unknown

WebMCP is a new Chrome Canary feature that enables AI agents to interact with websites through structured APIs. It supports two approaches: declarative (HTML attributes like tool-name and tool-description on form elements) and imperative (JavaScript execution with registered tools). The demo shows an AI agent controlling a kanban board and filling out a pizza order form by understanding available tools and their schemas, making autonomous decisions about DOM manipulation without traditional web scraping or excessive token usage.