🎵 MCP Spotify Player

---

Control Spotify from your assistant (Claude/Cursor/Void/IntelliJ).

📖 Table of Contents

• Overview
• Key Features
• Prerequisites
• Installation
• Configuration
  • Spotify Configuration
  • MCP Configuration with Claude Desktop
• Usage
• API Reference (Python)
• Architecture
• Troubleshooting
• Contribute
• License

🎯 Overview

MCP Spotify Player is an innovative tool that integrates Spotify control with AI assistants via the Model Context Protocol (MCP). It enables users to control their music using natural language commands processed by an AI assistant, delivering a seamless music interaction experience.

Why MCP Spotify Player?

• AI Integration: Full Spotify control via natural language
• Automatic Authentication: Secure and simplified OAuth flow
• High Performance: Fast responses to commands
• Extensible: Modular architecture for adding features
• Open Source: Active community and transparent development

✨ Key Features

🎵 Playback Control

• ▶️ Play / Pause music
• ⏭️ Skip tracks
• 🔄 Queue management
• 🔊 Volume control
• 🔁 Repeat modes

🔍 Search & Discovery

• Search tracks, artists and albums
• Browse playlists
• Personalized recommendations
• Playback history

📱 Device Management

• Switch between available devices
• Multi-device synchronization
• Remote playback control

🎨 Playlist Management

• Create and edit playlists
• Add / remove tracks
• Automatic organization
• Share playlists

📋 Prerequisites

Before you begin, make sure you have installed:

• Python 3.10+
• Spotify premium account - Premium recommended
• Spotify App - Create an app

🔧 Installation

1. Clone the repository:

   git clone <your-repository>
   cd mcp-spotify-player

2. Install:

  pip install .

For development:

  pip install -e .

3. Set up environment variables:

cp env.example .env

Edit the .env file with your Spotify credentials:

   SPOTIFY_CLIENT_ID=your_client_id_here
   SPOTIFY_CLIENT_SECRET=your_client_secret_here
   SPOTIFY_REDIRECT_URI=http://127.0.0.1:8000/auth/callback
   # Optional: custom path to store OAuth tokens
   # Defaults to ~/.config/mcp_spotify_player/tokens.json
   MCP_SPOTIFY_TOKENS_PATH=/path/to/tokens.json

Note: dependencies are managed with pyproject.toml.

If MCP_SPOTIFY_TOKENS_PATH is not set, tokens will be stored in ~/.config/mcp_spotify_player/tokens.json by default.

⚙️ Configuration

Spotify Configuration

Create an application on the Spotify Developer Dashboard

1. Go to Spotify Developer Dashboard
2. Create a new application
3. Get your CLIENT_ID and CLIENT_SECRET
4. In the app settings, add http://127.0.0.1:8000/auth/callback as a redirect URI

Configure credentials

Create a .env file at the project root:

SPOTIFY_CLIENT_ID=your_client_id_here
SPOTIFY_CLIENT_SECRET=your_client_secret_here
SPOTIFY_REDIRECT_URI=http://localhost:3000/callback
MCP_SERVER_PORT=3000

Token file format

When tokens are stored on disk they use a simple JSON structure. Example:

{
  "access_token": "ACCESS",
  "refresh_token": "REFRESH",
  "expires_at": 1700000000,
  "scopes": [
    "user-read-playback-state",
    "user-modify-playback-state",
    "user-read-currently-playing",
    "user-read-recently-played",
    "user-read-playback-position",
    "user-top-read",
    "playlist-read-private",
    "playlist-read-collaborative",
    "playlist-modify-private",
    "user-library-read",
    "user-library-modify"
  ]
}

By default tokens are saved to ~/.config/mcp_spotify_player/tokens.json unless MCP_SPOTIFY_TOKENS_PATH is configured.

MCP Configuration with Claude Desktop

Add the server to your Claude Desktop configuration or other MCP client:

Windows (%APPDATA%\Claude\claude_desktop_config.json):

