The Google Chrome team announced its early preview of WebMCP, "a standard way for exposing structured tools, ensuring AI agents can perform actions on your side with increased speed, reliability, and precision," André Cipriani Bandarra from Google wrote.
WebMCP basically gives you a way to communicate via your website to LLMs what actions certain buttons or links take on your website. WebMCP is designed to help AI agents interact with websites.
Currently, if an AI agent (like a browser assistant) wants to "book a flight" or "add an item to a cart" on a website, it often has to guess which buttons to click by looking at the pixels or HTML (screen scraping). This is error-prone.
WebMCP fixes this by allowing websites to explicitly publish a "Tool Contract." It uses a new browser API (navigator.modelContext). Instead of the AI guessing, the website provides a structured list of tools (e.g., function buyTicket(destination, date)). The AI can then "call" these functions directly.
André Cipriani Bandarra from Google explained, "By defining these tools, you tell agents how and where to interact with your site, whether it's booking a flight, filing a support ticket, or navigating complex data. This direct communication channel eliminates ambiguity and allows for faster, more robust agent workflows."
WebMCP proposes two new APIs that allow browser agents to take action on behalf of the user:
- Declarative API: Perform standard actions that can be defined directly in HTML forms.
- Imperative API: Perform complex, more dynamic interactions that require JavaScript execution.
- These APIs serve as a bridge, making your website "agent-ready" and enabling more reliable and performant agent workflows compared to raw DOM actuation.
Dan Petrovic called this the biggest shift in technical SEO since structured data. Glenn Gabe called this a big deal.
SEOs should 100% read up on this...
Forum discussion at X.
