Submit Log in

XcodeBuildMCP

by Sentry

v2.7.0

MIT

Repository npm

Developer Tools

xcodeiosmacosswiftbuild

Summary

Sentry's MCP server and CLI for iOS and macOS work — build, run on simulators and devices, capture logs, and drive tests from inside a coding agent.

Features

Build and test against iOS simulators from inside the agent loop
Device operations for code-signed builds
Log capture, so failures are diagnosable rather than just red
Dual mode: MCP server or standalone CLI with a per-workspace daemon
Skips xcodebuild macro validation so Swift Macros projects build

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

Open your agent's MCP configuration file.
Merge the snippet below into it.
Restart the agent — the "XcodeBuildMCP" tools will be available.

{
  "mcpServers": {
    "xcodebuild-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "xcodebuildmcp@latest",
        "mcp"
      ]
    }
  }
}

Claude Code

Add the server with a single CLI command.

claude mcp add xcodebuild-mcp -- npx -y xcodebuildmcp@latest mcp

Claude Desktop

Add the server to claude_desktop_config.json (Settings → Developer → Edit Config).

{
  "mcpServers": {
    "xcodebuild-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "xcodebuildmcp@latest",
        "mcp"
      ]
    }
  }
}

Cursor

{
  "command": "npx",
  "args": [
    "-y",
    "xcodebuildmcp@latest",
    "mcp"
  ]
}

Description

Xcode is the part of iOS development an agent cannot reach. It can write Swift, but it cannot build the project, boot a simulator, install the app or read the crash log — so the loop breaks exactly where verification should happen. XcodeBuildMCP closes that gap by exposing xcodebuild and the simulator/device toolchain as MCP tools.

The tool surface covers building and testing against simulators, device operations for signed builds, build management, and log capture for debugging. That is enough for an agent to change code, build it, run it, watch what the app actually printed, and iterate — instead of stopping at "this should compile".

It ships in two modes from one package. As an MCP server it plugs into Claude Code, Cursor and any MCP client; as a CLI it is usable directly in a terminal, backed by a per-workspace daemon that keeps stateful operations alive between invocations. Optional agent skills are included to teach an agent how to use either mode well. A practical detail worth knowing: it asks xcodebuild to skip macro validation, which is what otherwise breaks builds that use Swift Macros.

Maintained by Sentry and MIT licensed. Requires macOS 14.5+, Xcode 16.x or later, and Node 18+ unless installed through Homebrew. Sentry telemetry is on by default and can be opted out of.