Watch Q2 Community Update | Learn about MCP, AI Agents & v5.0 Early Access
Watch now 🚀

Access Your System’s Data and Trigger Scenarios via Remote MCP Server

Note: Coming Soon

The Remote MCP (Model Context Protocol) Server is a powerful and secure solution that enables AI-powered agents—like ChatGPT, Claude, or any other MCP-compatible chatbot—to communicate directly with your system’s data and even trigger automated scenarios. Unlike local implementations, the Remote MCP Server runs on Boost.space infrastructure, meaning there’s no need to host or manage the server yourself.

Once connected, you can interact with your system in real time, receive structured data, run queries, and even initiate workflows—all through a natural AI chat interface.

In this guide, we’ll walk you through three simple steps to get started:

  1. Connecting your AI agent to MCP
  2. Chatting with your AI agent, who can now access and work with your system’s data
  3. Taking action—triggering scenarios and launching automated workflows directly from the conversation

Let’s get started.

Step 1: Connect MCP to Your AI Agent (Example with ChatGPT)

To show how the Remote MCP Server works, let’s walk through a practical example using OpenAI’s ChatGPT interface with MCP support.

To connect your Boost.space system to an AI agent using the Remote MCP Server, follow this example with OpenAI’s Playground environment:

  1. First, go to https://platform.openai.com/playground.
  2. In the Prompts section, find the Tools column, click the plus (+) icon, and select MCP Server from the list.
  3. Once you’ve added the tool, you’ll need the following:


    MCP Token
    – A secure MCP token you generate in your Boost.space instance to authenticate the communication.URL – A dedicated connection URL, which you’ll find in the MCP Token section after creating the token.

    You’ll find out how to get your MCP token and URL in the step-by-step guide below.
    With these two components, the AI agent can access your data in a controlled and secure way. Simply paste the URL and your MCP Token into any chatbot interface that supports MCP (like OpenAI, Claude, or other tools), and start chatting with your data

  4. Tool Selection

    Once the AI agent is connected using the URL and MCP token, you’ll be asked to select which tools (i.e., specific functions or data modules) the agent can use during the conversation.This step determines what the AI will be able to do—for example:

    • Search and filter data from a module
    • Retrieve contact details or invoices
    • Summarize or analyze records
    • Trigger automation workflows

You can adjust these tools at any time to control the scope of access.

Step 2: Chat with Your AI Agent and Access System Data

Imagine you manage your sales records inside Boost.space. By connecting the Remote MCP Server to ChatGPT, you can type a question like:

“Show me all sales deals over $10,000 from the past quarter.”

The AI agent will use the MCP protocol to access your live data, filter it according to the query, and return a clean, structured answer—without writing a single line of SQL or leaving the chat window.

You can then follow up with:

“Which salesperson closed the most high-value deals?”

“Start the upsell analysis scenario.”

This transforms your AI chatbot into a fully conversational interface for working with your internal system.

Step 3: From Data to Action – Trigger Scenarios Directly from the Chat

In addition to working with data, the Remote MCP Server can also trigger scenarios—automated workflows built inside your Boost.space system.

To enable this, the scenario must be marked as “On Demand.” Only scenarios with this setting can be executed via AI chat interfaces using the MCP protocol.

Passing Inputs into the Scenario

When triggering a scenario via MCP, the AI needs to send the correct input values into the workflow—these are defined as Scenario Inputs in the scenario builder.

These inputs are created by you in the scenario editor

and mapped using the Scenario Inputs feature.

For example, let’s say you ask ChatGPT:

“Create the best offer for company X and send it via email.”

To complete this task, the scenario might require three inputs:

  • The email address
  • The subject line
  • The bullet points of the offer

Once defined, the AI can dynamically fill in those inputs during the conversation and pass them directly to the scenario at runtime.

This enables a whole new layer of automation—letting your AI agent not only read and summarize your system’s data but also take action by triggering predefined processes with context-aware inputs.

With just a few simple steps, you can turn any AI agent into a powerful interface for working with your system—securely accessing data, answering questions, and triggering automated workflows. Ready to explore what’s possible?

If you still need to clarify anything, please reach out to us at [email protected].