The newly introduced standard by Google and Microsoft called WebMCP
February 11, 2026•9 min read min read

WebMCP: The New Browser Standard for AI Agents

Your website doesn't speak AI? Now it does. Google and Microsoft have jointly introduced a new web standard that fundamentally changes the relationship between websites and AI agents: WebMCP. The early preview has been available in Chrome Canary 146 since February 10, 2026 — and the implications for digital business models are enormous.

Until now, AI agents had to read websites like a human: analyzing HTML code, clicking through forms, and hoping the layout wouldn't change. That was slow, error-prone, and expensive. WebMCP puts an end to that. Websites can now expose structured tools directly to AI agents — cleanly, reliably, and in a standardized way.

This article explains what this means for your business and how to position yourself now.

Contents

  1. Why DOM Scraping Is a Dead End

  2. WebMCP Explained: Two APIs for Two Use Cases

  3. The Imperative API in Detail

  4. The Declarative API: AI-Ready Forms in HTML

  5. Human or Machine? New Events and CSS Classes

  6. Limitations: What WebMCP Can't Do (Yet)

  7. Recommendations for SMBs

  8. Conclusion: The Browser Becomes an AI Interface

Why DOM Scraping Is a Dead End

When an AI agent needs to book a flight, fill out a form, or compare prices today, it typically does so via DOM scraping: it reads the website's HTML code, identifies buttons and input fields by their CSS classes or IDs, and simulates clicks. It works — until the website owner changes the design.

The result: Fragile automations that break regularly. Companies that rely on web scraping know the problem all too well. A single website update can bring entire processes to a halt. According to a Gartner study, over 40% of all RPA (Robotic Process Automation) projects fail due to unstable interfaces — and DOM scraping is one of the primary causes.

WebMCP solves this problem at its root: instead of AI agents scraping the interface, the website actively offers its functionality as structured tools. The agent doesn't have to guess where the booking button is — it simply calls the search_flights tool.

WebMCP Explained: Two APIs for Two Use Cases

WebMCP is a proposal by the W3C Web Machine Learning Working Group, backed by Google and Microsoft. The standard defines two complementary interfaces:

  1. Imperative API — for developers who need full control over tool logic. They register JavaScript functions with clearly defined input schemas.

  2. Declarative API — for anyone who wants to make existing HTML forms AI-ready. Two HTML attributes are all it takes for the browser to automatically generate a tool schema.

Both approaches use the same underlying principle: the website tells the browser which actions are available and which parameters they expect. The AI agent in the browser can then call these tools directly {EM} without parsing the DOM.

The Imperative API in Detail

The Imperative API is designed for developers who want to expose complex business logic as AI tools. The central object is navigator.modelContext, which is used to register tools.

Practical example: flight search as an AI tool

navigator.modelContext.registerTool({
  name: "search_flights",
  description: "Search for available flights",
  inputSchema: {
    type: "object",
    properties: {
      origin: { type: "string" },
      destination: { type: "string" },
      date: { type: "string" }
    },
    required: ["origin", "destination", "date"]
  },
  execute: async (params) => {
    const results = await searchFlightsAPI(params);
    return { flights: results };
  }
});

Here's what happens step by step:

  • name and description tell the AI agent what the tool does. The agent uses this description to decide whether to invoke the tool.

  • inputSchema defines the expected parameters in JSON Schema format. This tells the agent exactly what data it needs to provide: origin, destination, and date.

  • execute is the actual function that runs when the agent calls the tool. In this case, it calls an internal API and returns the results in a structured format.

The key advantage: The website retains full control. Business logic (pricing, availability checks, validation) stays server-side. The AI agent only sees what you want it to see.

For advanced scenarios, the API also offers:

  • navigator.modelContext.provideContext() {EM} replaces the entire toolset (e.g., when navigating between page sections)

  • navigator.modelContext.clearContext() {EM} removes all registered tools (e.g., on logout)

The Declarative API: AI-Ready Forms in HTML

Not every company has a development team that writes JavaScript tools. For this scenario, WebMCP offers the Declarative API: existing HTML forms become AI-ready with just two attributes.

Practical example: booking a consultation

<form toolname="book_consultation"
      tooldescription="Book a consultation appointment"
      toolautosubmit>
  <input name="name" type="text" required />
  <input name="email" type="email" required />
  <select name="service">
    <option value="strategy">KI-Strategie</option>
    <option value="chatbot">Chatbot</option>
  </select>
  <button type="submit">Absenden</button>
</form>

Here's how it works:

  • toolname gives the tool a unique name that the AI agent can reference.

