macPilotCli
MacPilot CLI is a mcp server, It provides a collection of system tools that allow AI assistants to perform various operations on macOS systems.
claude mcp add --transport stdio huhu415-macpilotcli /path/to/downloaded/binary
How to use
MacPilot CLI exposes a set of macOS system tools via the MCP protocol so AI assistants can perform tasks on a Mac. The available tools include: getCursorPosition, controlMouse, pasteText, captureScreen, shell, openApp, listApps, getWindowsInfo, and getWindowA11yInfo. To use the server, run the provided binary and configure your MCP-compatible client to point at the macPilotCli server entry. Once connected, send requests corresponding to the tool you want to use; the server will execute the requested action on the host Mac and return results or status information. This enables automated workflows where an LLM can read cursor coordinates, move the mouse, paste text, take screenshots, run shell commands, launch apps, enumerate installed applications, and inspect window information and accessibility details.
How to install
Prerequisites:
- A Mac with Apple Silicon running macOS 14.0 Sonoma or later
- Access to the internet to download the binary from the Releases page
Installation steps:
- Download the macPilot Cli binary from the project Releases page.
- Place the binary somewhere on your Mac (e.g., /usr/local/bin/macPilotCli).
- Ensure proper permissions are granted for screen recording and accessibility in System Settings for the terminal or host application that will run the MCP client.
- Start or configure your MCP-compatible application to connect to the macPilotCli server using the command path provided below.
Example configuration snippet (to be used by your MCP client):
{
"mcpServers": {
"macPilotCli": {
"command": "/path/to/downloaded/binary"
}
}
}
Notes:
- The binary should be executed on macOS with the required permissions granted to the host application (e.g., Cursor orClaude Desktop app) that will use the MCP interface.
- If you encounter permission prompts, follow the macOS dialog prompts to grant screen recording and accessibility access to the host application.
Additional notes
Tips and tips:
- Ensure the host macOS app has the necessary Accessibility and Screen Recording permissions for the best experience.
- When using shell commands via the MCP interface, be mindful of security and avoid exposing sensitive data.
- The MacPilot CLI currently implements a subset of MCP tool capabilities; monitor the project for new tools such as drag-and-drop mouse interactions as they are implemented.
- If the MCP client requires specific environment variables, you can extend the mcp_config with an env field under the macPilotCli server block as needed.
Related MCP Servers
rtfmbro
rtfmbro provides always-up-to-date, version-specific package documentation as context for coding agents. An alternative to context7
apple-mail
MCP server giving AI assistants full access to Apple Mail - read, search, compose, organize & analyze emails via natural language
mirroir
MCP server for controlling a real iPhone via macOS iPhone Mirroring...and any MacOs app. Screenshot, tap, swipe, type — from any MCP client.
finance
LLM-powered MCP server for building financial deep-research agents, integrating web search, Crawl4AI scraping, and entity extraction into composable analysis flows.
macos-tools
A Model Context Protocol server that brings macOS system tools directly inside your LLM client. Monitor system performance in real-time, analyze resource usage, search files with advanced patterns, and manage file tags using native macOS APIs.
instruments
MCP server for Xcode Instruments — AI agents get structured performance profiling (CPU, SwiftUI, memory, hitches) instead of raw xctrace XML