MCPlex-AI v1.0

This document provides an introduction to MCPlex-AI v1.0, a system that integrates Google’s Gemini large language model with external tools through a Model Context Protocol (MCP). Unlike typical chatbot frontends, MCPlex-AI runs entirely from the command line, giving users fine-grained control and a minimalistic interface to execute intelligent queries and trigger tool-based actions — all from a single terminal. The document covers the system’s purpose, high-level architecture, key components, and interaction patterns. For detailed information about specific components, please refer to their respective sections:

Purpose

MCPlex-AI v1.0 is designed to enhance the capabilities of large language models (Google Gemini-2.0) by enabling them to interact with external tools and APIs. This allows the AI to:

1. Access real-time information (news, Wikipedia, etc.)
2. Perform specific actions (post to Twitter, search files)
3. Calculate results (truth tables, mathematical operations)
4. Retrieve data from external services (GitHub, OMDB)

System Architecture

The system follows a client-server architecture where the client handles user interaction and AI communication, while the server manages tool execution. MCPlex-AI consists of three primary components:

• MCP Client (client/index.js): Handles user chat with Google Gemini AI, sends tool execution requests to the server
• MCP Server (server/index.js): Manages tool registration and execution, handles client requests
• Tool System (server/tools.js): Collection of tools that extend the AI’s capabilities

The diagram below illustrates the high-level architecture of the system:

Available Tools

Usage

1. Print Menu Tool

Name: print-menu Description: Prints what the MCP server can do with the available tools, provides descriptions of tools.

Input Parameters:

• title (string, optional): Optional title for the menu