Status: Emerging Standard — Experimental
WebMCP is an experimental web-level adaptation of Anthropic's Model Context Protocol (MCP). It is designed for autonomous AI agents and developer tooling. It is NOT evaluated as a direct search engine ranking factor by Googlebot or Bingbot.
Traditional web pages are designed for visual rendering to human users. When an autonomous AI agent queries a website, parsing full HTML consumes substantial token context and can introduce parsing errors. WebMCP provides a structured alternative.
Architecture of WebMCP
- Discovery Manifest (
/.well-known/mcp.json): Declares available tool endpoints, authentication requirements (if any), and schema definitions. - Structured Endpoints: Lightweight API routes returning predictable JSON (e.g.
/api/mcp/services,/api/mcp/contact). - LLM Reference: Linked from
/llms.txtto enable seamless agent discovery.
Experimental Deployment on InfinitusNow
To evaluate agentic interaction, InfinitusNow published an experimental discovery file at /.well-known/mcp.json. In synthetic testing with MCP-compatible clients, agent retrieval completed in under 120ms with zero HTML parsing overhead. However, standard search crawlers currently ignore these files, meaning WebMCP should complement—never replace—standard technical SEO and JSON-LD schema.