orchestrator-mcp-server
The Workflow Orchestrator MCP Server leverages AI to manage dynamic workflows, supporting complex task breakdowns and adaptable processes. Its modular architecture facilitates intelligent decision-making powered by a Large Language Model, with persistent state management and resumption capabilities.
Workflow Orchestrator MCP Server
Overview
This project implements an AI-Powered Workflow Orchestrator as a Model Context Protocol (MCP) server. It is designed to manage and execute complex workflows using a Large Language Model (LLM) for intelligent decision-making. The orchestrator breaks down complex tasks into manageable steps defined in workflows, with real-time feedback from clients.
Features
- Intelligent, non-linear workflows
- Reusable and modular steps in Markdown
- Human-readable and editable workflows
- Persistent state management using SQLite
- Workflow resumption capability
Workflows
Workflows are specified in subdirectories within a directory defined in the MCP server settings. Workflows include definitions and steps in Markdown files.
MCP Tools
- List, start, and manage workflows
- Resume workflows seamlessly
Configuration
Configurable through environment variables specifying paths and AI service settings.
Quickstart
- Ensure Python environment and directories are set up
- Install dependencies and run the server