{
  "mcpServers": {
    "spotify-player": {
      "command": "C:/path/to/python/distribution/python.exe",
      "args": [
        "-m",
        "mcp_spotify_player"
      ],
      "cwd": "C:/path/to/cloned/repo/mcp-spotify-player",
      "env": {
        "SPOTIFY_CLIENT_ID": "your_client_id",
        "SPOTIFY_CLIENT_SECRET": "your_client_secret",
        "SPOTIFY_REDIRECT_URI": "http://127.0.0.1:8000/auth/callback"
      }
    }
  }
}

macOS/Linux (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "spotify-player": {
      "command": "/path/to/python/distribution/python.exe",
      "args": [
        "-m",
        "mcp_spotify_player"
      ],
      "cwd": "/path/to/cloned/repo/mcp-spotify-player",
      "env": {
        "SPOTIFY_CLIENT_ID": "your_client_id",
        "SPOTIFY_CLIENT_SECRET": "your_client_secret",
        "SPOTIFY_REDIRECT_URI": "http://127.0.0.1:8000/auth/callback"
      }
    }
  }
}

Configuring MCP Server in PyCharm / IntelliJ

JetBrains IDEs (PyCharm, IntelliJ, etc.) with GitHub Copilot have native support for MCP servers.

1. Go to File → Settings → GitHub Copilot → Model Context Protocol (MCP).
2. On the right side, click Configure to edit the mcp.json file.
3. Add the following snippet, replacing the paths and dummy credentials as needed:

{
  "servers": {
    "spotify-player": {
      "command": "C:/Users/YourUser/AppData/Local/Programs/Python/Python311/python.exe",
      "args": ["-m", "mcp_spotify_player"],
      "cwd": "/path/to/cloned/repo/mcp-spotify-player",
      "env": {
        "SPOTIFY_CLIENT_ID": "your-client-id",
        "SPOTIFY_CLIENT_SECRET": "your-client-secret",
        "SPOTIFY_REDIRECT_URI": "http://127.0.0.1:8000/auth/callback"
      }
    }
  }
}

Replace your-client-id and your-client-secret with the credentials from your Spotify Developer Dashboard. Do not commit your real credentials into version control.

Once configured, the server will appear in the Copilot Agent inside PyCharm/IntelliJ, and tools like /auth, /play_music, and /pause_music can be invoked directly.

Other MCP clients

Use mcp-spotify-player.yaml as a template (same fields: command, args, cwd, env).

🎮 Usage

Available Commands

Command	Description	Example
play_music	Play music	play_music — Play Where Is My Mind by The Pixies
pause_music	Pause playback	pause_music — Pause the music
skip_next	Next track	skip_next — Next song
skip_previous	Previous track	skip_previous — Previous song
set_volume	Set playback volume (0-100)	set_volume — Set volume to 50%
set_repeat	Set repeat mode (off, track, context)	set_repeat — Set repeat mode to off, track, or context (e.g., "Repeat current track")
get_current_playing	Get the currently playing track	get_current_playing — What's playing?
get_playback_state	Get full playback state	get_playback_state — What's the playback state?
get_devices	List available playback devices	get_devices — List my available devices
search_music	Search for tracks/artists/albums	search_music — Search for songs by Queen
search_collections	Search for playlists or albums	search_collections — Search for playlists or albums
get_playlists	List user playlists	get_playlists — List my playlists
get_playlist_tracks	Get tracks in a playlist	get_playlist_tracks — Show tracks in playlist 'Road Trip'
get_artist	Get artist information by ID	get_artist — Show info about artist with a given ID
get_artist_albums	Get albums for an artist by ID	get_artist_albums — Show albums of an artist by ID
get_artist_top_tracks	Get top tracks for an artist by ID	get_artist_top_tracks — Show top tracks of an artist by ID
get_album	Get album information by ID	get_album — Show info about album 'The Dark Side of the Moon'
get_albums	Get multiple albums information	get_albums — Show info about multiple albums
get_album_tracks	Get tracks in an album	get_album_tracks — Show tracks in album 'The Dark Side of the Moon'
get_saved_albums	List saved albums in library	get_saved_albums — List my saved albums
check_saved_albums	Check if albums are saved	check_saved_albums — Check if these albums are saved
save_albums	Save albums to library	save_albums — Save these albums to my library
delete_saved_albums	Remove albums from library	delete_saved_albums — Remove these albums from my library
create_playlist	Create a new playlist	create_playlist — Create playlist 'Road Trip' with these songs...'
rename_playlist	Rename an existing playlist	rename_playlist — Rename playlist 'Road Trip' to 'Vacation'
clear_playlist	Remove all tracks from a playlist	clear_playlist — Remove all songs from playlist 'Road Trip'
add_tracks_to_playlist	Add tracks to a playlist	add_tracks_to_playlist — Add these songs to playlist 'Road Trip'
queue_add	Add a track to the queue	queue_add — Add this track to the queue
queue_list	Show upcoming queue	queue_list — Show the upcoming queue
diagnose	Display diagnostic information	diagnose — Display diagnostic information