  • tooldescription describes what the form does — the agent uses this description for its decision-making.

  • toolautosubmit allows the agent to automatically submit the form without waiting for human confirmation.

  • The browser generates the input schema automatically from the form fields: name (text, required), email (email, required), and service (select).

This is the fastest way to make an existing website AI-ready. You change two attributes in your HTML — the rest happens automatically. No JavaScript, no build process, no new infrastructure.

Human or Machine? New Events and CSS Classes

A common challenge with AI interactions: how does the website distinguish whether a human or an AI agent submitted a form? WebMCP solves this elegantly:

  • SubmitEvent.agentInvoked — a boolean that indicates on every form submission whether the action was triggered by an AI agent. This lets you add extra validations or route agent requests to a separate system.

  • SubmitEvent.respondWith(Promise) — enables a structured response back to the agent. Instead of an HTML page, the agent receives JSON data.

Additionally, WebMCP provides CSS pseudo-classes for visual feedback:

  • :tool-form-active — activates when an AI agent is currently interacting with a form

  • :tool-submit-active — activates while the agent is performing a submission

This lets you show users in real time that an AI is performing an action — for example, with a blue border around the form or a status message.

Developers also have access to the toolactivated and toolcancel events, which signal when an agent activates a tool or cancels the operation.

Limitations: What WebMCP Can't Do (Yet)

WebMCP has been available as an early preview in Chrome Canary 146 since February 10, 2026. This means: The standard is functional but not yet production-ready. Here are some limitations you should be aware of:

  1. No headless mode: The browser tab must be visible and active. Background automations without an open browser are not (yet) possible.

  2. No discovery mechanism: There's no standard for an AI agent to automatically discover which websites offer WebMCP tools. The agent needs to already know and visit the page.

  3. UI must stay in sync: The website's visual interface must reflect the current state of tool interactions. This requires careful development.

  4. Chrome Canary only: Currently, WebMCP is only available in Google's preview browser. Microsoft Edge will follow (Kyle Pflug from Microsoft Edge confirmed support in an interview with The New Stack), but broad browser support is still pending.

These limitations are typical for an early preview. The crucial point: Google and Microsoft are jointly backing the standard. In the browser world, that's a powerful signal. When both Chromium-based browsers (Chrome and Edge) support WebMCP, that already covers over 80% of the desktop market.

Recommendations for SMBs

WebMCP isn't a production feature yet — but smart companies are preparing now. Here are concrete next steps:

  1. Audit your forms and interfaces: List all forms on your website that customers or partners use: contact forms, booking systems, product configurators, calculators. These are your potential WebMCP tools.

  2. Pilot project with the Declarative API: Choose a simple form (e.g., appointment booking or quote request) and add toolname and tooldescription. That's 30 minutes of effort {EM} with potentially massive impact.

  3. Think API-first: If you're modernizing your website or online store anyway, plan for clean APIs. WebMCP tools are essentially API calls with browser integration. Those who build good APIs today will have ready-made AI tools tomorrow.

  4. Get your team on board: Share this article with your development team and product managers. WebMCP isn't some abstract future concept {EM} it's testable in Chrome Canary right now.

  5. Watch the competition: Over the coming months, check whether your competitors or industry platforms are offering WebMCP tools. First-mover advantages are particularly strong in the AI economy.

Rule of thumb: If your website has forms that customers fill out, WebMCP is relevant to you. The sooner you engage with it, the better positioned you'll be when the standard hits stable browser releases.

Conclusion: The Browser Becomes an AI Interface

WebMCP marks a turning point: The browser is evolving from a passive display device into an active AI interface. Websites are no longer just visited by humans — they're also used by AI agents, in a standardized and reliable way.

For SMBs, this means: making your digital presence AI-ready today secures access to a new channel. AI agents will increasingly prepare purchasing decisions, compare offers, and complete bookings. Companies whose websites can serve these agents have a clear competitive advantage.

The standard is coming — the only question is whether you're ready.

Want to make your website ready for AI agents? At hypescale, we advise SMBs on AI strategy and help with implementation {EM} from the initial audit to a production-ready WebMCP setup. Get in touch.