Provides tools for controlling an Android device via adb, including screen capture, tap, swipe, text input, key events, package management, app installation and launching, battery diagnostics, and logcat.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@adb-mcp check the battery level"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A single star is the cheapest thank-you ā ā the repo if this is useful. It's the only thing that helps other people find it.
Turn
adbinto a full Model Context Protocol surface. Tap, swipe, screencap, install, read packages, watch logcat ā directly from Claude, Cursor, or any MCP-capable agent. Your phone becomes an AI-controlled device.
"AI controls your phone" is the 2026 demo ā and adb is the closest thing to a universal phone API.
One command to mount your device as a set of tools any agent can call.
Works over WiFi (adb pair/connect) or USB on Linux, macOS, Windows, and Termux on Android.
Zero cloud ā everything stays on your local machine. Private by design.
Related MCP server: scrcpy-mcp
pipx install adb-mcp # or: pip install adb-mcp
adb-mcp # stdio transport, targets default deviceOr run with an explicit device:
adb-mcp --device <serial>Add this to claude_desktop_config.json:
{ "mcpServers": { "adb": { "command": "adb-mcp" } } }Add a new MCP server entry with command adb-mcp and type stdio.
Try it against the built-in mock adb, fully offline:
adb-mcp --mock
python -m pytest tests/ -q # 5 passing tests, zero hardware neededCategory | Tools |
Devices |
|
Screen |
|
Input |
|
Apps |
|
Diagnostics |
|
Each tool runs a scoped, time-limited adb shell/adb exec-out call ā no root, no
persistent anything.
adb-mcp/
āā pyproject.toml # hatchling build, `adb-mcp` console script
āā adb_mcp/
ā āā server.py # FastMCP server + all tool registrations
ā āā adb.py # Adb / MockAdb transport
ā āā __init__.py
āā tests/test_tools.py # mock-based, device-free tests
āā README.mdover-screenrecord GIF framing for the README demo
files/ module: scoped push/pull
CE/ streaming logcat
pairing wizard for wireless adb pair
topic badges + GH Actions test on push
flowchart LR
Agent[Any AI Agent] <--> MCP[MCP stdio]
MCP --> adbc{adb-mcp server}
adbc --> Shell[adb shell]
Shell --> Phone((Android Phone))
Phone --> T[tap / swipe / screencap]
Phone --> A[install / launch]
Phone --> L[logcat]
Phone --> P[packages]Control real Android and iOS devices with LLM agents ā tap, swipe, type, automate flows.
Melaya is a remote MCP server. It gives an assistant hands on your own Android phone and browser: it reads the screen through the accessibility tree, then taps, types and navigates inside the apps and sites you allow-list, with no per-app API. It also builds, schedules and runs agent pipelines across 6k+ connected tools. OAuth 2.1, nothing to install.
Drive real devices from your AI Coding tool. Embed a client SDK (Unity, Godot, Flutter, iOS/macOS, Android, React Native, Web) in your app, then capture screenshots, traverse the UI tree, inject taps and key events, and run automated test tasks on the physical device over a secure relay.
Control Android TV from any AI. 38 MCP tools: playback, recap, recommend, smart-home, schedules.