Usage Examples

With Claude or your AI assistant:

User: "Play relaxing music for work"
Claude: [Performs search and plays a relaxing playlist]

User: "Raise the volume to 80%"
Claude: [Sets volume to 80%]

User: "Create a playlist with the best 80s songs"
Claude: [Creates playlist and adds popular 80s tracks]

Programmatic usage (Python)

from pathlib import Path
from mcp_spotify_player.spotify_controller import SpotifyController
from mcp_spotify_player.client_auth import load_tokens

# Load tokens from the default path (adjust if you store them elsewhere)
tokens_path = Path.home() / '.config' / 'mcp_spotify_player' / 'tokens.json'
# load_tokens returns a Tokens object or raises if missing
tokens = load_tokens(tokens_path)

# SpotifyController expects a zero-arg callable that returns tokens
controller = SpotifyController(lambda: tokens)

if not controller.is_authenticated():
    raise SystemExit("Run the auth tool first (use the auth command in your MCP client)")

# Play a song by query (search + play)
controller.play_music(query="Bohemian Rhapsody")

# Get currently playing track
current = controller.get_current_playing()
if current and current.get("item"):
    name = current["item"]["name"]
    artist = current["item"]["artists"][0]["name"]
    print(f"Now playing: {name} — {artist}")

# Create a playlist and add tracks (example)
playlist_resp = controller.create_playlist(playlist_name="Road Trip 2024", description="Road trip mix")
# create_playlist returns {"success": True, "playlist": {...}}
playlist_id = None
if isinstance(playlist_resp, dict) and playlist_resp.get("success") and playlist_resp.get("playlist"):
    playlist_id = playlist_resp["playlist"].get("id")

if playlist_id:
    controller.add_tracks_to_playlist(playlist_id=playlist_id, track_uris=["spotify:track:TRACK_URI_1", "spotify:track:TRACK_URI_2"])
else:
    print("Could not create playlist or retrieve playlist id")

📚 API Reference (Python)

This project exposes a Python-first API via the SpotifyController facade. The controller delegates functionality to PlaybackController, PlaylistController, Albums and Artists controllers. Signatures and return values below reflect the current implementation.

SpotifyController

A facade that aggregates playback, playlist, album and artist operations.

Constructor

SpotifyController(tokens_provider: Callable[[], Optional[Tokens]])

• tokens_provider: zero-argument callable that returns a Tokens object or None.

Methods (delegated to underlying controllers)

• is_authenticated() -> bool
  • Returns True if tokens are present and not expired.

Playback-related (via PlaybackController)

• play_music(query: Optional[str] = None, playlist_name: Optional[str] = None, track_uri: Optional[str] = None, artist_uri: Optional[str] = None) -> Dict[str, Any]

  • Play a track by search query, specific track URI, playlist name or artist URI. Returns a dict with success/message and optional details.

• pause_music() -> Dict[str, Any]

  • Pause playback.

• skip_next() -> Dict[str, Any]

  • Skip to next track.

• skip_previous() -> Dict[str, Any]

  • Skip to previous track.

• set_volume(volume_percent: int) -> Dict[str, Any]

  • Set device volume (0-100).

• set_repeat(state: str) -> Dict[str, Any]

  • Set repeat mode. state is one of 'off', 'track', 'context'.

• get_current_playing() -> Dict[str, Any]

  • Returns a dict with success and track info (name, artist, album, uri, duration_ms, external_url), is_playing and progress_ms.

• get_playback_state() -> Dict[str, Any]

  • Returns full playback state including is_playing, current_track (TrackInfo dict) volume_percent, device_name, shuffle_state and repeat_state.

• get_devices() -> Dict[str, Any]

  • Returns a list of available playback devices.

• search_music(query: str, search_type: str = "track", limit: int = 10) -> Dict[str, Any]

  • Search for tracks or artists. For track searches returns a list of TrackInfo dicts.

• search_collections(q: str, type: str, limit: int = 20, offset: int = 0, market: Optional[str] = None) -> Dict[str, Any]

  • Search for playlists or albums. Returns container info and items.

• queue_add(uri: str, device_id: Optional[str] = None) -> Dict[str, Any]

  • Add a track/episode URI to the active device queue.

• queue_list(limit: Optional[int] = None) -> Dict[str, Any]

  • Get the upcoming queue (delegates to the client).

PlaylistController

• get_playlists() -> Dict[str, Any]

  • Returns user's playlists as a list of PlaylistInfo dicts.

