AI agents
WebMCP
WebMCP is a proposed browser API that lets web applications offer JavaScript-based tools to AI agents.
In plain words
A tool is a JavaScript function with a plain-language description and a structured schema. An agent can call it instead of clicking its way through the page.
The specification is published by the W3C Web Machine Learning Community Group. It states that it is not a W3C Standard and not on the W3C Standards Track.
Example
A travel site could offer a "search flights" tool. Instead of filling in the form field by field, an agent calls that tool with the dates and the destination.
Why it matters for AI search
The specification says pages using WebMCP can be thought of as MCP servers that run their tools in the browser instead of on a back end.
Lighthouse 13.3 added a WebMCP audit next to its new agentic browsing category.
What you can do
- Treat it as experimental: it is a community group specification, not a standard.
- First make your forms and buttons work well for agents, with clear labels.
- Read our guide to MCP for the server-side version of the same idea.
Often confused with
- MCP
- MCP connects AI applications to outside systems, usually through a server. WebMCP brings the same idea into the web page itself.
Sources
- W3C Web Machine Learning Community Group: WebMCP
- Model Context Protocol: What is the Model Context Protocol (MCP)?
- Lighthouse v13.3.0 release notes (GitHub)
Facts and sources last checked on 19 October 2026.
Related terms
Go deeper on GEO-Score
How does your page do in AI search?
GEO-Score checks one page on 22 signals, from AI crawler access to structured data, and shows what to fix first.
Analyze a page