What does mcp-server-macos-use do?
It exposes MCP tools that let you open apps, click, type, and press keys, then traverse the accessibility tree—allowing programmatic control of macOS apps from an MCP client.
MCP Server Macos Use
@mediar-ai
基本信息
关于 MCP Server Macos Use
AI agent that controls computer with OS-level tools, MCP compatible, works with any model
MCP Server Macos Use 配置
MCP Server Macos Use 工具
10** Opens or activates a specified application and then traverses its accessibility tree.
`identifier` (String, Required): The application's name, bundle ID, or file path.
** Simulates a mouse click at specific coordinates within the window of the target application (identified by PID) and then traverses its accessibility tree.
`pid` (Number, Required): The Process ID (PID) of the target application.
** Simulates typing text into the target application (identified by PID) and then traverses its accessibility tree.
`pid` (Number, Required): The Process ID (PID) of the target application.
** Simulates pressing a specific keyboard key (e.g., 'Enter', 'Tab', 'a', 'B') with optional modifier keys held down, targeting the application specified by PID, and then traverses its accessibility tree.
`pid` (Number, Required): The Process ID (PID) of the target application.
** Only performs the accessibility tree traversal for the specified application (identified by PID). Useful for getting the current UI state without performing an action.
`pid` (Number, Required): The Process ID (PID) of the application to traverse.
概览
What is mcp-server-macos-use?
An MCP server in Swift that controls macOS applications via accessibility APIs through the MacosUseSDK. It is designed for users of MCP-compatible clients like Claude Desktop who need to automate or interact with desktop apps programmatically.
How to use mcp-server-macos-use?
Build the server with swift build -c debug, then run the executable. It communicates over standard input/output (stdio). Configure your MCP client (e.g., Claude Desktop) by pointing it to the server binary path in the client's configuration file.
Key features of mcp-server-macos-use
- Open or activate an application and traverse its accessibility tree
- Simulate mouse clicks at specified coordinates
- Simulate typing text into a target application
- Press keyboard keys with optional modifier flags
- Refresh traversal to get current UI state without action
- Control traversal behavior with optional parameters (diff, visibility, animation, delay)
Use cases of mcp-server-macos-use
- Automating repetitive UI tasks in any macOS application
- Querying the current UI state of a running desktop app
- Building AI-powered assistants that interact with native macOS software
- Accessibility tree inspection and testing
常见问题
What dependencies does mcp-server-macos-use require?
It requires the MacosUseSDK Swift package and Swift toolchain to build. No other runtime dependencies are listed.
How does mcp-server-macos-use communicate?
It uses the stdio transport: the server reads and writes MCP messages over standard input and output.
How do I integrate mcp-server-macos-use with Claude Desktop?
Add an entry to Claude Desktop's configuration with the command set to the absolute path of the built server executable, under the `mcpServers` key.
Where can I get help or request features?
Reach out via email (matt@mediar.ai) or Discord (m13v_). The maintainer is open to tailoring the server for specific needs.
生产力 分类下的更多 MCP 服务器
Etincel
AIStoryHubFind the AI tells in your prose. Deterministic, local, runs in CI. MCP server + CLI + Action.
6
Weft
AI PRODUCTSOne board for every AI you use: agents claim tasks and finish with proof you accept or return. Weft is a Scrumban task manager for solo founders and small teams — one board, your sprints, and the AI you already use driv
WhatsMCP
whatsmcpWhatsMCP — 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
fixRAgent maintenance triage
AR Logic LLCOne photo of a building problem becomes a Triage Profile in about ten seconds: what the asset is, how urgent it is (EMERGENCY, TODAY, THIS WEEK or WHENEVER), which trade to call, and what to tell the resident. For proper
ReactLive
Alan SunleyRun live event Q&A, polls and moderation from your AI assistant. Create and configure events, approve and answer audience questions, launch polls, manage moderators and pull post-event reports — in plain language.
评论