• create_playlist(playlist_name: str, description: str = "") -> Dict[str, Any]

  • Create a new playlist (private by default). Returns the created playlist info.

• get_playlist_tracks(playlist_id: str, limit: int = 20) -> Dict[str, Any]

  • Return tracks in a playlist as TrackInfo dicts.

• rename_playlist(playlist_id: str, playlist_name: str) -> Dict[str, Any]

  • Rename a playlist.

• clear_playlist(playlist_id: str) -> Dict[str, Any]

  • Remove all tracks from a playlist.

• add_tracks_to_playlist(playlist_id: str, track_uris: List[str]) -> Dict[str, Any]

  • Add valid Spotify track URIs (spotify:track:...) to a playlist.

Data models returned (summary)

• TrackInfo: dict with keys: name, artist, album, uri, duration_ms, external_url
• PlaylistInfo: dict with playlist metadata (id, name, description, owner, track_count, uri)

Example (Python)

from pathlib import Path
from mcp_spotify_player.client_auth import load_tokens
from mcp_spotify_player.spotify_controller import SpotifyController

tokens = load_tokens(Path.home() / '.config' / 'mcp_spotify_player' / 'tokens.json')
controller = SpotifyController(lambda: tokens)

if controller.is_authenticated():
    controller.play_music(query='Where Is My Mind by The Pixies')
    state = controller.get_playback_state()
    print(state)

[See implementation in src/mcp_spotify_player/* controllers]

🏗️ Architecture

Below are three conceptual views to help understand the system: a components diagram, the OAuth (PKCE) authorization flow, and a sequence diagram for a typical MCP command (e.g., play_music).

1) Components (logical)

This diagram shows the main modules and how an MCP client (Claude) connects to the MCP stdio server, which delegates to the SpotifyController and underlying clients. Tokens are persisted on disk.

graph TB
  subgraph Client
    C["Claude / MCP client"]
  end

  subgraph Host
    S["MCP stdio server<br>(mcp_spotify_player)"]
    SC["SpotifyController"]
    PC["PlaybackController"]
    PLC["PlaylistController"]
    CL["SpotifyClient"]
    TS["Tokens store<br>(~/.config/.../tokens.json)"]
  end

  C -->|JSON-RPC over stdio| S
  S --> SC
  SC --> PC
  SC --> PLC
  SC --> CL
  CL -->|HTTPS| Spotify["Spotify Web API"]
  CL --> TS

2) OAuth (PKCE) flow

PKCE flow used by the server when running the interactive /auth command. Important note: when the server runs as an MCP stdio process inside a client (e.g., Claude Desktop), the browser callback must be reachable from the same host.

sequenceDiagram
  participant User
  participant Claude as Claude (MCP client)
  participant Server as MCP stdio server
  participant Browser
  participant Spotify
  participant TokenStore as Tokens (filesystem)

  User->>Claude: /auth
  Claude->>Server: auth request (JSON-RPC)
  Server->>Server: generate code_verifier & code_challenge (PKCE)
  Server->>Browser: open authorization URL (Spotify)
  User->>Browser: login & authorize
  Spotify->>Server: redirect to local callback with code
  Server->>Spotify: exchange code + code_verifier -> access/refresh tokens
  Server->>TokenStore: save tokens.json
  Server->>Claude: auth result (success)
  Claude->>User: authentication completed

3) Command sequence (play_music)

Sequence of a typical user command from Claude to Spotify and back.

sequenceDiagram
  participant Claude
  participant MCPServer as MCP stdio server
  participant SpotifyController
  participant PlaybackClient as SpotifyClient (playback)
  participant SpotifyAPI as Spotify Web API

  Claude->>MCPServer: play_music(query="Where Is My Mind")
  MCPServer->>SpotifyController: dispatch play_music(query)
  SpotifyController->>PlaybackClient: search / play (choose track URI)
  PlaybackClient->>SpotifyAPI: POST /v1/me/player/play (uris/context)
  SpotifyAPI-->>PlaybackClient: 204 / playback started
  SpotifyController-->>MCPServer: {"success": true, "message": "Playing: ..."}
  MCPServer-->>Claude: result (JSON-RPC response)

Note: when using Claude (or any MCP client running the server as stdio), ensure the MCP server process and the browser used for PKCE callback run on the same host or that the redirect URI is reachable from that host.

