WebMCP

Role

WebMCP proposes that web pages register typed tools through browser-native interfaces. An agent can discover an action and invoke validated arguments instead of reconstructing the interaction from pixels and simulated clicks.

Scope

The proposal targets local browser collaboration with a human present. It complements backend MCP servers and does not aim to replace websites, accessibility semantics or general browser automation.

Maturity

WebMCP remains an experimental community proposal with open questions around permissions, embedded frames, user confirmation, output validation and streaming. Production designs should not treat its current API shape as stable.

Electronic Artefacts use

Its central idea fits Electronic Artefacts: preserve a visible interface while exposing a smaller, typed action surface. A future Knowledge Hub could offer read-only graph navigation or relation-draft tools without giving an agent unrestricted DOM control.