Zesty.io MCP. Manage CMS models and items from natural language.

This MCP server connects your AI client directly to Zesty.io's API, giving your agent full control over content models, data entries, and site instances. You treat it like you're logged into the CMS yourself; your agent reads, writes, and updates structured content—think product specs or blog posts—using natural language commands without ever touching a web interface.

If you're building anything that needs to interact with enterprise-level digital assets, this is what you use. It’s developer-grade API wrapper designed for managing headless CMS data.

When your agent starts up, it first checks which sites are connected. You can list every Zesty.io instance associated with your account using list_zesty_instances. After that, if you need to know what kind of content you're dealing with, you run list_content_models to get a rundown of all available content structures and their unique Model ZUIDs.

For site-specific technical details, the agent can pull the full configuration settings for your entire website instance via get_instance_settings. This tells you exactly how that specific property is set up.

Reading data is straightforward. You don't have to guess what content exists; you first query all entries for a specific model and site using list_content_items, which returns a full list of records. If you know the exact ID, you can get every single field detail for that one item directly with get_content_item.

Writing new stuff is just as easy. When you need to create an entirely new content piece, you pass in all the necessary field values inside a JSON object, and the agent builds it out using create_content_item. If something changes—say, you correct text or update a date on an existing post—you modify that data using update_content_item.

And if an old piece of content is garbage, you wipe it clean from the system with delete_content_item by referencing its unique ID.

Your agent handles this whole process natively. For example, to update a title, your client first gets the existing data using get_content_item, then passes the new text into update_content_item. It's a clean loop that keeps all your structured content accurate and current across all connected sites.

This setup means you don't need manual logins or clicking through menus. You just tell your agent what to do—like, 'Change the main image on the Q3 report page to this new file,'—and it executes the sequence of API calls required to make it happen. It handles all that messy backend work for you.

It's built around specific actions: getting a list of sites (list_zesty_instances), mapping models (list_content_models), reading single items (get_content_item), listing batches (list_content_items), creating new content (create_content_item), modifying old entries (update_content_item), and finally, deleting records (delete_content_item). This combination gives you total control over your CMS data structure.