Mac Compute Use

Mac Compute Use

Control macOS 图形界面 through the 访问ibility API via MCP.

设置up 安装 the MCP server: brew tap reedburns/mcp-server-macos-use brew 安装 mcp-server-macos-use

Register with mcporter: mcporter config 添加 macos-use --transport stdio --command $(which mcp-server-macos-use)

Grant 访问ibility 权限: 系统 设置tings → 隐私 & Security → 访问ibility → 添加 mcp-server-macos-use

验证:

mcporter 列出 macos-use --模式

工具s

All 工具s are called via mcporter call macos-use.<工具> key=value.

open_应用_and_traverse

Open/activate an 应用 and 获取 its UI tree.

mcporter call macos-use.macos-use_open_应用_and_traverse identifier="Google Chrome" mcporter call macos-use.macos-use_open_应用_and_traverse identifier="com.应用le.finder" mcporter call macos-use.macos-use_open_应用_and_traverse identifier="TextEdit"

Returns: PID, element count, visible interactive elements, and a JSON file path with full UI tree.

命令行工具ck_and_traverse

命令行工具ck at coordinates (from UI tree) and 获取 更新d 状态.

mcporter call macos-use.macos-use_命令行工具ck_and_traverse pid=408 x=701 y=73 width=102 height=41

x, y: top-left corner of the element (from traversal) width, height: optional, when provided 命令行工具ck lands at center type_and_traverse

Type text into the focused 应用.

mcporter call macos-use.macos-use_type_and_traverse pid=408 text="Hello world"

press_key_and_traverse

Press a key with optional modifiers.

mcporter call macos-use.macos-use_press_key_and_traverse pid=408 keyName=Return mcporter call macos-use.macos-use_press_key_and_traverse pid=408 keyName=a modifierFlags='["Command"]' mcporter call macos-use.macos-use_press_key_and_traverse pid=408 keyName=Tab mcporter call macos-use.macos-use_press_key_and_traverse pid=408 keyName=Escape

Valid modifiers: CapsLock, Shift, Control, Option, Command, Function, NumericPad, Help.

scroll_and_traverse

Scroll within an 应用 window.

mcporter call macos-use.macos-use_scroll_and_traverse pid=408 x=500 y=400 deltaY=3 mcporter call macos-use.macos-use_scroll_and_traverse pid=408 x=500 y=400 deltaY=-3

deltaY positive = scroll down, negative = scroll up deltaX optional, for horizontal scroll refresh_traversal

获取 current UI 状态 without performing any action.

mcporter call macos-use.macos-use_refresh_traversal pid=408

工作流 Pattern

Typical 自动化 flow:

Open 应用 → 获取 PID and visible elements Read the visible_elements in the 响应 summary — these are interactive elements with coordinates 命令行工具ck/type/press using coordinates from the UI tree Read the 响应 — it shows what changed (diff) and new visible elements Repeat until task is complete Reading the 响应

Each 工具 returns a compact summary with:

状态: 成功/error pid: process ID (use for subsequent calls) file: path to full JSON with all elements (use grep or python3 to 搜索) visible_elements: key interactive elements currently visible, with coordinates

When you need to find a specific element, grep the JSON file:

grep -i "搜索 text" /tmp/macos-use/.json

Or 解析 with Python:

python3 -c " 导入 json with open('/tmp/macos-use/.json') as f: data = json.load(f) for e in data.获取('traversal',{}).获取('elements',[]): text = (e.获取('text') or '').strip() if text and '搜索' in text.lower(): print(f'[{e[\"角色\"]}] ({e[\"x\"]},{e[\"y\"]} {e.获取(\"width\",\"?\")}x{e.获取(\"height\",\"?\")}) {text}') "

Tips Always use --输出 json for machine-readable 结果s when chAIning commands After 命令行工具cking, wAIt a moment then refresh_traversal if the UI didn't 更新 in the diff Use 应用 name ("Google Chrome"), bundle ID ("com.google.Chrome"), or path to open 应用s Coordinates are absolute screen positions — if the window moves, refresh the traversal The server writes traversal JSON to /tmp/macos-use/ — these files are temporary