Overview

Content Hub has a built-in Model Context Protocol (MCP) server. Once connected, an AI assistant like Claude can read and act on your Content Hub data directly: checking sprint progress, pulling GEO citation data, auditing content freshness, or creating a new sprint, all from a chat, without you switching tabs or writing a single API call.

What you can do with it

Once connected, you can ask things like:

• "List my sprints in ContentHub"
• "Show me the GEO dashboard"
• "Which topics are stale, older than 6 months?"
• "Create a sprint called Q3 Launch from Aug 1 to Aug 14"

The assistant calls the same underlying tools your team uses in the product itself: sprints, topics, GEO citations, performance reports, dashboard metrics, coach insights, integrations, and settings are all available as individual, purpose-built actions rather than one generic "run a query" tool. See Available Tools for the full catalog.

How it's designed to behave

• Nothing changes without your say-so. Any action that creates, updates, or deletes data shows you a preview first and waits for explicit confirmation: see Security & Access.
• Scoped to your company. Every request is tied to your signed-in account; you only ever see and modify your own company's data.
• Read-heavy, not a database backdoor. There's no arbitrary query tool: every action maps to a specific, well-defined piece of functionality, the same way the product's own UI does.
• No surprise automation. The MCP server doesn't run scheduled jobs, trigger scrapes/crawls, or expose admin operations. It only responds to what you explicitly ask for in the moment.

Getting started

You'll need a Content Hub account and an MCP-compatible client. Connect Claude Desktop walks through the full setup: it takes about two minutes.