WebMCP

Websites usually show buttons for people. WebMCP adds explicit action descriptions: what they do, which inputs they need and what they return. Think of it as an instruction menu for an assistant, not a marketing app you need to buy.

Hypothetical example: a clinic describes ‘Check availability’ as an action taking a date and returning time slots. Checking must not be confused with confirming a booking: each action needs a clear scope.

Chrome documents two ways to register actions: form attributes or JavaScript. For marketing, a useful first step is a brief: action, inputs, outcome and limits. Your technical team must assess compatibility and implementation.

WebMCP and Lighthouse’s agentic browsing category are experimental. They are not universal website requirements or guarantees of SEO, citations or sales. Registering more actions is not always helpful: Chrome’s informational audit warns when there are more than 40 tools.

Our editorial recommendation is to start with a low-risk enquiry in a test environment. Keep identity and permission checks, and separate enquiries from actions that create commitments or send personal information.