WordPress.com Now Lets AI Agents Build and Run Your Website
WordPress powers over 43% of all websites on the internet. On March 20, 2026, the hosted version at WordPress.com took a major step by allowing AI agents to draft, edit, and publish content directly on customer sites. The feature also supports managing comments, fixing metadata, and organizing categories, all through natural language commands. This is not a simple autocomplete. It is a full content operation layer powered by AI agents connected through the Model Context Protocol (MCP).
How AI Publishing on WordPress.com Works
- AI agents can create posts, landing pages, and About pages from natural language descriptions
- Agents can approve, reply to, and clean up comments across a site
- They can fix alt text, captions, and titles to improve SEO
- Users connect their preferred AI client (Claude, Cursor, ChatGPT) via MCP at wordpress.com/mcp
- All changes are tracked in the site’s Activity Log, and AI-written posts default to draft status
What MCP Brings to WordPress.com Publishing
This builds on WordPress.com’s MCP support, introduced in October 2025. MCP, or Model Context Protocol, is a standard that lets applications provide context to large language models. Previously, MCP only let AI assistants read site content, settings, and analytics. Now the connection is read-write.
Before creating content, the AI agent scans the site’s theme and design. It learns the colors, fonts, spacing, and block patterns already in use. That means generated pages should match an existing site’s look without manual formatting.
Users can author drafts for their AI agent to publish, tag, and categorize, or they can describe what they want and let the agent create it from scratch. All changes require user approval.
What This Means for the Web
WordPress.com’s hosted network sees 20 billion page views and 409 million unique visitors every month. Even though it represents a fraction of all WordPress installations, the scale is significant. Opening that network to AI-generated content at the publishing layer changes the math for content creation.
The barrier to launching and maintaining a website drops sharply. A small business owner could describe a blog post in a sentence and have it drafted, tagged, and ready for review in seconds. The tradeoff is obvious: more machine-generated content across the web.
Meta recently acquired Moltbook, a social network where AI agents posted and interacted with each other. Anthropic has experimented with AI-written blogs under human oversight. WordPress.com joining this trend makes AI publishing a mainstream tool, not just an experiment.
Practical Limits to Watch
WordPress.com built in guardrails. Posts written by AI default to draft status. Users toggle which capabilities to enable. The Activity Log tracks every change. These are reasonable safety nets, but they depend on the site owner actually reviewing what the AI produces.
For site owners who already struggle to find time for content, handing the keys to an AI agent is tempting. The risk is publishing content that looks polished but lacks depth, accuracy, or a real point of view. The tool is powerful. How it gets used will determine whether it raises the floor for web publishing or floods it with noise.
