mcp-on-android-tv
A Model Context Protocol (MCP) community server running on Android TV with on-device ADB integration - Control your TV with natural language.
Documentation
MCP on Android TV
A Model Context Protocol (MCP) server running directly on your Android TV devices.
By leveraging ADB (Android Debug Bridge), it allows local network control of your Android TV through any LLM agent or client that supports remote MCP servers. This makes it possible to open apps, list apps, and run other commands on your TV using natural language.
> ⚠️ Work in Progress
> This project is in an early stage of development. Features may be incomplete, unstable, or subject to change.
> Contributions and feedback are very welcome as the project evolves.
Why Android TV?
- Simpler ADB setup: On Android TV, ADB debugging can be activated and connected to without requiring pairing.
- Stable local connection: Unlike wireless debugging on phones, the connection does not cycle, giving you a direct link to the device.
- MCP integration: With MCP layered on top of ADB, any LLM client supporting custom MCP servers (such as Goose or Claude Desktop) can act as a controller for your TV.
⚠️ Note: For security reasons this should only be made to work on your local network, exposing your TV to the internet is a big risk.
Features
- Direct ADB integration without pairing.
- MCP server running on Android TV devices.
- Works with the MCP Kotlin SDK 0.7.2 (official SDK).
- Local network device control via LLM clients.
- Current communication method: SSE (deprecated).
Exposed MCP Tools
The MCP server currently supports the following tools:
- [x] launch_tv_app — Launch a TV application by its package name.
- [x] get_installed_tv_apps — Get a list of installed TV applications.
- [x] take_screenshot — Take a screenshot of the device screen.
- [x] send_key_event — Send a key event to the device.
- [x] send_nav_key_event — Send a navigational key event (up, down, left, right, select, back, home).
- [x] type_text — Type text into the current input field.
- [x] tap_screen — Tap the screen at specific coordinates.
Installation & Usage
1. Build & install the app on your Android TV device or set-top box.
2. Ensure ADB debugging is enabled and working.
3. Launch the app — it will display the MCP Ready screen with:
4. On your LLM client (e.g., Goose, Claude Desktop):
5. Start using natural language to control your TV:
Contributing
Contributions are very welcome!
- Submit a pull request with new features, tools, or fixes.
- Ideas for extended control and utilities are encouraged.
- Please follow typical open-source etiquette when submitting changes.
License
This project is licensed under the MIT License—see the LICENSE file for details.
Frequently asked questions
What is mcp-on-android-tv?
mcp-on-android-tv is A Model Context Protocol (MCP) community server running on Android TV with on-device ADB integration - Control your TV with natural language.
How do I install mcp-on-android-tv?
Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.
Is mcp-on-android-tv open source?
Yes — it is hosted on GitHub at https://github.com/MiddlePoint-Solutions/mcp-on-android-tv and has 12 stars.
Related MCP tools
The official Kotlin SDK for Model Context Protocol servers and clients. Maintained in collaboration with JetBrains Trusted by 1100+ developers.
Kotlin Multiplatform sample with SwiftUI, Jetpack Compose, Compose for Wear, Compose for Desktop, and Compose for Web clients along with Ktor backend.
Kotlin/Compose Multiplatform project to show climate related emission data from https://climatetrace.org/data. Built for the Model Context Protocol to enhance A
🙌 OpenHands: Code Less, Make More for the Model Context Protocol. Enhance AI assistants with powerful integrations. Python-based implementation.
The all-in-one Desktop & Docker AI application with built-in RAG, AI agents, No-code agent builder, MCP compatibility, and more.
Universal memory layer for AI Agents; Announcing OpenMemory MCP - local and secure memory management. Python-based implementation.
Run your own MCP server? See who uses it and what to fix.
Measure it with TrackMCP