Open-Websearch provides an MCP server that enables live web searching and content retrieval across multiple search engines (Bing, Baidu, DuckDuckGo, Brave, Exa, and others) without requiring API keys or authentication. The server offers structured search results with titles, URLs, and descriptions, along with content fetching capabilities for specific sources like CSDN, GitHub, and generic web pages, with support for HTTP proxy configuration. It solves the problem of integrating real-time web search into AI applications and agents while maintaining flexibility through multiple deployment options: MCP integration with clients like Claude Desktop, CLI for direct terminal usage, and a local daemon for reusable HTTP-based access.
claude mcp add --transport stdio aas-ee-open-websearch uvx open-websearch