About the Attentive MCP Server
Status: Beta. This document is a first draft of external, customer-facing developer documentation. Availability, tool names, and behavior may change during the beta.
Last updated: 8/13/2026
The Attentive MCP server lets you connect your own AI assistant — such as Claude or ChatGPT — directly to Attentive. Once connected, your assistant can pull reporting, draft and edit campaigns, build journeys and segments, and retrieve existing configuration, all scoped to what you can already do in the Attentive UI.
The MCP server is not itself an AI. It is a governed set of tools that your AI assistant calls. You decide how to combine those tools, from answering a one-off question to running a multi-step workflow.
What you will learn
- What the Model Context Protocol (MCP) is and how Attentive uses it
- How access and permissions work
- How to connect the Attentive MCP server to Claude Code, ChatGPT Codex, or another MCP client
- The configuration options available at connection time
- The full set of tools the server exposes today
- Example prompts, security guidance, limitations, and troubleshooting
What is MCP?
The Model Context Protocol (MCP) is an open standard that lets AI applications connect to external tools and data sources in a structured, governed way.
There are two roles:
- MCP client — the AI application you already use (for example, Claude Code or ChatGPT Codex). This is where the model runs.
- MCP server — a service that exposes a specific, controlled set of tools. The Attentive MCP server is one of these.
An MCP server is similar to an API, but instead of exposing every capability, it defines a specific set of governed tools that control what the connected assistant can see and do. The Attentive MCP server does not reason, decide, or act on its own. The intelligence comes from your own AI client; Attentive provides the tools it can call.
Updated about 17 hours ago