WebMCP, or Web Model Context Protocol, is an emerging JavaScript API specification that is revolutionizing the way web applications expose their functionalities to AI agents, browser assistants, and assistive technologies. Backed by industry giants like Google and Microsoft, WebMCP allows developers to register JavaScript functions with natural language descriptions and JSON schemas, effectively transforming web pages into Model Context Protocol servers on the client side.
WebMCP introduces two APIs to facilitate more reliable interactions: a Declarative API for standard HTML form actions and an Imperative API for more complex JavaScript-based interactions. These APIs allow AI agents to interact with websites without the need to scrape HTML or process screenshots, thus reducing resource consumption.
Google’s Chrome browser currently offers WebMCP as an early preview, showcasing its potential in real-world applications such as customer support ticket creation, e-commerce shopping flows, and travel booking. This standard not only makes it possible for agents to collaborate with users within the same interface, maintaining a shared context and user control, but also significantly reduces the cost of AI-agent interactions by minimizing token usage.
The W3C Web Machine Learning Community Group published a draft specification for WebMCP, marking a significant step towards a broader rollout expected by mid-year. Developers can already start experimenting with the potential that WebMCP brings, using the early preview in Chrome or leveraging the polyfill implementation, MCP-B.
In essence, by enabling structured tool exposure for AI agents, WebMCP is setting the stage for more sophisticated and seamless interactions between humans and AI in the web environment, making web applications truly AI-agent ready.

Collections

WebMCP (Web Model Context Protocol) is an emerging JavaScript API specification backed by Google and Microsoft that enables web applications to expose their functionality to AI agents in a structured way. It introduces two APIs—a Declarative API for HTML form actions and an Imperative API for complex JavaScript interactions—allowing AI agents to interact with websites without scraping HTML or processing screenshots. Chrome offers an early preview, and the W3C Web Machine Learning Community Group has published a draft specification. WebMCP reduces token usage and resource consumption while enabling agents to collaborate with users in a shared interface, with broader rollout expected mid-year. Developers can experiment via Chrome's early preview or the MCP-B polyfill.

Exploring WebMCP: A New Standard for AI Agent and Website Interaction

I fell like pretty soon you won’t need the website. If it’s just a tool for AI to get the information. Soon we’ll just be building branded MCP Servers for marketing. Why worry about responsiveness, UI design, etc. For some people this will be all they need.

What would be the motivation for a website provider to support this? Don’t websites lose revenue if you do not interact with the site but just with ai tools?

This is obviously the way things are headed, it will be fun to play around with though.

This is the first stage of the change of the world under AI.

Am I missing something? HTML is for humans, and APIs are for systems. Instead of creating WebMCP, why can’t the website owner simply create a simple MCP server?