Loxone MCP Server
About
An MCP server for Loxone home automation systems, allowing AI assistants to control lights, blinds, sensors, and weather.
Details
- Author
- avrabe
- Categories
- Other
Jump to
Setup
Install Loxone MCP Server in your MCP client (Claude Desktop, Cursor, Windsurf, and others).
Repository: https://github.com/avrabe/mcp-loxone
Follow the installation instructions in the repository README, then restart your MCP client.
Control your Loxone smart home through the Model Context Protocol
An async Rust MCP server that connects AI assistants to Loxone Miniservers. Control lights, blinds, climate, security, audio, and more — all through standardized MCP tools and resources.
[!NOTE] Built on thePulseEngine MCP frameworkv0.17.0 for standardized protocol handling, authentication, and transport layers.
- 🔌17 MCP Tools— Control lights, blinds, HVAC, security, audio, door locks, intercoms, and scenes — all wired to real Miniserver commands
- 📊25+ MCP Resources— Read-only access to rooms, devices, sensors, energy, weather, and system status with live state
- 🚀Three Transports— stdio (Claude Desktop), HTTP/SSE (n8n, web clients), and Streamable HTTP
- 🔐Security by Default— SSL verification on, UUID validation, rate limiting, input sanitization, dev-mode restricted to localhost
- ⚡Async Rust— Connection pooling, intelligent caching, batch operations
- 🧊Nix Flake— Reproducible builds with OpenClaw plugin integration
- 🔑Credential Management— Credential ID system, environment variables, or Infisical vault
# Build cargo build --release # Setup credentials cargo run --bin loxone-mcp-setup -- --generate-id --name "My Home" # Run with Claude Desktop cargo run --bin loxone-mcp-server -- stdio --credential-id <your-id>
git clone https://github.com/avrabe/mcp-loxone cd mcp-loxone cargo build --release
nix build github:avrabe/mcp-loxone # Or run directly nix run github:avrabe/mcp-loxone
docker build -t loxone-mcp . docker run -e LOXONE_HOST=192.168.1.100 \ -e LOXONE_USER=admin \ -e LOXONE_PASS=secret \ -p 3001:3001 loxone-mcp
# Interactive setup — generates a unique credential ID cargo run --bin loxone-mcp-setup -- --generate-id --name "Main House" # Or store manually cargo run --bin loxone-mcp-auth -- store \ --name "Office" --host 192.168.1.100 \ --username admin --password secure123 # Manage credentials cargo run --bin loxone-mcp-auth -- list cargo run --bin loxone-mcp-auth -- test <credential-id>
export LOXONE_HOST="192.168.1.100" export LOXONE_USER="admin" export LOXONE_PASS="password"
export INFISICAL_PROJECT_ID="your-project-id" export INFISICAL_CLIENT_ID="your-client-id" export INFISICAL_CLIENT_SECRET="your-client-secret"
[!TIP] Migrating from environment variables? See theCredential Migration Guide.
Add to yourclaude_desktop_config.json:
{ "mcpServers": { "loxone": { "command": "loxone-mcp-server", "args": ["stdio", "--credential-id", "abc123def-456-789"] } } }
# Standard HTTP/SSE loxone-mcp-server http --port 3001 --credential-id <id> # Streamable HTTP (new MCP Inspector) loxone-mcp-server streamable-http --port 3001 --credential-id <id>
The flake exports anopenclawPluginfornix-openclaw:
customPlugins = [ { source = "github:avrabe/mcp-loxone"; config.env = { LOXONE_HOST = "/path/to/secrets/loxone-host"; LOXONE_USER = "/path/to/secrets/loxone-user"; LOXONE_PASS = "/path/to/secrets/loxone-pass"; }; } ];
AI Assistant (Claude, n8n, OpenClaw) │ MCP Protocol (stdio / HTTP / Streamable HTTP) │ ┌────▼─────────────────────────┐ │ loxone-mcp-server │ │ ┌─────────────────────────┐ │ │ │ Security Layer │ │ │ │ Auth · Rate Limit · TLS │ │ │ ├─────────────────────────┤ │ │ │ Tool Handlers │ │ │ │ 17 tools → send_command │ │ │ ├─────────────────────────┤ │ │ │ Loxone Client │ │ │ │ HTTP · WebSocket · Cache│ │ │ └─────────────────────────┘ │ └────┬─────────────────────────┘ │ HTTP /jdev/sps/io/{uuid}/{cmd} │ Loxone Miniserver │ Physical Devices
- Rust 1.85+ (2024 edition)
- Loxone Miniserver (Gen 1 or Gen 2)
cargo build # Dev build cargo test --lib # Unit tests cargo fmt && cargo clippy -- -D warnings # Format + lint cargo audit # Security audit
# Requires network access to Miniserver LOXONE_LIVE_TEST=1 cargo test \ --test live_miniserver_tests \ --features test-utils -- --nocapture
src/ ├── server/ # MCP protocol, tool handlers (macro_backend.rs) ├── client/ # HTTP/WebSocket clients with UUID validation ├── config/ # Credentials, master key auto-persistence ├── security/ # Input sanitization, rate limiting, CORS ├── monitoring/ # Metrics, dashboards, InfluxDB ├── history/ # Time-series data storage ├── discovery/ # mDNS network discovery └── main.rs # CLI, transport selection, startup
Built onPulseEngineMCP framework — async Rust for the Model Context Protocol
A free, private, and secure remote MCP server for Home Assistant.
MCP server for Google Cast — discover devices, play media, control volume, launch apps, and manage queues over stdio
A safe-by-default MCP server that exposes your Control4 home automation (lights, scenes, locks, thermostats, and media) as structured tools over HTTP and Claude Desktop STDIO for reliable AI-powered control on your local network.
Control and query the status of Ecovacs cleaning robots using the MCP protocol.
Control AVM FRITZ!Box routers - manage devices, WiFi, network settings, parental controls, and schedule time-delayed actions
Allows an LLM agent to control your Gaggimate espresso machine
An MCP server for the Gaggiuino open-source espresso machine, providing real-time local network access to machine status and shot data.
A Model Context Protocol (MCP) server that provides AI assistants with access to Home Assistant, enabling smart home control and automation management.
Read-only MCP (Model Context Protocol) server for Home Assistant. Gives AI assistants (Claude Desktop, LibreChat, Cline) full observability into your smart home — entity states, automations, scripts, devices, logs, diagnostics — without any write access. Also generates static AI context snapshots for RAG systems, ChatGPT Projects, Qwen, and other tools that accept custom knowledge files. Built in Python, runs anywhere — locally, in Docker, or as an MCP integration.
Native macOS HomeKit integration for AI assistants via Model Context Protocol
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
