Draft:WebMCP

Review waiting, please be patient.

This may take 3 months or more, since drafts are reviewed in no specific order. There are 4,805 pending submissions waiting for review.

If the submission is accepted, then this page will be moved into the article space.
If the submission is declined, then the reason will be posted here.
In the meantime, you can continue to improve this submission by editing normally.

Where to get help

If you need help editing or submitting your draft, please ask us a question at the AfC Help Desk or get live help from experienced editors. These venues are only for help with editing and the submission process, not to get reviews.
If you need feedback on your draft, or if the review is taking a lot of time, you can try asking for help on the talk page of a relevant WikiProject. Some WikiProjects are more active than others so a speedy reply is not guaranteed.

How to improve a draft

Wikipedia:Contributing to Wikipedia – a basic overview on how to edit Wikipedia.
Help:Wikitext – how to use the markup
Help:Referencing for beginners – how to include references
Wikipedia:Article development – how to develop your article
Wikipedia:Writing better articles – how to improve your article
Wikipedia:Verifiability – make sure your article includes reliable third-party sources

You can also browse Wikipedia:Featured articles and Wikipedia:Good articles to find examples of Wikipedia's best writing on topics similar to your proposed article.

Improving your odds of a speedy review

To improve your odds of a faster review, tag your draft with relevant WikiProject tags using the button below. This will let reviewers know a new draft has been submitted in their area of interest. For instance, if you wrote about a female astronomer, you would want to add the Biography, Astronomy, and Women scientists tags.

Add tags to your draft

Editor resources

Easy tools: Citation bot (help) | Advanced: Fix bare URLs

Reviewer tools

Instructions · What links here · WebMCP (talk: + · bio) · (log) · Copyvios report · reFill · Citation Bot · (Search: Google, Wikipedia) · Submitted 30 days ago by Idan nekuda (talk: D · +) · Last edited 29 days ago by Idan nekuda

Comment: If the sources are accessible, please link them. Greenman (talk) 22:55, 21 June 2026 (UTC)

The Web Model Context Protocol (WebMCP) is an open web standard and browser API that allows web applications to expose their functionality as structured tools to artificial intelligence (AI) agents. Proposed jointly by engineers from Google and Microsoft in 2025, the standard is incubated through the World Wide Web Consortium (W3C) Web Machine Learning Community Group.^[1] WebMCP serves as a client-side, browser-native complement to the broader Model Context Protocol (MCP).^[2]

Background

Before the introduction of WebMCP, AI agents interacting with web pages primarily relied on DOM scraping and simulated actuation — reading raw HTML, guessing the purpose of buttons or input fields, and synthesizing mouse clicks or keystrokes.^[3] This approach was prone to breakage when a site's layout or markup changed, and required additional inference steps that increased latency and cost.^[4]

The conceptual foundation of WebMCP originated in early 2025 at Amazon. Alex Nahas, an Amazon engineer, encountered authentication and authorization challenges (relating to OAuth) when attempting to deploy Anthropic's Model Context Protocol (MCP) across Amazon's internal services.^[4] To address these issues, Nahas developed MCP-B (Model Context Protocol for Browser), a Chrome extension and client-side architecture that ran MCP entirely within the browser, leveraging the user's existing federated Single Sign-On (SSO) session.^[4]

Concurrently, engineers from the Google Chrome and Microsoft Edge teams were prototyping structured agent-web interaction models, including Chrome's "script tools" proposal.^[4] The teams eventually converged, bringing Nahas into the W3C Web Machine Learning Community Group to formalize the approach into the WebMCP specification.^[1] In August 2025, Microsoft's Edge team publicly introduced the joint proposal,^[5] and by February 2026, Google launched an early developer preview of WebMCP in Chrome 146.^[6]

Architecture and Implementation

WebMCP provides two distinct APIs for developers to make their websites agent-ready:

Declarative API: Developers add specific HTML attributes (toolname and tooldescription) to existing standard HTML forms. The browser automatically translates these fields into a structured JSON Schema that AI agents can interpret.^[3]
Imperative API: Developers register complex or dynamic tools programmatically via JavaScript using the document.modelContext interface. This allows tools to be registered or unregistered based on the current page state — for example, a "checkout" tool surfacing only when items are present in a shopping cart.^[7]

When an agent invokes a WebMCP tool, the browser handles the execution in a single round trip, returning structured data rather than requiring the agent to parse a new page load.^[4]

Browser Implementation Status

As of June 2026, WebMCP is progressing through Google Chrome's standard feature launch process. The specification is formally tracked on the Chrome Platform Status dashboard (Feature ID: 5117755740913664) and described as a specification being incubated in a Community Group.^[8] The estimated milestones for Chrome on Desktop and Android are as follows:

