The Google Scholar MCP Server is a TypeScript-based MCP (Model Context Protocol) server built with the @modelcontextprotocol/sdk. It exposes Google Scholar search functionality through a streamable HTTP transport (StreamableHTTPServerTransport), enabling HTTP POST RPC-style interactions and HTTP GET SSE streams for real-time updates. The server uses a session-based architecture to support multiple simultaneous connections via unique session IDs, a pluggable Tool System for registering and executing tools, and robust error handling with JSON-RPC error responses. The primary tool currently exposed is search_google_scholar, which supports configurable search parameters (query, filters, etc.) and returns structured results with paper details. Transport capabilities include persistent connections and SSE-driven updates. The project demonstrates integration with AI models such as Google's Gemini via a client that discovers tools and performs function calling, enabling end-to-end chat workflows. The repository includes server and client components, a development outline, example usage, and guidance on adding new tools, configuring environment variables, and running the server on port 3000.