Automattic adds authoring capabilities to WordPress.com’s MCP integration, allowing AI agents like Claude and ChatGPT to create posts, create pages, manage comments, and restructure content.
For most of the last six months, connecting an AI agent to your WordPress.com site meant giving it a window. You can ask Claude or ChatGPT questions about your content, pull up site analytics, or check which posts haven’t been updated in a year. Useful, but mostly passive.
On Friday, Automattic added a door.
WordPress.com has enabled writing capabilities for its Model Context Protocol (MCP) integration allows AI agents to create and modify content directly on your site.
The update adds 19 new operations across six content types: posts, pages, comments, categories, tags, and media. From a single natural language query, an agent can create and publish a post, create a landing page using your theme’s block patterns, approve and reply to comments, rearrange category structures, or fix missing subtext in your entire media library.
The core architecture, MCP, an open protocol that standardizes how applications provide context to large language models, was first introduced on WordPress.com in October 2025. At the time, it was read-only: agents could query your site, but not touch it.
A second update in January 2026 added OAuth 2.1 authentication, making it easier to connect AI clients securely. In February, Automattic launched the official Claude Connector, which was again read-only at the time. Today’s writing capabilities are a step up for the platform.
The feature is clearly designed around human verification. Before creating, updating, or deleting something, the agent describes exactly what it plans to do and asks for approval. New posts have a draft status by default, giving users a chance to review anything before it’s published; changing a published post triggers a notification that the changes will be visible immediately.
Deleting posts, pages, comments, and media sends items to the trash, where they can be restored within 30 days. Categories and tags that WordPress does not trash will trigger an additional confirmation warning that the deletion is permanent. Each action is logged in the site’s Activity log.
User role permissions are fully enforced: Editor can create and edit posts, but cannot change site settings; A contributor can draft but not publish.
One of the more technically interesting aspects of the app is its topic awareness. Before creating a page or post, an agent can read the site’s design system, colors, fonts, spacing, block patterns, and create content that inherits these specifications.
Posting features are available on all WordPress.com paid plans today. Users activate the specific operations they want to allow on each site via the MCP dashboard at wordpress.com/me/mcp.
Compatible clients include Claude, ChatGPT, Cursor and any other MCP enabled tool. WordPress.com powers a significant portion of the internet, according to figures presented at Automattic’s Word of the State event in December 2025, with WordPress running more than 43% of all websites globally and a 60.5% share of the content management system market.
The scale at which AI agents with typing capabilities can now run on that infrastructure is immense.
The MCP ecosystem is expanding rapidly. The WordPress MCP Adapter, which provides similar functionality on self-hosted WordPress installations, is making its way into WordPress Core.
Automattic’s other products, including WooCommerce and Beeper, have their own MCP implementations. For example, standardized AI agent access becomes an architectural assumption rather than an experiment rather than a one-off integration into application functionality.
A practical question for WordPress.com users is trust. Allowing an AI agent to write to a production site is a different proposition than asking it to aggregate your traffic. Automattic has clearly leaned into this, making the validation model a central part of the announcement and changing the default configuration for each transaction.
