WebMCP lets websites clearly tell AI what they can do.
The extension adds a browser sidebar, WebMCP, and web search.
WebMCP (Web Model Context Protocol) is a proposed browser API from Google and Microsoft. It lets websites expose their capabilities to AI as structured tools, describing what each tool does and what information it needs.
Simply put, it is designed to help AI assistants work with websites more intelligently and precisely. Websites can tell AI directly: “what I can do and how to use it.”
AI infers what a site can do from its appearance and structure. This can take time and is sensitive to interface changes or ambiguous labels.
Infer each control's purpose and operation from screenshots or page structure.
Find controls, click, and enter information step by step.
Repeated attempts can increase token and compute use.
Ambiguous labels or page changes can lead to missed steps or mistakes.
The website provides a structured “guide to its capabilities”, so AI can call the right action directly.
Open a supported site to discover its available tools.
Use available actions directly with fewer searches and clicks.
Reduce repeated page analysis and save tokens.
Use clearly defined actions that are less affected by page changes.
EO2Weave connects AI conversations, websites, and local files in your browser sidebar. With WebMCP, it discovers website tools so AI can complete tasks using the actions the site provides.
Discover WebMCP tools on the current site and how AI can call them.
Provide tools through integrations for selected sites without native WebMCP support.
Review available website tools and decide whether to enable them before AI calls them.
Discover tools, grant access, and complete tasks from your browser conversation.
Try the full workflow of using website tools with WebMCP.
Install and try the extension in a desktop browser. Use online →