Mcp Android Agent Python

Created By

nim4447 months ago

This project provides an **MCP (Model Context Protocol)** server for automating Android devices using [uiautomator2](https://github.com/openatx/uiautomator2). It's designed to be easily plugged into AI agents like GitHub Copilot Chat, Claude, or Open Interpreter to control Android devices through natural language.

# adb

# android

Overview Content Tools Comments

Overview

🛠️ Setup

1. Clone the repo

git clone https://github.com/nim444/mcp-android.git
cd mcp-android

2. Create and activate virtual environment

# Using uv (https://github.com/astral-sh/uv)
uv venv
source .venv/bin/activate

3. Install dependencies (from pyproject.toml)

uv pip install

4. Run with uvicorn (recommended)

uvicorn server:app --factory --host 0.0.0.0 --port 8000

Alternatively, run with MCP stdio (for AI agent integration):

python server.py

🤖 Available MCP Tools

Tool Name	Description
`mcp_health`	Check if the MCP server is running properly
`connect_device`	Connect to an Android device and get basic info
`get_installed_apps`	List all installed apps with version and package info
`get_current_app`	Get info about the app currently in the foreground
`start_app`	Start an app by its package name
`stop_app`	Stop an app by its package name
`stop_all_apps`	Stop all currently running apps
`screen_on`	Turn on the screen
`screen_off`	Turn off the screen
`get_device_info`	Get detailed device info: serial, resolution, battery, etc.
`press_key`	Simulate hardware key press (e.g. `home`, `back`, `menu`, etc.)
`unlock_screen`	Unlock the screen (turn on and swipe if necessary)
`check_adb`	Check if ADB is installed and list connected devices
`wait_for_screen_on`	Wait asynchronously until the screen is turned on
`click`	Tap on an element by `text`, `resourceId`, or `description`
`long_click`	Perform a long click on an element
`send_text`	Input text into currently focused field (optionally clearing before)
`get_element_info`	Get info on UI elements (text, bounds, clickable, etc.)
`swipe`	Swipe from one coordinate to another
`wait_for_element`	Wait for an element to appear on screen
`screenshot`	Take and save a screenshot from the device
`scroll_to`	Scroll until a given element becomes visible
`drag`	Drag an element to a specific screen location
`get_toast`	Get the last toast message shown on screen
`clear_app_data`	Clear user data/cache of a specified app
`wait_activity`	Wait until a specific activity appears

Server Config

{
  "mcpServers": {
    "mcp-android": {
      "type": "stdio",
      "command": "bash",
      "args": [
        "-c",
        "cd <DIRECTORY> && source .venv/bin/activate && python -m server"
      ]
    }
  }
}

Recommend Servers

TraeBuild with Free GPT-4.1 & Claude 3.7. Fully MCP-Ready.

Context7Context7 MCP Server -- Up-to-date code documentation for LLMs and AI code editors

Howtocook Mcp基于Anduin2017 / HowToCook （程序员在家做饭指南）的mcp server，帮你推荐菜谱、规划膳食，解决“今天吃什么“的世纪难题； Based on Anduin2017/HowToCook (Programmer's Guide to Cooking at Home), MCP Server helps you recommend recipes, plan meals, and solve the century old problem of "what to eat today"

DeepChatYour AI Partner on Desktop

CursorThe AI Code Editor

WindsurfThe new purpose-built IDE to harness magic

Baidu Map百度地图核心API现已全面兼容MCP协议，是国内首家兼容MCP协议的地图服务商。

MCP AdvisorMCP Advisor & Installation - Use the right MCP server for your needs

Jina AI MCP ToolsA Model Context Protocol (MCP) server that integrates with Jina AI Search Foundation APIs.

BlenderBlenderMCP connects Blender to Claude AI through the Model Context Protocol (MCP), allowing Claude to directly interact with and control Blender. This integration enables prompt assisted 3D modeling, scene creation, and manipulation.

MiniMax MCPOfficial MiniMax Model Context Protocol (MCP) server that enables interaction with powerful Text to Speech, image generation and video generation APIs.