PC-MCP

Overview

The PC‑MCP server is a lightweight Model Context Protocol (MCP) endpoint that exposes a handful of high‑value system and weather utilities to AI assistants. By running locally on a developer’s machine, it allows Claude or similar agents to fetch real‑time weather alerts and forecasts for the United States, schedule power events such as shutdown or restart, launch browsers with intelligent search capabilities, and capture screenshots—all through simple, well‑typed MCP commands. This eliminates the need for external API calls or custom scripting when an assistant needs to interact with a PC’s operating system or access up‑to‑date meteorological data.

At its core, the server bridges two worlds: operational control and environmental awareness. For developers building productivity or monitoring tools, the ability to trigger a graceful shutdown after a task completes, or to pre‑emptively check for severe weather alerts before initiating data transfers, can be critical. The server’s design focuses on safety and clarity: each command is validated with Zod, ensuring that malformed requests are rejected early, while the log file () provides a transparent audit trail of every action taken.

Key capabilities include:

• Weather Alerts: Query the National Weather Service for active warnings in a specified U.S. state, enabling agents to advise users on safety or adjust workflows during hazardous conditions.
• Weather Forecasts: Retrieve detailed, coordinate‑based forecasts, allowing assistants to schedule outdoor tasks or adjust HVAC settings intelligently.
• System Shutdown/Restart: Schedule power events with optional delays, force flags, or restart options—useful for automating maintenance windows or ensuring data is flushed before a reboot.
• Smart Browser Search: Open any major browser and perform keyword searches or directly navigate to inferred URLs, streamlining information retrieval without leaving the assistant’s context.
• Screenshot Capture: Take a visual snapshot of the current desktop, facilitating debugging, reporting, or evidence collection directly from an AI prompt.

Typical real‑world scenarios involve a developer using Claude to orchestrate nightly builds that must respect local weather conditions, or an operations team employing the server to automatically shut down a workstation after a critical alert is issued. In a research setting, an AI could gather weather data for a field study while also managing the laptop’s power state to conserve energy. The server’s cross‑platform support (Windows, macOS, Linux) and minimal dependency footprint make it a plug‑and‑play addition to any MCP‑enabled workflow, giving developers a powerful, secure interface to their own machines without exposing sensitive system internals.