WebMCP

WebMCP is a JavaScript library that bridges the gap between websites and AI. It lets developers add a small, configurable widget to their webpages. When a user connects via an MCP client—like Claude Desktop—the AI can then interact with the page. The AI isn't just reading the page; it can call specific functions the developer has exposed, like fetching data or triggering actions.

The library works by letting developers register three types of components: tools, prompts, and resources. Tools are functions the AI can execute, like getting weather info or submitting a form. Prompts are reusable templates for common LLM queries. Resources expose chunks of the page's data for the AI to read as context. This turns a static or dynamic website into an interactive platform for AI agents.

This is most useful for developers building prototypes, interactive tutorials, or complex web applications where AI assistance could streamline tasks. Imagine a dashboard where an AI can pull specific reports, or a form-heavy application where an agent can help fill fields. WebMCP provides the plumbing to make those AI-web interactions possible.