Project structure

mcp-spotify-player/
├── src/
│   ├── mcp_logging/
│   │   └── __init__.py
│   ├── mcp_spotify/
│   │   ├── __init__.py
│   │   └── errors.py
│   └── mcp_spotify_player/
│       ├── __init__.py
│       ├── __main__.py
│       ├── album_controller.py
│       ├── cli.py
│       ├── client_albums.py
│       ├── client_auth.py
│       ├── client_playback.py
│       ├── client_playlists.py
│       ├── client_artists.py
│       ├── config.py
│       ├── mcp_manifest.py
│       ├── mcp_models.py
│       ├── mcp_stdio_server.py
│       ├── playback_controller.py
│       ├── playlist_controller.py
│       ├── artists_controller.py
│       ├── spotify_client.py
│       └── spotify_controller.py
├── pyproject.toml
└── requirements.txt

MCP stdio server

• This project exposes an MCP stdio server for integration with MCP clients (Claude and other MCP-compatible tools).
• Communication protocol: JSON-RPC over stdio.
• Recommended server command for integration and development:

python -m mcp_spotify_player

Run in development mode

mcp-spotify-player
# or
python -m mcp_spotify_player

🔧 Troubleshooting

If you encounter issues when using the MCP server or commands, check these common cases below.

Authentication error

Issue: "Invalid client credentials" or other errors when exchanging the OAuth code

Fix:

1. Verify SPOTIFY_CLIENT_ID and SPOTIFY_CLIENT_SECRET are correct in your .env or MCP server configuration. Ensure those environment variables are available to the process that runs the server (your MCP client must pass env when launching the server).
2. Make sure the Redirect URI registered in the Spotify Dashboard exactly matches (scheme, host and port) the SPOTIFY_REDIRECT_URI used by the server.
3. Check the system clock (a desynchronized clock can cause OAuth token exchange failures).
4. If problems persist, regenerate the credentials in the Dashboard and try again.
5. Inspect server logs for error details (see "Check logs" below).

Cannot control playback

Issue: "No active device found" or API responds that playback/transfer is not possible

Fix:

1. Ensure Spotify is open and there is an active session on some device (desktop app, mobile app or Web Player). Often starting playback manually on the device is enough.
2. Verify the account is Premium — remote playback control via the Web API requires a Premium account.
3. Run the get_devices command (or the equivalent in your MCP client) and confirm which devices are available and which device_id is active.
4. Confirm the authorized scopes include user-modify-playback-state and user-read-playback-state — if not, re-authorize the app with the needed scopes.
5. If needed, start playback on the target device and use the API to transfer playback to it.

Commands not recognized / not working in the MCP client

Issue: Commands sent from the client (e.g. Claude Desktop) have no effect or the client reports errors

Fix:

1. Verify your MCP server configuration in the client: command, args, cwd, and env. Ensure the path to Python and the module (-m mcp_spotify_player) are correct and that necessary environment variables (SPOTIFY_*) are present.
2. Ensure the server process is running. If the server is started as a stdio process, the client must have started it successfully.
3. Check the server output (stdout/stderr). By default logs are printed to the console; to write logs to a file set the MCP_LOG_FILE environment variable.
4. If using Claude Desktop, restart the app after changing the server configuration.
5. Verify tokens exist and are not expired (tokens file at ~/.config/mcp_spotify_player/tokens.json by default) and that the server's token provider is returning valid tokens.

Check logs

• By default logs are sent to standard error (stderr). To save logs to a file, export MCP_LOG_FILE=/path/to/file.log before starting the server.
• Adjust verbosity with MCP_LOG_LEVEL (DEBUG, INFO, WARNING, ERROR).
• When sharing logs for debugging, include only relevant fragments and redact credentials or tokens.

Contribute

Contributions are welcome! Please read our CONTRIBUTING.md before opening a PR.

How to contribute

1. Fork the repository
2. Create a branch for your feature (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to your branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Local development

# Install package in editable mode and project dependencies
python -m pip install -e .

# (Optional) install test/dev dependencies listed in requirements.txt
python -m pip install -r requirements.txt

# Run tests
python -m pytest

# Run linter/formatters if configured (example using ruff)
# python -m pip install ruff
# ruff check .

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

MIT License

Copyright (c) 2025 Víctor Sáez

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction...

🙏 Acknowledgements

• Spotify Web API for providing the API
• Anthropic for the Model Context Protocol
• The community of contributors and users

📞 Contact & Support

• Author: Víctor Sáez
• Email: victor.saez.dev@gmail.com
• Issues: GitHub Issues

---