kef-mcp
MCP server for controlling KEF wireless speakers (LSX II, LS50 Wireless II, LS60) and playing Spotify on them via local network and Spotify Web API.
README
kef-mcp
An MCP server for controlling KEF wireless speakers (LSX II / LS50 Wireless II / LS60) and playing Spotify on them. Exposes tools any MCP client can call to power the speaker on/off, switch inputs, set volume, control transport, and start Spotify playback — with one combined tool that does the whole "wake → switch to wifi → play" flow.
Built to be driven by Poke (see poke/), but it's a
standard MCP server (streamable HTTP) usable by any MCP client.
MCP client ──► kef-mcp ──┬─ HTTP ─► KEF speaker (LAN)
└─ HTTPS ► Spotify Web API
Tools
| Tool | What it does |
|---|---|
play_on_kef(query, volume?) |
Main one. Wake KEF → source=wifi → find on Spotify → play on KEF. |
kef_power_on / kef_standby |
Power on / off. |
kef_set_source(source) |
wifi, bluetooth, tv, optic, coaxial, analog. |
kef_set_volume(level) / kef_get_volume |
Hardware volume 0-100. |
kef_play_pause / kef_next / kef_previous |
Transport control. |
kef_status |
Power, source, volume, now playing. |
spotify_pause / spotify_skip |
Spotify-side control. |
spotify_set_volume(level) |
Volume on the active Spotify device. |
spotify_now_playing |
Track / artist / device. |
Prerequisites
- Spotify Premium (the Web API blocks playback control on Free accounts).
- A KEF LSX II / LS50 Wireless II / LS60 on your LAN, ideally with a DHCP reservation so its IP is stable.
- Docker (to run the server), or just Python 3.12+ to run it directly.
Setup
1. Configure
cp .env.example .env
Fill in .env:
KEF_HOST— the speaker's LAN IP (from the KEF Connect app or your router).KEF_DEVICE_NAME— substring of its Spotify Connect name (e.g.KEF).SPOTIFY_CLIENT_ID/SPOTIFY_CLIENT_SECRET— from a Spotify app (Redirect URI:http://127.0.0.1:8888/callback).MCP_AUTH_TOKEN— a random string clients must send as a Bearer token. Generate:python -c "import secrets; print(secrets.token_urlsafe(32))"
2. Get a Spotify refresh token (once)
pip install requests
export SPOTIFY_CLIENT_ID=... # PowerShell: $env:SPOTIFY_CLIENT_ID="..."
export SPOTIFY_CLIENT_SECRET=...
python auth_spotify.py
Log in, approve, and copy the printed SPOTIFY_REFRESH_TOKEN=... line into .env.
3. Run
Docker (recommended):
docker compose up -d --build
docker compose logs -f kef-mcp
docker-compose.ymlalso runs apoke-tunnelsidecar that forwards the MCP server up to your Poke agent (seepoke/README.md). SetPOKE_TUNNEL_TOKENin.envto enable it (runnpx poke@latest loginonce to get it); it's only needed for Poke.
Or directly:
pip install -r requirements.txt
python server.py # serves streamable HTTP at http://0.0.0.0:8000/mcp
4. Connect a client
- Poke: set
POKE_TUNNEL_TOKENin.env(fromnpx poke@latest login) and rundocker compose up -d— thepoke-tunnelcontainer connects it automatically. Full guide + the recipe:poke/README.md. - Other MCP clients: point them at
http://<host>:8000/mcp. (If you expose the server beyond the tunnel, setMCP_AUTH_TOKENand sendAuthorization: Bearer <token>.)
Sanity checks
KEF reachable from the container:
docker compose exec kef-mcp python -c \
"import asyncio, os; from kef_client import KefClient; \
print(asyncio.run(KefClient(os.environ['KEF_HOST']).get_status()))"
Should print the speaker status (e.g. standby / powerOn).
Notes & troubleshooting
- "KEF Spotify Connect device not found" — the KEF only advertises as a Connect
target once it's awake and has played Spotify since boot.
play_on_kefwakes it first; on a cold boot, cast to it once from the Spotify app, then retry. - Spotify 404 / "No active device" — same cause; ensure
KEF_DEVICE_NAMEmatches the name Spotify shows. - Container can't reach the KEF — verify
KEF_HOSTand that the speaker has a DHCP reservation. - KEF API — writes are
POST /api/setData(JSON body), reads areGET /api/getData. Verified against LSX II / LS50 W II / LS60 firmware. Very old firmware used GET-only; update firmware if writes fail.
Project layout
kef-mcp/
├── server.py # MCP server: defines the tools
├── kef_client.py # KEF local HTTP API client
├── spotify_client.py # Spotify Web API client
├── auth_spotify.py # one-time Spotify OAuth helper
├── Dockerfile
├── docker-compose.yml # server + poke-tunnel sidecar
├── .env.example
├── tunnel/ # headless poke-tunnel container
│ ├── Dockerfile
│ └── entrypoint.sh
└── poke/ # Poke-specific: connection guide + recipe
├── README.md
└── RECIPE.md
Security
.envholds your Spotify secret/refresh token and auth token — gitignored, never commit it.- Anyone with
MCP_AUTH_TOKEN+ the URL can control your speakers and Spotify. Keep it secret; consider a Cloudflare Access policy for remote exposure.
Credits
KEF local-API payload formats verified against pykefcontrol (MIT).
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.