WebMCP · document.modelContext
Tools agents can use on your site
Browser agents can click and interact with pages. WebMCP adds explicit tools for compatible agents, making supported actions easier to discover and describe. It is an optional interaction capability, not a requirement for search visibility.
What just changed
WebMCP — the W3C Web Model Context API, document.modelContext — lets a page hand agents real tools instead of making them scrape. A tool is a named function with a description and a schema ("search this catalogue", "book this slot") that an agent can call directly, in the page, while the human watches.
The idea moved fast: proposed by Microsoft's Edge team in August 2025, co-authored by Google's Chrome team weeks later, accepted into a W3C community group that September, and shipped as a developer preview in Chrome in February 2026. It remains an evolving API: check current browser support and the tools your intended agents can actually use before implementing it.
What the free scan can tell you
The free AIEO scan scores any domain across four groups: Traditional SEO, AI Discoverability, Schema Coverage — and Agent Interaction (WebMCP), worth 20% of the overall score. It looks for WebMCP registration source signals and reports text reference counts. These are not counts of distinct or working tools: comments, library definitions or inactive code can contain references. JavaScript is not executed, so a pass does not verify runtime registration, discovery or successful tool calls. Missing evidence within the scan limits does not prove WebMCP is absent. Failed checks include suggested next steps.
AIEO registers 7 WebMCP tools on its own pages. With a compatible browser and agent, you can request a scan on the homepage and see its score card appear there. Our own source evidence passes this part of the scanner's rules. Test tool discovery and deliberate calls in your intended browser and agent to establish that they work; the score does not demonstrate stronger search rankings or more enquiries.
A real before-and-after
Theory is cheap, so we instrumented a live production site: enm-compass.life, a relationship-style quiz — 33 questions, three axes, radar-chart results, all computed in the browser.
An agent can now take the quiz with you: you talk through your preferences in plain language and the answers sweep visibly across the page. The quiz calculations run in the browser. That alone does not establish where an agent or its provider processes tool inputs and outputs: check those data flows before using personal or sensitive information. The score change above reflects detected source evidence, not runtime verification or a search or business outcome.
What we learned scanning the real web
One finding worth sharing with anyone adopting WebMCP: modern bundlers can hide tool source from static HTML checks. Vite, Astro and webpack routinely split the code that calls registerTool into a chunk that no <script src> tag ever names — it's reachable only through an import statement inside another script. A naive scanner can miss those registrations; we encountered this in our own testing. The AIEO scanner follows same-site import chains within its scan limits. Looking only at the HTML's script tags can therefore miss tools defined in imported code.
Try it with your own agent
Enable WebMCP in Chrome at chrome://flags/#enable-webmcp-testing (or use the MCP-B extension), open the homepage, and ask your agent to scan a site. Prefer raw data? The endpoint below returns the full scan as JSON — CORS open, built for agents.
curl "https://aieo.fyi/api/agent-scan?domain=example.com"