Is your WordPress site visible to AI agents?
AI assistants like ChatGPT, Claude, and Gemini don't browse websites the way humans do. They can't execute JavaScript, click buttons, or parse prices from HTML. Enter your domain to see exactly what they read — and what they miss.
Why your WordPress site is invisible to AI agents
When someone asks ChatGPT, Claude, or Gemini to find a product or look up information on a website, the assistant does not open a browser. It connects to structured data sources — APIs, MCP servers, or machine-readable documents — and reads from those directly. If your site does not expose this kind of interface, the agent has nothing to work with. It may attempt to scrape your HTML, but without JavaScript execution it will retrieve an incomplete, unstructured document that contains no actionable product data, no prices, no stock levels, and no purchase path.
This is not a future problem. AI assistants are already being used to shop, research, and compare products. Sites that are not agent-readable are already missing this traffic. The gap will only grow as AI-assisted search becomes the default entry point for product discovery.
What the Model Context Protocol (MCP) is
MCP is an open standard released by Anthropic in 2024. It defines a common interface that AI agents use to connect to external tools and data sources. Think of it as a universal connector: any MCP-compatible agent can discover what tools a server exposes, call those tools with structured inputs, and receive structured outputs. No custom integration per agent, no scraping, no ambiguity.
KaliCart Bridge and KaliCart MCP both implement this standard on WordPress. Once installed, your site becomes natively connectable by any MCP-compatible AI client — including Claude, ChatGPT, Cursor, and a growing list of AI tools that adopt the protocol.
How this scanner works
The scanner performs a server-side probe on the domain you enter. It fetches the homepage and captures the raw HTML — the same incomplete view an agent gets without structured data support. It then checks for the KaliCart Bridge discovery document, the catalog meta endpoint, a sample structured product, and the KaliCart MCP plugin endpoint.
The result is a readability score from 0 to 100. A score of 0 means agents see only raw HTML with no actionable information. A score near 100 means your site exposes a discovery document, a full catalog API with live prices and stock, shipping policy data, purchase readiness signals, and an MCP server — everything an agent needs to serve your products to a user.
Frequently asked questions
Can AI agents like ChatGPT read my WooCommerce store?
Not by default. AI agents cannot execute JavaScript or parse product data from HTML. They need a structured API like the one provided by KaliCart Bridge — a free WooCommerce plugin that exposes your catalog as a machine-readable REST API and MCP server.
What is KaliCart Bridge?
KaliCart Bridge is a free WordPress plugin that makes your WooCommerce product catalog readable by AI agents. It exposes prices, stock levels, variants, and shipping data as a structured REST API and MCP server — no JavaScript required, no external accounts, no API keys.
What is the KaliCart MCP plugin?
KaliCart MCP is a free WordPress plugin that exposes your posts, pages, and custom post types as a Model Context Protocol server. AI assistants can browse your site structure, search content, and read articles in clean Markdown — entirely on your server, no cloud dependency, no external calls.
How do I make my WordPress site visible to AI agents?
Install KaliCart Bridge if you have WooCommerce products, or KaliCart MCP if you publish blog posts and pages. Both are free plugins on WordPress.org. If you have both WooCommerce and editorial content, install both for full coverage.
What is an AI agent readability score?
The score measures how completely an AI agent can understand and act on your site. It checks seven signals: site reachability, discovery endpoint, catalog meta, structured product data, purchase readiness, shipping policy, and MCP endpoint. Each signal contributes points to a total out of 100.