LLM/MCP Personal Assistant

A personal assistant application built using the Model Context Protocol (MCP) that enables sophisticated interactions between AI models and external tools/resources. Currently uses the Anthropic API and supports Wikipedia search and GitHub search, but is easily extendable.

Overview

This application is a personal assistant that leverages the Model Context Protocol to provide enhanced AI capabilities. It consists of three main components:

1. Client: A React-based web interface that provides the user interface for interacting with the assistant
2. API Server: An Express.js backend that handles client requests and manages the communication between components
3. MCP Server: Implements the Model Context Protocol to facilitate communication between AI models and external tools/resources

Features

• Modern React-based web interface
• RESTful API for handling client requests
• MCP server implementation for AI model integration
• Support for sophisticated AI interactions and tool usage
• Current tool support:
  • Wikipedia search
  • GitHub search
• Extendable architecture for adding new tools and capabilities

Note: This is a work in progress. While currently supporting Wikipedia and GitHub search capabilities, the architecture is designed to be easily extended with additional tools and features.

Prerequisites

• Node.js (v14 or higher)
• pnpm package manager (v10.9.0 or higher)

Installation

1. Clone this repository:

git clone https://github.com/mikefey/LLM-MCP-personal-assistant.git
cd LLM-MCP-personal-assistant

2. Install dependencies:

pnpm install

3. Create a .env file in the root directory with the following variables:

ANTHROPIC_API_KEY=your_anthropic_api_key_here
VITE_API_PORT=3001
VITE_API_HOST=localhost

Replace your_anthropic_api_key_here with your actual Anthropic API key. The default port and host values can be adjusted if needed.

Usage

1. Start the development server:

pnpm dev

This command will:

• Build the MCP server
• Start the API server
• Launch the Vite development server for the client

Architecture

Client

The client is a React application built with Vite that provides the user interface for interacting with the personal assistant.

API Server

The API server is built with Express.js and handles:

• Client requests
• Communication with the MCP server
• Session management
• Data persistence

MCP Server

The MCP server implements the Model Context Protocol specification, enabling:

• Communication with AI models
• Tool integration
• Context management
• Resource access

License

This project is licensed under the ISC License.

Acknowledgments

• Based on the Model Context Protocol specification
• Inspired by the tutorials from modelcontextprotocol.info