP | TriggerAgent tool |
| Prompt In | Robomotion.MCP.Server.PromptIn | Agent tool |
| Prompt Out | Robomotion.MCP.Server.PromptOut | Agent tool |
| Resource In | Robomotion.MCP.Server.ResourceIn | Agent tool |
| Resource Out | Robomotion.MCP.Server.ResourceOut | Agent tool |
| Tool In | Robomotion.MCP.Server.ToolIn | Agent tool |
| Tool Out | Robomotion.MCP.Server.ToolOut | Agent tool |
| Connect Stdio | Robomotion.MCP.Client.ConnectStdio | Agent tool |
| Connect SSE | Robomotion.MCP.Client.ConnectSSE | Agent tool |
| Connect HTTP | Robomotion.MCP.Client.ConnectHTTP | Agent tool |
| Call Tool | Robomotion.MCP.Client.CallTool | Agent tool |
| Get Prompt | Robomotion.MCP.Client.GetPrompt | Agent tool |
| List Prompts | Robomotion.MCP.Client.ListPrompts | Agent tool |
| List Resources | Robomotion.MCP.Client.ListResources | Agent tool |
| List Tools | Robomotion.MCP.Client.ListTools | Agent tool |
| Read Resource | Robomotion.MCP.Client.ReadResource | Agent tool |
| Toolkit | Robomotion.MCP.Agents.Toolkit | Toolkit · 6 tools |
Platforms
The MCP package runs on Windows, macOS and Linux robots. Any Robomotion robot on these operating systems can run a flow that uses it.
How to install the MCP package
- Open Packages in the Robomotion Flow Designer.
- Search for MCP or
Robomotion.MCP, then click Install on its card. - The MCP nodes appear in the node palette for everyone in the workspace, because installation is per workspace.
From a terminal, the Robomotion CLI installs the same package:
robomotion install Robomotion.MCPTemplates that use MCP
2 ready-made templates use MCP nodes in their flows. Open one to start from a working flow.
Clinic Support Assistant
A customer support assistant for a website, served to the page as one MCP tool. A visitor and a signed-in patient are answered by two different Hermes Agents, and the difference is not a prompt or a flag, it is two wires: the visitor's agent is wired to the practice documents and to nothing else, so nothing on that side of the canvas can reach a patient record whatever the model is asked. The signed-in tools take no arguments at all. They read the record number the website established from a session cookie, so the model cannot name a patient, and every turn is written to a dialog log the practice can read in the morning.
Tax Portal MCP Server
Serves an ordinary RPA automation to an AI assistant as MCP tools, so a model can ask a question of a system it could never sign into. One Listen HTTP node makes the flow a stateless MCP server; three Tool In nodes declare the contracts a model reads, with no arguments, one required argument and one optional. Behind them a sign in subflow answers a government portal's tax number, password and one-time code from the vault, and two more subflows read VAT returns, penalties and electronic invoices off the page. The caller is given a bearer token and nothing else, and Tool Out's Is Error flag lets the flow refuse a company it holds no mandate for instead of returning an empty answer the model will fill in for itself.
Related integrations
Packages that share a category or keywords with MCP. It is filed under Artificial Intelligence.
ADK AgentAn AI-powered automation agent that helps automate applications across different platforms
AI AgentsAn AI-powered automation agent that helps automate applications across different platforms
ClaudeProvides a simple interface to access the Anthropic Claude API. Claude is Anthropic's powerful AI assistant capable of advanced reasoning, coding, analysis, and creative tasks.
ElevenLabs AIIt connects to ElevenLabs AI. It is an natural-sounding speech synthesis and text-to-speech software, using artificial intelligence and deep learning
Google GeminiProvides a simple interface to access the Google Gemini API (Gemini 3.5 / 3.1, Nano Banana image models, and more), plus an agent Toolkit for AI-driven image generation.
Hermes AgentSelf-improving Hermes Agent (Nous Research) with skills, sessions, MCP, and visual tool wiring
MCP documentation
The MCP package reference on docs.robomotion.io documents each node and its properties.
https://docs.robomotion.io/reference/packages/mcp/Package data from the Robomotion package index, synced October 31, 2026.