As AI continues to evolve, WordPress is exploring what it means to build smarter, more agent-friendly interfaces. Two emerging experimental projects, the WordPress Feature API and the WordPress MCP Project are at the heart of that conversation. While they sound similar and even work together, they serve different purposes.
I’m Elliott Richmond and in this post I’ll break down the difference between the Feature API and the WordPress MCP project, how they relate, and what roles they might play in the future of AI + WordPress.
What Is the WordPress Feature API?
The WordPress Feature API is a proposed system, currently in RFC (Request for Comments) stage that lets developers register WordPress features in a structured, machine-readable way. These “features” might include things like:
- Editing a post
- Searching for content
- Getting WooCommerce product info
- Navigating to a specific admin screen
Each feature is defined with:
- A name and description
- Input and output schemas (based on JSON Schema)
- A type (tool or resource)
- An optional callback function to run when executed
It’s designed to work inside WordPress, both server-side (PHP) and client-side (JavaScript). Once a feature is registered, it becomes discoverable and executable, not just by AI, but also by UIs like the command palette, custom interfaces, or automation scripts.
In essence, the Feature API gives WordPress a way to describe what it can do, so other systems (including AI) can understand and use those capabilities.
What Is the WordPress MCP Project
The WordPress MCP project is a separate, standalone project that acts as a bridge between the WordPress Feature API and the Model Context Protocol (MCP) an open protocol designed to let large language models (LLMs) discover and use external tools.
MCP defines how AI agents can:
- Discover what tools are available
- Understand what inputs they need
- Execute those tools safely and predictably
The WP MCP project takes the features defined by the Feature API and exposes them through a standard MCP-compliant interface, making them usable by LLMs like GPT-4, Claude, or custom AI agents.
How the WordPress Feature API and MCP Work Together
Here’s the key relationship:
- The Feature API defines what WordPress can do (internally).
- The MCP project takes those definitions and makes them usable externally by AI systems.
You can think of it like this:
Layer | Role |
Feature API | WordPress-native feature registry |
WP MCP | AI-facing wrapper using MCP format |
The WP MCP project depends on the Feature API. But the Feature API does not depend on MCP, it can be used by other things too (like UI command tools or workflows).
How This Might Be Used in the Future
In a future where AI assistants are common in WordPress dashboards, these two systems could work together to power:
- Context-aware chat interfaces that can perform real actions on your site
- LLM-powered admin tools that help you edit content, settings, or WooCommerce products
- Custom developer workflows where AI agents assist with block editing, debugging, or deployment
- Automation and orchestration, chaining features into workflows triggered by user commands or backend rules
And importantly, this system allows for safe, permission-based, and structured execution, far more reliable than natural language commands without context.
Watch the Deep Dive
If you’d prefer a visual walk-through, I’ve also created a YouTube video on my channel that explores the WordPress Feature API in detail, including what it is, how it works, and what kind of functionality it could unlock.
Watch the video here
Note: This video focuses on the Feature API only. I’ll cover the WordPress MCP project in a future video.
Final Thoughts
The WordPress Feature API and the WordPress MCP project aren’t here to “put AI in WordPress.” Instead, they’re building the infrastructure that lets WordPress and AI communicate, safely, clearly, and extensibly.
- The Feature API is the toolbox.
- The MCP Adapter is the translator.
- The AI agent is the user.
Together, they could define a whole new class of WordPress experiences.
If you’re a developer or curious about where this is going, check out the RFC for the Feature API and the MCP GitHub repo, and follow along. This is open source at its best, forward-thinking, collaborative, and experimental.
Leave a Reply