# Multi-server

An MCP client application based on Streamlit, supporting connection to multiple LLM providers (such as OpenAI, Anthropic, etc.) and interaction with MCP servers, providing tool testing and chat interface functions.

This project demonstrates how to use the MCP protocol to connect multiple servers with different transport methods (stdio and SSE), including mathematical calculation and weather query functions, and integrates with LangChain to create intelligent agents.

MCP Bridge is a lightweight, LLM-independent RESTful proxy for connecting to multiple Model Context Protocol (MCP) servers and exposing their functions through a unified REST API. It solves the problem that platforms such as edge devices, mobile devices, and web browsers cannot efficiently run MCP servers, and provides optional risk-based execution levels, including security controls such as standard execution, confirmation workflows, and Docker isolation.

AI agent robot project based on multi-MCP servers

Doroseek is an AI application built on Deno and Fresh, providing OpenAI-compatible API endpoint management, multi-server MCP functions, and proxy services, supporting key sharing and model alias settings.

The MetaTool MCP Server is a proxy server used to connect multiple MCP servers and forward tool calls to the appropriate servers. It is used in conjunction with the metatool-app to provide GUI tool management functions. The project is deprecated, and the Node.js alternative version is recommended.

The MCP Gateway is a reverse proxy server used to centrally manage and forward client requests to multiple MCP servers, supporting SSE streams and tool calls.

This project demonstrates a real-time multi-tool agent architecture based on LangGraph and the MCP protocol, achieving decoupled AI agents and remote tool execution, supporting asynchronous, modular, and scalable real-time interaction.