spotify-mcp
Enables AI assistants to search Spotify and manage playlists, library, and playback via a remote MCP server running on Cloudflare Workers.
README
spotify-mcp
A remote MCP (Model Context Protocol) server for Spotify, running on Cloudflare Workers. It lets an AI assistant search Spotify and manage your playlists, library and playback — including moving tracks around inside playlists.
Auth is handled entirely over the web: the server is its own OAuth provider to
the MCP client, and performs the Spotify OAuth flow upstream. There is no local
binary and no authenticate tool — connecting the server in your MCP client
opens the Spotify consent screen in your browser.
Architecture
McpAgent(SpotifyMCP, a Durable Object) hosts the MCP server and tools, served over Streamable HTTP at/mcp(and legacy SSE at/sse).@cloudflare/workers-oauth-providerwraps the Worker. It issues tokens to MCP clients and stores the upstream Spotify tokens (access + refresh) encrypted in the grantprops.SpotifyHandler(Hono app) implements/authorizeand/callback, driving the Spotify authorization-code flow and showing the consent dialog.- The agent persists a working access token in its Durable Object state and refreshes it automatically using the stored refresh token.
MCP client ──/mcp──▶ OAuthProvider ──▶ SpotifyMCP (Durable Object)
│ │ │
/authorize /callback Spotify Web API
└────── Spotify consent (browser) ────────┘
Tools
| Tool | What it does |
|---|---|
get_me |
Current user profile |
search |
Search tracks / albums / artists / playlists |
list_playlists |
Your playlists |
get_playlist_tracks |
Playlist details + tracks with their positions |
create_playlist |
Create a playlist |
update_playlist_details |
Rename / re-describe / change public state |
add_tracks_to_playlist |
Add tracks, optionally at a specific position |
remove_tracks_from_playlist |
Remove tracks |
reorder_playlist_tracks |
Move a track (or range of tracks) to a new position |
unfollow_playlist |
Unfollow / delete a playlist |
get_saved_tracks / save_tracks / remove_saved_tracks |
Liked-songs library |
get_playback_state |
What's playing, on which device |
control_playback |
play / pause / next / previous (Premium + active device) |
add_to_queue |
Queue a track |
Deploy it yourself
Runs on the Cloudflare Workers free tier.
Prerequisites
- A Cloudflare account and the Wrangler CLI logged in (
npx wrangler login). - A Spotify Developer account.
- Node.js 18+.
0. Get the code
git clone https://github.com/lassejlv/spotify-mcp.git
cd spotify-mcp
npm install --ignore-scripts # --ignore-scripts avoids an unused native build (sharp)
1. Create a Spotify app
At https://developer.spotify.com/dashboard, create an app and note the Client ID and Client Secret. Under Redirect URIs, add your Worker's callback URL:
https://spotify-mcp.<your-subdomain>.workers.dev/callback
(The workers.dev subdomain is shown after the first wrangler deploy. Add the
URI, then deploy again if needed. A custom domain's /callback works too.)
2. Create the KV namespace
The OAuth provider stores grants, tokens and clients in KV.
npx wrangler kv namespace create OAUTH_KV
Copy the returned id into wrangler.jsonc, replacing the existing OAUTH_KV
id value (the checked-in one belongs to the original author's account and won't
work for you).
3. Set secrets
npx wrangler secret put SPOTIFY_CLIENT_ID # your Spotify Client ID
npx wrangler secret put SPOTIFY_CLIENT_SECRET # your Spotify Client Secret
npx wrangler secret put COOKIE_ENCRYPTION_KEY # any random string, e.g. `openssl rand -hex 32`
Optional — restrict who can use the server:
# Comma-separated allowlist of Spotify account emails. Only these accounts can
# authorize; everyone else is rejected at the callback (no token issued).
# Leave unset to allow any Spotify account.
npx wrangler secret put ALLOWED_EMAILS # e.g. me@example.com,friend@example.com
4. Deploy
npx wrangler deploy
Your server is now live at https://spotify-mcp.<your-subdomain>.workers.dev/mcp.
Connect an MCP client
Point any remote-MCP-capable client at the /mcp URL. Clients that only speak
stdio can bridge via mcp-remote:
{
"mcpServers": {
"spotify": {
"command": "npx",
"args": ["mcp-remote", "https://spotify-mcp.<your-subdomain>.workers.dev/mcp"]
}
}
}
On first connect, your browser opens: approve the MCP client, then log in and grant Spotify access. Tokens are refreshed automatically thereafter.
Local development
cp .dev.vars.example .dev.vars # then fill in the three values
npx wrangler dev
wrangler dev simulates KV locally. Use http://localhost:8788/callback as an
additional Spotify redirect URI for local testing.
Notes
- Tracks and playlists can be referenced by bare ID or full
spotify:URI in any tool. reorder_playlist_tracksuses zero-based positions; callget_playlist_tracksfirst to see current positions.- Playback control endpoints require Spotify Premium and an active device.
- Requested scopes: playlist read/modify (public + private), library read/modify,
playback read/modify,
user-read-private, anduser-read-email(used for theALLOWED_EMAILSaccess gate). - Access control: set the
ALLOWED_EMAILSsecret to a comma-separated list to restrict the server to specific Spotify accounts; leave it unset to allow anyone.
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.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.