Stage	Chrome Milestone
Dev Trial (behind `chrome://flags`)	Chrome 146 (February 2026)
Origin Trial (opt-in for developers)	Chrome 149 to 156
Shipping (stable, on by default)	Chrome 157 (estimated)

The Dev Trial in Chrome 146 was the early preview program launched in February 2026, accessible by enabling the WebMCP for testing flag in chrome://flags. The Origin Trial phase, spanning Chrome 149 through 156 and announced at Google I/O in May 2026, allows web developers to opt their sites into the API without a flag, enabling real-world testing at scale.^[8] Shipping in Chrome 157 represents the estimated stable release. These milestones are Chrome's current estimates and are subject to change before final release. No other browser vendor has announced a formal implementation timeline, though Microsoft's co-authorship of the specification suggests Microsoft Edge support is probable.

Ecosystem and Adoption

Unlike traditional MCP, which operates on a client-server architecture via JSON-RPC, WebMCP runs entirely within the browser tab.^[3] This allows it to inherit the user's existing authentication state, session cookies, and Single Sign-On (SSO) credentials without requiring separate API keys.^[2]

As the standard gained developer attention following the Chrome 146 preview, the developer community began cataloging WebMCP-enabled websites and tooling. The W3C Web Machine Learning Community Group maintains an official "Awesome WebMCP" curated list on GitHub, and Google Chrome Labs maintains a parallel list of third-party demos and implementations.^[1] Independent directories also emerged, including webmcp.com, which catalogs live websites exposing WebMCP tools and allows AI agents to discover site capabilities prior to navigation.^[9] The broader ecosystem includes polyfill packages and developer tooling, such as the @mcp-b/global npm package, which implements the WebMCP API and converts it to JSON-RPC for compatibility with existing MCP clients.^[4]

Early adoption has been most prominent in sectors requiring complex multi-step workflows, such as e-commerce, travel booking, and B2B SaaS applications.^[6]

References

1 2 3 "WebMCP Draft Community Group Report". Web Machine Learning Community Group. W3C. Retrieved June 21, 2026.
1 2 "Model Context Protocol: The next big step in generating value from AI". Engineering.com. May 9, 2025.
1 2 3 McKenzie, Leigh (March 11, 2026). "WebMCP: What It Is, Why It Matters, and What to Do Now". Semrush.
1 2 3 4 5 6 Nabors, RL (February 11, 2026). "WebMCP: Making Every Website a Tool for AI Agents". Arcade.dev.
↑ Brosset, Patrick (August 28, 2025). "WebMCP: A proposal to let web developers control how AI agents interact with their pages". Windows Developer Blog.
1 2 Bandarra, André Cipriani (February 10, 2026). "WebMCP is available for early preview". Chrome for Developers.
↑ "WebMCP Explainer". Web Machine Learning Community Group. W3C. Retrieved June 22, 2026. The imperative API registers tools via document.modelContext.registerTool().
1 2 "WebMCP". Chrome Platform Status. Google. Feature ID: 5117755740913664. Retrieved June 21, 2026.
↑ "WebMCP Directory". webmcp.com. Retrieved June 21, 2026.

[w3c-spec-1] 1 2 3 "WebMCP Draft Community Group Report". Web Machine Learning Community Group. W3C. Retrieved June 21, 2026.

[engineering-mcp-2] 1 2 "Model Context Protocol: The next big step in generating value from AI". Engineering.com. May 9, 2025.

[semrush-3] 1 2 3 McKenzie, Leigh (March 11, 2026). "WebMCP: What It Is, Why It Matters, and What to Do Now". Semrush.

[arcade-4] 1 2 3 4 5 6 Nabors, RL (February 11, 2026). "WebMCP: Making Every Website a Tool for AI Agents". Arcade.dev.

[brosset-5] Brosset, Patrick (August 28, 2025). "WebMCP: A proposal to let web developers control how AI agents interact with their pages". Windows Developer Blog.

[chrome-epp-6] 1 2 Bandarra, André Cipriani (February 10, 2026). "WebMCP is available for early preview". Chrome for Developers.

[webmcp-explainer-7] "WebMCP Explainer". Web Machine Learning Community Group. W3C. Retrieved June 22, 2026. The imperative API registers tools via document.modelContext.registerTool().

[chromestatus-8] 1 2 "WebMCP". Chrome Platform Status. Google. Feature ID: 5117755740913664. Retrieved June 21, 2026.

[webmcp-dir-9] "WebMCP Directory". webmcp.com. Retrieved June 21, 2026.

[1]