sonos-mcp
MCP server for controlling Sonos speakers. Currently early scaffolding with a temporary discover tool.
README
sonos-mcp
MCP server for controlling Sonos speakers.
Using it
Connect the server to an MCP client (see Connecting to Claude below) and ask for music in plain language — "play some jazz in the kitchen", "turn the living room down to 20", "DJ a party then wind down with chill vibes". The server exposes these tools:
topology— list speakers on the network, or the current room groupings (run this first to learn valid room names)search— search the Sonos-linked Spotify account for tracks, albums, playlists, or artistsplay_media— play a Spotify URI on a speaker/group (replace the queue, or enqueue)transport— play, pause, stop, skip, and playback statusvolume— get/set volume for a room or a whole bonded groupqueue— list, jump to, remove, or clear queue entriesgrouping— join, unjoin, or dissolve speaker groupsdj— turn free-form vibe strings (e.g.["party", "focus", "chill"]) into a playlist queue: each vibe resolves to the first matching Spotify playlist, played in orderstatus— check that the underlyingsonosCLI is installed and reachable
You'll need Sonos speakers on your local network, the sonos CLI on your PATH, and a Spotify account linked to your Sonos system.
Prefer skipping the server entirely? skills/sonos is a Claude skill that teaches Claude to drive the sonos CLI directly — copy it into ~/.claude/skills/ and it covers the same discover/search/play/queue/volume/grouping workflows from the terminal.
Requirements
- Node.js 20+
- The
sonosCLI installed and onPATH
Development
Install dependencies:
npm install
Generate a self-signed TLS cert for local dev (once, or whenever it expires):
npm run certs
This writes certs/key.pem and certs/cert.pem, which are gitignored.
Run the server (TypeScript, no build step):
npm run dev
The server listens on https://localhost:3000/mcp (override with PORT).
Build and run the compiled output:
npm run build
npm start
Connecting to Claude
The dev server uses a self-signed cert, which Claude's HTTP client rejects by default
(DEPTH_ZERO_SELF_SIGNED_CERT). curl -k bypasses this, but Claude has no such flag, so
each client needs to be told to trust certs/cert.pem explicitly.
Claude Code
Set NODE_EXTRA_CA_CERTS to the cert's path whenever you launch claude:
export NODE_EXTRA_CA_CERTS="$(pwd)/certs/cert.pem"
claude mcp add --transport http sonos-mcp https://localhost:3000/mcp
Add the export to your shell profile if you want it to persist — note it's a global env
var that affects every Node/Bun process you launch from that shell, not just this project,
so it's opt-in rather than something this repo sets for you.
Claude Desktop
Desktop's Settings → Connectors → "Add custom connector" doesn't work for a local dev
server: that flow goes through claude.ai's cloud connector setup, which can't reach
localhost on your machine at all — clicking Add just silently does nothing.
Desktop's local claude_desktop_config.json also can't reference a URL directly; it only
supports spawning local stdio commands. The fix is to bridge HTTP to stdio with
mcp-remote, passing the CA cert via that
process's own env (no global/shell env var needed). Add this to
~/Library/Application Support/Claude/claude_desktop_config.json, under mcpServers:
"sonos-mcp": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://localhost:3000/mcp"],
"env": {
"NODE_EXTRA_CA_CERTS": "/absolute/path/to/sonos-mcp/certs/cert.pem"
}
}
Then fully quit (Cmd+Q) and reopen Claude Desktop to pick up the config change.
Manually testing
The server speaks MCP over Streamable HTTP. With the dev server running, exercise it with curl (the -k flag skips validation of the self-signed cert):
# initialize a session
curl -sk -X POST https://localhost:3000/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"manual-test","version":"1.0.0"}}}'
# list tools
curl -sk -X POST https://localhost:3000/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'
# call the topology tool to list speakers
curl -sk -X POST https://localhost:3000/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"topology","arguments":{"action":"discover"}}}'
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
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.