Desktop Automation MCP Server

@tanob

リポジトリ

基本情報

カテゴリ

生産性

ライセンス

MIT

トランスポート

stdio

作者

tanob

投稿者

Adriano Bonat

Desktop Automation MCP Server について

A Model Context Protocol server that provides desktop automation capabilities using RobotJS and screenshot capabilities

#desktop-automation#robotjs#llm-control

Desktop Automation MCP Server の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "desktop-automation": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-desktop-automation"
      ]
    }
  }
}

Desktop Automation MCP Server のツール

Gets the screen dimensions

Captures the current screen content

Presses a keyboard key or key combination

Types text at the current cursor position

Performs a mouse click

Moves the mouse to specified coordinates

概要

What is Desktop Automation?

A Model Context Protocol (MCP) server that provides desktop automation capabilities using RobotJS and screenshot functions. It enables Large Language Models to control mouse movements, keyboard inputs, and capture screenshots of the desktop environment.

How to use Desktop Automation?

Configure Claude Desktop with the NPX command as shown in the README. The server requires system-level permissions for screen capture, mouse control, and keyboard simulation, which may need to be granted in your operating system's security settings.

Key features of Desktop Automation

Desktop mouse control (move, click)
Keyboard input simulation (keys, text typing)
Screen size detection
Screenshot capture and listing
Simple JSON response format

Use cases of Desktop Automation

Automate repetitive desktop tasks via an LLM
Test GUI applications programmatically
Assist users by performing actions on their desktop
Capture screen content for analysis or documentation

よくある質問

What runtime does Desktop Automation require?

Node.js version 14.x or higher is required.

Which MCP clients are supported?

The server has been primarily tested with Claude Desktop, though it may work with other MCP clients.

Is there a limit on screenshot size?

Yes, the current implementation has a 1MB response size limit. High-resolution screenshots may fail; 800x600 resolution has been tested to work reliably.

What permissions does Desktop Automation need?

It needs system-level permissions to capture screenshots, control mouse movement and clicks, and simulate keyboard input.

What tools does Desktop Automation provide?

It provides tools: `get_screen_size`, `screen_capture`, `keyboard_press`, `keyboard_type`, `mouse_click`, and `mouse_move`.

「生産性」の他のコンテンツ

Local MCP (LMCP)

lanchuske

Local tools that let your AI use Mail, Calendar, Contacts, Teams, Slack, WhatsApp, OneDrive, Google Drive, Notion, Outlook and Office files on Mac and Windows. Runs on your computer.

ZenSched

ZenShows LLC

Field workforce scheduling for AI agents. Connect over MCP for account create, verified GPS punches, worker invites, shifts, forms, webhooks, and timesheets — no human dashboard. MCP: https://mcp.zensched.com/mcp Site: h

post bridge

jack friks

post bridge is a remote MCP server that lets your AI agent schedule and publish social media posts across 10 platforms: Instagram, TikTok, YouTube, X, LinkedIn, Facebook, Pinterest, Threads, Bluesky and Google Business P

EQIQs

EQIQs brings working-style and compatibility insights into Claude. Managers and coaches can take or send a short assessment, compare two people, get a read on a team of up to 12, prepare for meetings with tips that fit t

spacesheep

Publish pages your agent writes to a private, shareable URL, then read, comment on and update them. A hosted remote MCP server (OAuth) that works with Claude, ChatGPT, Codex, Cursor and any MCP client.