MCP Web Research Server

Example Use Cases

• 1

  Perform up-to-date Google searches and extract structured search results to inform responses.

• 2

  Visit webpages, extract full page content, and optionally capture screenshots for citation or visual reference.

• 3

  Maintain a research session history (queries, visited pages, extracted content, screenshots, timestamps) to support iterative investigations.

Description

This server integrates Google search, webpage content extraction, screenshot capture, and session tracking into Claude via the Model Context Protocol. It exposes agent tools like search_google, visit_page, and take_screenshot, and stores resources including captured screenshots and a research session history (queries, visited pages, extracted content, timestamps). The package is installed/started through Claude Desktop and is intended to help Claude perform iterative, source-citing web research. It is described as pre-alpha software, so users should expect bugs.