Complete AI Training

MCP server · Developer tools

agent-device MCP server

by callstackincubator

Let your AI open, tap through, and check mobile apps on simulators and real devices, then save proof.

Flow diagram: you ask your AI “Open the Contacts app and add Ada as a new contact”, on your own computer the agent-device MCP server works with app on your device, and you get back A short report and proof.

This is a helper for people who build or test apps. It lets your AI assistant actually open an app on a phone simulator or a real device, look at the screen, tap buttons, and take screenshots. If you work on iOS or Android apps, this saves you from describing every screen by hand.

What is an MCP server? The 30-second version

On its own, your AI can only chat with you. An MCP server is a small helper program that gives your AI a new skill or a connection to another tool. This one connects your AI to agent-device, a tool that drives apps on phones, tablets, and computers. Once it is connected, you can ask your AI to open an app, check what is on screen, and do things there for you.

What this MCP server does

You ask your AI to do something with an app, like open it or tap a button. The AI uses this helper, which talks to the app running on a simulator, emulator, or real device. The helper reads the screen and reports back what it sees, then performs the taps or typing you asked for. You get back a short report of what happened, plus screenshots or logs if you want them. You can also save a run so you can repeat it later.

Flow diagram: you ask your AI “Open the Contacts app and add Ada as a new contact”, on your own computer the agent-device MCP server works with app on your device, and you get back A short report and proof. Click to zoom

What you can do with it

  • Open an app on an iOS simulator or Android emulator
  • Tap buttons, fill in fields, and scroll through screens
  • Take screenshots and record video as proof of what happened
  • Read the app's screen so your AI knows what is there
  • Capture logs and crash details when something goes wrong
  • Save a working run as a script you can replay later
  • Check performance and see if the app is doing extra work

Try asking your AI

  • “Open the Contacts app on the iOS simulator and add a new contact named Ada”
  • “Run the checkout flow on Android and take a screenshot at each step”
  • “Reproduce this crash and show me the logs right before it happens”
  • “Check if my latest change causes the screen to re-render too often”

What it gives back to you

You get short, readable answers in the chat, like what the app is showing and what changed after each action. It can also hand back screenshots, video files, logs, and performance numbers. If you asked it to save a run, you get a script file you can use again. Everything shows up as normal chat messages and files you can open.

Before you start

What you need

  • Node.js 22.12 or newer installed on your computer (24 or newer for web testing)
  • The agent-device tool installed, which you do with one command in your terminal
  • A simulator, emulator, or real device set up for the app you want to test
  • An AI assistant that can connect to MCP servers, like Claude Code, Cursor, or Codex

Good to know

It can tap, type, and change things inside your app, so be careful when pointing it at a real device or a live account.

Install it with your AI

Add agent-device MCP server to your AI, no technical skills needed

You don't install anything by hand. You copy one prompt, paste it into an AI that can work on your computer, and it checks, installs and connects the server for you, asking you when it needs something.

Sign in to get the install prompt

Members get a ready-made prompt that lets the Claude desktop app check agent-device MCP server, install it and connect it for them, step by step. You don't need any technical skills: you copy, paste and answer a few questions. Your connected AI can also find and install any of the 4,066 MCP servers here for you.

Sign in Become a member

Who it's for

Mobile app developers and QA testers who want their AI assistant to check apps on real devices or simulators.