The Chrome team has officially released the early preview of WebMCP in Chrome 126 Canary. Developed in collaboration between Google and Microsoft, this protocol aims to standardize how websites interact with AI agents.
What is WebMCP?
WebMCP is a proposed standard that allows websites to expose structured actions directly to AI. Instead of relying on fragile methods like HTML parsing or UI screenshots, sites can now publish a "tool contract." This enables AI agents to discover and call specific functions through a browser API—specifically navigator.modelContext.
Key Features and Benefits
- Reduced Fragility: Automations won't break when a website's UI or DOM changes.
- Lower Costs: Reduces the computational overhead required for visual-based AI processing.
- Implementation Options:
- Declarative API: For simple actions based on HTML forms.
- Imperative API: For complex interactions requiring client-side JavaScript.
- User-Centric: Unlike backend MCP, WebMCP focuses on the browser environment where the user is present. It is not intended for headless or fully autonomous automation.
Current Status
WebMCP is currently being incubated at the W3C. While a public timeline for other browsers is not yet available, developers can test the protocol by enabling the test flag in Chrome Canary and joining the early preview program.
