About

Do you want to streamline your editorial processes? Do you want to create, edit and publish content with AI support, trigger workflows, or simply get information about your CoreMedia content repository — all from within the AI assistant you already use, such as Claude, ChatGPT or Gemini, whether on desktop or mobile? Then the adesso CoreMedia MCP Server is the missing piece that connects your familiar AI assistant to your CoreMedia installation.

The adesso CoreMedia MCP Server turns your CoreMedia CMS into a content platform that can be operated by external AI assistants, in particular, occasional content editors can orchestrate their content with great ease. Deployed as a native Studio plugin, it exposes 51 production-ready tools over the Model Context Protocol (MCP) — the open standard that lets AI assistants, or any other MCP-capable client, talk to external systems. This allows AI assistants to read, author, publish, drive through workflow and analyze CoreMedia content — all using the same permissions and the same editorial processes your team already trusts.

There is no separate service, no data replication and no parallel permission model. Every call runs in the context of the respective authenticated user. It follows that: if an editor can't do it in Studio, the AI can't do it either!

The tool catalog of the adesso CoreMedia MCP Server covers the full editorial lifecycle:

• Discovery & Read (10 tools) – from a single ID lookup to a structured, multi-clause query across an entire site, including full-text search, taxonomy browsing and structured content export.
• Authoring & Write (18 tools) – the full editorial write surface: checkout/checkin discipline, page and section building, link-list manipulation, taxonomy tagging, image ingestion and batch edits. Every one of these tools supports a preview run (dryRun).
• Publication & Lifecycle (4 tools) – publish, take offline, check status and schedule publication dates, with dependent content published automatically alongside.
• Workflow (5 tools) – AI assistants can start, monitor, approve or reject editorial approval processes instead of routing around them.
• Governance & Insight (13 tools) – content health score, broken-link detection, orphaned content, pre-deletion impact analysis, storage and permission analysis, stale-content detection and more: answers to the questions nobody has time to answer manually.
• Diagnostics (1 tool) – a simple connectivity and identity check.

Safety is built in from the start: 24 of 26 mutating tools support a dryRun preview before anything is actually changed. Destructive operations such as content deletion require explicit confirmation (confirm). And because every tool runs in the context of the authenticated user, the AI can never do more than that person could already do in Studio.

The MCP Server can be deployed both self-hosted and on Content Cloud installations.

Here are some example requests from everyday editorial work and their results:

• "Find all articles under /Sites/EN/Campaigns/Summer that are missing SEO metadata (title, description), draft suitable text and add it." → The AI finds the gaps, writes the title and description, and saves them directly to the content — editors only need to proofread instead of writing it themselves.
• "Is it safe to delete the old 'Summer-Sale' teaser, or is other content still attached to it?" → Impact and reference analysis provide a reliable answer in advance — cleaning up without unpleasant surprises.
• "Show me all articles in the guides section that haven't been updated in over 6 months, prioritized by importance." → Instead of manually combing through the content inventory, a ready-made prioritized list for the next update round is generated on request.
• "Create a new section for the autumn campaign with three landing pages from the campaign template, link the matching product images, and schedule publication for October 1st." → The entire campaign scaffold is built in one conversation instead of half a day of clicking — editors only need to review the finished result.

Installation

The MCP Server is delivered on a project-specific basis as a CoreMedia plugin and runs natively inside Studio Server — no sidecar, no proxy, no additional infrastructure.

Requirements

• CoreMedia 11/12/13
• Java and Spring Boot version, that are already part of the Studio Server runtime
• An MCP-capable AI assistant (e.g. Claude Desktop/Claude Code, OpenAI GPT-series, Google Gemini, a custom MCP client, or any other MCP-compatible tool)

Support

The plugin we provide is a production-ready foundation for exploring — and productively using — AI-driven control of CoreMedia via the Model Context Protocol (MCP) within your system landscape.

We welcome your feedback on further use cases and future developments! If you would like to adapt the MCP Server to your editorial processes, your security and permission requirements, or your individual CoreMedia configuration, we would be very happy to discuss this with you. Feel free to contact us at coremedia-support@adesso.de.