MCP Scheduler

What is MCP Scheduler?

MCP Scheduler is a robust task scheduler server built with the Model Context Protocol (MCP) for scheduling and managing automated tasks. It supports shell commands, API calls, AI content generation via OpenAI, and desktop reminders with sound. The server uses cron expressions for flexible timing and provides a complete execution history.

How to use MCP Scheduler?

Install via uv or standard pip, then run python main.py with optional arguments (e.g., --transport sse --port 8080). Integrate with MCP clients like Claude Desktop by adding the server to its configuration file. Tasks are added via MCP tools such as add_command_task, add_api_task, add_ai_task, and add_reminder_task, each requiring a cron schedule and optional do_only_once flag.

Key features of MCP Scheduler

• Supports shell commands, API calls, AI tasks, and desktop notifications
• Uses standard cron syntax for precise scheduling
• Allows run-once or recurring task execution
• Tracks successful and failed task executions
• Cross-platform (Windows, macOS, Linux)
• Desktop alerts with sound for reminder tasks
• Seamless MCP integration with AI assistants
• Robust error handling with comprehensive logging

Use cases of MCP Scheduler

• Schedule daily database backups with a shell command
• Periodically fetch data from external APIs (e.g., weather)
• Generate weekly reports automatically using OpenAI
• Set recurring desktop reminders for team meetings

FAQ from MCP Scheduler

What are the prerequisites for MCP Scheduler?

Python 3.10 or higher and either uv (recommended) or standard pip are required.

How do I configure MCP Scheduler?

Configuration can be done via command-line arguments (e.g., --port, --db-path), a JSON configuration file, or environment variables (e.g., MCP_SCHEDULER_PORT, OPENAI_API_KEY).

What transport modes does MCP Scheduler support?

The server supports SSE (Server-Sent Events) and stdio (standard input/output) transport modes.

How do I set up AI tasks in MCP Scheduler?

Provide an OpenAI API key via the OPENAI_API_KEY environment variable or the configuration file, and optionally set the AI model with --ai-model (default gpt-4o).

Can I schedule a one-time task with MCP Scheduler?

Yes, set the do_only_once parameter to True when adding a task, and it will execute only once at the scheduled time.