mcp-pyatv
MCP server for controlling Apple TV, HomePod, and AirPlay devices via pyatv, offering 32 tools for playback, navigation, volume, apps, and more.
README
mcp-pyatv
MCP server for controlling Apple TV, HomePod, and AirPlay devices via pyatv.
What It Does
mcp-pyatv is an MCP (Model Context Protocol) server that bridges any MCP client to Apple TV, HomePod, and AirPlay devices on your local network. It wraps pyatv -- the open-source Python library that implements the actual Apple TV and AirPlay protocols -- and exposes its functionality as MCP tools.
This lets you control your devices using natural language through Claude Desktop, Claude Code, Cursor, or any other MCP-compatible client.
This project does not implement any device protocols. All protocol-level communication is handled by pyatv. mcp-pyatv is purely the MCP bridge layer.
Supported Devices
- Apple TV -- all generations, including tvOS 15+
- HomePod / HomePod Mini
- AirPort Express
- Third-party AirPlay speakers
- macOS (Music/iTunes)
Quick Start
Install
pip install mcp-pyatv
Or run directly without installing:
uvx mcp-pyatv
Claude Desktop Configuration
Add this to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"apple-tv": {
"command": "uvx",
"args": ["--python", "python3.13", "mcp-pyatv"]
}
}
}
Note: Claude Desktop may not inherit your shell's PATH. If you get a "command not found" error, use the full path to
uvx(runwhich uvxin your terminal to find it):{ "mcpServers": { "apple-tv": { "command": "/Users/you/.local/bin/uvx", "args": ["--python", "python3.13", "mcp-pyatv"] } } }
Local Development
If you've cloned the repo and want to run from source:
{
"mcpServers": {
"apple-tv": {
"command": "/path/to/venv/bin/python",
"args": ["-m", "mcp_pyatv.server"],
"env": {
"PYTHONPATH": "/path/to/mcp-pyatv/src"
}
}
}
}
First Use / Pairing
On first use, just say "Pair my Apple TV" to your MCP client. The server handles the rest:
- It scans your network and discovers devices.
- A PIN appears on your TV screen.
- You tell the client the PIN.
- Credentials are stored locally and persist across sessions.
You only need to pair once per device. HomePods do not require pairing.
Available Tools
mcp-pyatv exposes 32 tools organized by category:
Discovery
| Tool | Description |
|---|---|
scan_devices |
Scan the local network for Apple TV and AirPlay devices |
device_info |
Get detailed information about a specific device |
Pairing
| Tool | Description |
|---|---|
start_pairing |
Begin the pairing process with a device |
finish_pairing |
Complete pairing by submitting the PIN shown on screen |
Playback
| Tool | Description |
|---|---|
play |
Resume playback |
pause |
Pause playback |
play_pause |
Toggle play/pause |
stop |
Stop playback |
next_track |
Skip to next track |
previous_track |
Go to previous track |
skip_forward |
Skip forward by a number of seconds |
skip_backward |
Skip backward by a number of seconds |
set_position |
Seek to a specific position |
set_shuffle |
Set shuffle mode |
set_repeat |
Set repeat mode |
Navigation
| Tool | Description |
|---|---|
navigate |
Send remote control commands -- supports up, down, left, right, select, menu, home, top_menu with single_tap, double_tap, or hold actions |
Audio
| Tool | Description |
|---|---|
get_volume |
Get the current volume level |
set_volume |
Set volume to a specific level |
volume_up |
Increase volume |
volume_down |
Decrease volume |
Apps
| Tool | Description |
|---|---|
list_apps |
List all installed apps |
launch_app |
Launch an app by name or bundle ID |
Power
| Tool | Description |
|---|---|
turn_on |
Turn on the device |
turn_off |
Turn off / put the device to sleep |
power_state |
Check if the device is on or off |
Keyboard
| Tool | Description |
|---|---|
set_text |
Type text into a text field (e.g., search boxes) |
get_text |
Get the current text field contents |
clear_text |
Clear the current text field |
Streaming
| Tool | Description |
|---|---|
play_url |
Stream media from a URL to the device |
stream_file |
Stream a local file to the device |
Media Info
| Tool | Description |
|---|---|
now_playing |
Get information about what's currently playing |
get_artwork |
Get the artwork for the currently playing media |
Example Conversations
Checking what's playing:
"What's playing on my Apple TV?"
The server calls
scan_devicesto find your Apple TV, thennow_playingto get the current track/show info.
Launching an app:
"Open Netflix"
The server calls
launch_appwith the Netflix bundle ID.
Adjusting volume:
"Set the volume to 40"
The server calls
set_volumewith level 40.
Pairing a new device:
"Pair my Apple TV"
The server calls
scan_devices, thenstart_pairing. A PIN appears on your TV. You say "The PIN is 1234", and the server callsfinish_pairingto complete the process.
Configuration
| Variable | Description | Default |
|---|---|---|
MCP_PYATV_STORAGE_PATH |
Override the file path where pairing credentials are stored | ~/.pyatv.conf |
The default credential file (~/.pyatv.conf) is shared with pyatv's atvremote CLI tool. If you've already paired devices using atvremote, mcp-pyatv will pick up those credentials automatically.
Known Limitations
play_urlbroken on tvOS 26.x -- upstream pyatv issue (#2821)stopignored by some apps -- YouTube and certain other apps do not respond to the stop commandget_artworkdepends on the app -- not all apps expose artwork metadata- Same network required -- your machine must be on the same WiFi network as your devices
set_shuffle/set_repeat-- require an active music playback queue to work- Python 3.14 -- not currently supported due to asyncio compatibility issues in pyatv
Built On
This project is built on pyatv by Pierre Stahl -- the Python library that implements the Companion, AirPlay, RAOP, MRP, and DMAP protocols used by Apple devices. All protocol-level communication, device discovery, and pairing logic happens inside pyatv. mcp-pyatv provides only the MCP server layer on top.
For protocol details, bug reports related to device communication, or to contribute to the underlying library, visit:
License
MIT -- see LICENSE for details.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.