marcelmarais/spotify-mcp-server

Spotify MCP Server – What It Is

Spotify MCP Server is a small Node‑JS service that implements the Model Context Protocol (MCP) for Spotify. It lets AI assistants (Claude, Cursor, VS Code’s Cline extension, etc.) talk to your Spotify account: they can query what’s playing, search the catalog, read or edit playlists, control playback, adjust volume, and more – all through a standardized set of “tools” exposed over MCP.

Core Features (as described in the README)

Category	Operations (examples)	What they do
Read	`searchSpotify`, `getNowPlaying`, `getMyPlaylists`, `getPlaylistTracks`, `getQueue`, `getAvailableDevices`	Query Spotify’s catalog, fetch the current track, list your playlists, see the playback queue, discover connected devices, etc.
Playback control	`playMusic`, `pausePlayback`, `resumePlayback`, `skipToNext`, `skipToPrevious`, `addToQueue`, `setVolume`, `adjustVolume`	Start/stop music, move through the queue, add items, change volume (Premium required).
Album handling	`getAlbums`, `getAlbumTracks`, `saveOrRemoveAlbumForUser`, `checkUsersSavedAlbums`	Retrieve album metadata, list tracks, save/remove albums from your library.
Playlist handling	`createPlaylist`, `addTracksToPlaylist`, `removeTracksFromPlaylist`, `updatePlaylist`, `reorderPlaylistItems`	Create new playlists, add or delete tracks, edit metadata, reorder items.
User library	`getUsersSavedTracks`, `removeUsersSavedTracks`, `getFollowedArtists`	Read or modify your “Liked Songs” and followed artists.

All tools return structured JSON (IDs, names, status messages) that an AI can consume directly.

How to Get It Running

Prerequisites – Node v26+, a Spotify Premium account, and a Spotify developer app.
Install – Clone the repo, run npm ci, then npm run build.
Create a Spotify app – Obtain a client ID/secret and set a redirect URI.
Configure – Copy spotify-config.example.json to spotify-config.json and fill in the credentials.
Authenticate – Run npm run auth to complete the OAuth flow; tokens are saved in the config file and refreshed automatically.
Start the server – node build/index.js (default stdio transport) or, for HTTP‑based clients, set MCP_TRANSPORT=http and optionally MCP_AUTH=spotify for OAuth‑protected endpoints.

Using It with AI Tools

Claude Desktop – Add a MCP entry pointing to the built index.js.
Cursor – Add the same command in the MCP tab of settings.
VS Code Cline extension – Provide a cline_mcp_settings.json that lists the server command and any auto‑approved tools.

The server can run as a long‑lived HTTP process (port 3000 by default) so multiple clients can share it. When MCP_AUTH=spotify is set, the endpoint follows RFC 9728, letting OAuth‑aware clients obtain a Spotify token themselves, keeping your personal refresh token out of the server.

Who Might Find This Useful?

Developers building AI agents that need to control music playback as part of a broader workflow (e.g., “play background music while I code”).
Power users who want their LLM‑based assistants (Claude, Cursor) to manage playlists or fetch the current song without manual clicks.
Experimenters exploring MCP as a way to expose third‑party services to LLMs.

Bottom Line

Spotify MCP Server is a genuine, open‑source bridge between the Model Context Protocol and the Spotify Web API, providing a rich toolbox for AI assistants to read and manipulate a Spotify account programmatically. It is fully documented, includes OAuth handling, and can be run locally or behind an HTTP proxy for multiple clients.

Repository: marcelmarais/spotify-mcp-server
License / status: Not specified in the README, but the project appears actively maintained (build scripts, Docker notes, integration guides).