What permissions does π Automation MCP require?
On macOS, you must grant Accessibility (for mouse and keyboard control) and Screen Recording (for screenshots and analysis) permissions in System Settings.
@ashwwwin
Basic information
Control your Mac with detailed mouse, keyboard, screen, and window management capabilities.
Add this server to your MCP-compatible client using the configuration below.
{
"mcpServers": {
"automation-mcp": {
"command": "bun",
"args": [
"run",
"index.ts"
]
}
}
}Tool list not available yet
This listing does not currently have a tool list to display. This does not mean the server has no tools; check its documentation for supported capabilities.
We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.
π Automation MCP is a Model Context Protocol (MCP) server that gives AI models complete desktop automation capabilities on macOS. It enables AI assistants to control the mouse, keyboard, take screenshots, manage windows, interact with UI elements, and analyze screen content.
Install via the furi tool with furi add ashwwwin/automation-mcp and then furi start automation-mcp. Alternatively, clone the repository, install dependencies with bun install, and start the server using bun run index.ts (HTTP transport) or bun run index.ts --stdio (stdio transport). On first run, grant Accessibility and Screen Recording permissions in macOS System Settings.
On macOS, you must grant Accessibility (for mouse and keyboard control) and Screen Recording (for screenshots and analysis) permissions in System Settings.
The server requires the Bun runtime. Quick install: `curl -fsSL https://bun.sh/install | bash`.
Yes. Clone the repository, run `bun install`, then start the server with `bun run index.ts` (HTTP) or `bun run index.ts --stdio` (stdio).
The server is primarily designed for macOS. Its architecture mentions crossβplatform support (Windows/Linux) via nut.js and fallback systems, but full functionality is documented for macOS.
It runs locally only (no network access), requires explicit permission grants, logs all actions, and can be limited to specific applications.
WhatsMCP β WhatsApp for AI agents. Put an AI agent on a real WhatsApp number. Expose regular WhatsApp accounts to your LLM agents over JSON-RPC. A complete Model Context Protocol server that turns phone numbers into addr

Read and update live small-business inventory: stock levels, items, suppliers, reorder suggestions and purchase orders.

Run your TRMNL ePaper displays from any AI agent: check on devices, plan playlists and build new screens from your own data.
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.

Multiplayer coding in the agentic era through an MCP server for real-time codebase sync between your agents.
Comments