📌Overview

Purpose: The MCP Orchestrator Server enables task management and coordination among multiple MCP enabled LLM instances, facilitating task creation, sharing, and execution by AI agents.

Overview: The MCP Orchestrator Server is designed to streamline the management of tasks across various AI instances like Claude Desktop or Cline. By providing robust features for task coordination, it significantly enhances the operational efficiency of AI agents.

Key Features:

Task Management: Supports task creation with dependencies, allowing for complex workflows among multiple instances.
Multi-instance Coordination: Facilitates seamless coordination and collaboration between different LLM instances, ensuring efficient task execution.
Persistent Task Storage: Stores tasks persistently to maintain state across different sessions and instances, enhancing reliability.
Dependency Enforcement: Ensures that tasks are executed in the correct order by enforcing dependencies, mitigating risks of execution errors.
Enhanced State Management: Optimizes task state transitions, making it easier to track and manage tasks effectively throughout their lifecycle.

MCP Orchestrator Server

The MCP Orchestrator Server provides task management and coordination capabilities across MCP enabled LLM instances like Claude Desktop or Cline. In simpler terms, it allows AI agents to create, share, and execute tasks across instances.

Features

Version 1.1.0

Task Updates: Modify pending tasks
Safe Deletion: Delete tasks with dependency checks
Cycle Detection: Prevent dependency cycles
Tool Listing: Comprehensive tool documentation
Enhanced State Management: Improved task state transitions

Core Features

Task creation with dependencies
Multi-instance coordination
Persistent task storage
Dependency enforcement
Task status tracking

Installation

npm install
npm run build

Usage

Create a Task

await create_task({
  id: 'setup',
  description: 'Initial setup'
});

Get Next Task

const task = await get_next_task({
  instance_id: 'worker-1'
});

Complete Task

await complete_task({
  task_id: 'setup',
  instance_id: 'worker-1',
  result: 'System initialized'
});

Documentation

API Reference
Quick Start Guide
Changelog

Roadmap

Version 1.2.0

Task priorities
Timeouts
Instance management

Version 1.3.0

Task groups
Analytics
Dashboard

License

MIT

orchestrator-server