getsentry/xcodebuildmcp — explained in plain English
Analysis updated 2026-06-26
Let Claude Code or Cursor build and test your iOS app without you having to run terminal commands manually.
List available simulators and launch your app on a specific device from inside an AI coding conversation.
Integrate Xcode build feedback directly into your AI coding workflow so errors appear as structured tool results.
Install globally via Homebrew or npm to reuse across all your Xcode projects with any MCP-compatible AI client.
| getsentry/xcodebuildmcp | dushixiang/next-terminal | executeautomation/mcp-playwright | |
|---|---|---|---|
| Stars | 5,516 | 5,519 | 5,511 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | hard | moderate |
| Complexity | 2/5 | 4/5 | 3/5 |
| Audience | developer | ops devops | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires macOS 14.5 or later and Xcode 16 or later, won't run on older macOS or Xcode versions.
XcodeBuildMCP is a tool that connects AI coding assistants, such as Claude Code or Cursor, to Xcode, Apple's development environment for building iOS and macOS apps. It does this through a standard called the Model Context Protocol, which lets AI agents call tools and receive structured results rather than just reading or writing text. The practical effect is that an AI assistant can trigger builds, run tests, launch apps in a simulator, and inspect build logs without you having to copy-paste terminal output back and forth. The package ships as both a command-line tool and an MCP server. You can use the command line directly from a terminal to build for a simulator, list available tools, or check for updates. You point the CLI at your Xcode project file and specify a build scheme, and it calls Xcode's underlying build system on your behalf. The MCP server mode is what AI coding agents connect to when you configure them to use it, which turns those same capabilities into callable functions the agent can invoke during a conversation. Installation requires macOS 14.5 or later and Xcode 16 or later. You can install the package globally through Homebrew or npm. Clients that support MCP, including Cursor, Codex, and Claude Code, can also run it on demand without a global install using a single npx command. Configuration snippets for each supported AI client are linked from the documentation site. The project is maintained by Sentry, the error-monitoring company. It uses Sentry's own telemetry to collect internal runtime errors, with opt-out instructions in the privacy documentation. An optional "skills" feature lets you load instructions into an AI agent that explain how to use the CLI or the MCP server effectively. The project is released under the MIT license.
A tool that lets AI coding assistants like Claude Code or Cursor trigger Xcode builds, run tests, and launch iOS simulators directly, so you stop copying terminal output back and forth while building Apple apps.
Mainly TypeScript. The stack also includes TypeScript, Node.js, MCP.
Use freely for any purpose including commercial use, just keep the MIT copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.