MCP-Server

MCP-Server

0
knowledge_and_memoryai_chatbot

The project is a server implementing the Model Context Protocol for role-based context management of AI agents, allowing dynamic role-specific memory management and tone adaptation. It provides comprehensive tools and integration for managing roles and processing role-based queries.

Overview

This project is a Model Context Protocol (MCP) server aimed at providing role-based context management for AI agents. Key features include:

  • Establishing distinct instructions and domain knowledge for each AI role
  • Partitioned memory management to prevent cross-contamination between roles
  • Dynamic adaptation of tone and style per role

Features

  • Role Management: Create, update, delete custom roles, with pre-defined expertise
  • Memory Management: Role-specific memory storage with TTL and limits
  • MCP Integration: Expose roles as resources; tools for query processing

Getting Started

  • Prerequisites: Node.js 18+, OpenAI API key
  • Installation: Clone the repo, install dependencies, and set environment variables
  • Running the Server: Use npm start to run the MCP server and npm run start:http for the HTTP server

MCP Integration

  • Resources: Access role information and tone profiles
  • Tools: Process queries, manage roles, and memory actions
  • Prompts: Use roles to process requests or create roles