The Help Scout MCP Server, developed by Drewburchfield, functions as a Model Context Protocol (MCP) server, bridging AI assistants with the Help Scout customer support platform. Built with TypeScript, this free server serves as an integration layer, allowing AI agents to interact directly with Help Scout's API. It is distributed as an npm package and a custom .mcpb binary for one-click installation, compatible with AI clients such as Claude, ChatGPT, and Cursor.

What does Help Scout MCP Server connect to?

The Help Scout MCP Server connects directly to Help Scout, a widely used customer support platform. It operates as a Model Context Protocol (MCP) server, specifically designed to act as an integration layer. This setup allows various AI assistants to interact with Help Scout's API, enabling automated and intelligent management of customer interactions. The server is confirmed to be compatible with leading AI clients, including Claude (Desktop, Cowork, Code), ChatGPT, and Cursor, extending its utility across different AI environments.

Core features and tools

The server exposes callable tools that empower AI agents with extensive capabilities. These include searching conversations by diverse criteria and looking up customer or organization details. AI can inspect conversation metadata, summaries, threads, and attachments, or pull full thread history. It also supports searching Help Scout Docs articles and retrieving various Help Scout reports and metadata. With opt-in writes, it enables actions like drafting replies, adding internal notes, applying tags, and changing conversation status. Granular tools include Update Conversation, Send Reply, Get Conversation Details, and Create Customer.

Who is it for?

This program is intended for Help Scout customers, particularly enterprise support teams, who seek to enhance their efficiency in managing customer conversations through AI. It supports workflows where AI agents search conversations and retrieve customer context to answer questions, assist with drafting replies and internal notes, and perform triage actions. Additionally, it aids in analyzing customer interactions by identifying patterns and pulling various Help Scout reports. Skip this if your organization does not use Help Scout for customer support.

How to get started

The server authenticates using a standard OAuth2 Client Credentials flow. Setup requires a Help Scout App ID and App Secret, provided as HELPSCOUT_CLIENT_ID and HELPSCOUT_CLIENT_SECRET environment variables.

Use cases

• Enabling AI agents to search conversations and retrieve customer context to answer questions directly within their chat or editor interface.
• Assisting with drafting replies and internal notes, and performing triage actions like tagging or changing conversation status (when write actions are opted in).
• Facilitating the analysis of customer interactions by spotting patterns across tickets and pulling various Help Scout reports.

The Help Scout MCP Server provides a critical integration layer for AI agents to interact with Help Scout. Developed by Drewburchfield, this free MCP server enhances customer support operations. It enables AI to manage conversations, assist with replies, and analyze interactions. The server is built with TypeScript and distributed via npm and a custom binary.