lg-thermav
LG Therma V MCP server to communicate with heat pump via modbus
claude mcp add --transport stdio marinx-lg-thermav-mcp-server docker run -i marinx/lg-thermav-mcp-server \ --env MODBUS_URL="tcp://<modbus-ip>:<port>"
How to use
The LG Therma V MCP Server exposes Modbus-enabled control of LG Therma V heat pumps via the MCP (Model Context Protocol). It enables automation tools and MCP clients to query status, read and set parameters, and respond to events from a Therma V system over a Modbus TCP connection. By running the server (for example in a container), you can connect MCP clients and Claude Desktop-style tooling to issue commands, retrieve device state, and orchestrate automation flows that integrate with other devices in your smart home or building management stack.
Once running, you can use your MCP client or Claude-like tooling to browse the available LG Therma V tools exposed by the server, send queries, and apply control actions such as temperature setpoints, operation modes, and fault checks. The server is designed to be hardware-agnostic in its interface, focusing on MCP-style requests and responses to simplify integration for automation pipelines and local testing.
How to install
Prerequisites:
- A Modbus TCP connection to the LG Therma V heat pump (on the same network).
- Docker installed on the host (recommended for simplicity and portability).
- Optional: an MCP client (e.g., Claude Desktop) or any software that can speak MCP to interact with the server.
Installation steps:
-
Pull the MCP server image from the registry: docker pull marinx/lg-thermav-mcp-server
-
Run the server (adjust Modbus URL to point to your Therma V device): docker run -i -e MODBUS_URL=tcp://192.168.1.100:4196 marinx/lg-thermav-mcp-server
-
Verify the server is up by connecting an MCP client and listing available tools or querying status.
Optional: Build from source (if you prefer to run natively in Go and have Go tooling):
- Prerequisites: Go 1.20+ and Git
- Clone the repository: git clone https://github.com/MarinX/lg-thermav-mcp-server.git
- Change to the directory and build: cd lg-thermav-mcp-server go build -o lg-thermav-mcp-server ./...
- Run the binary locally (ensure MODBUS_URL is set): MODBUS_URL=tcp://<modbus-ip>:<port> ./lg-thermav-mcp-server
Note:
- The server is intended for development and experimentation; adjust Modbus URL and networking rules to fit your environment.
Additional notes
Environment variables and configuration:
- MODBUS_URL: The Modbus TCP endpoint for the LG Therma V unit (e.g., tcp://192.168.1.10:4196).
- If running via Docker, pass MODBUS_URL as an environment variable or set it at runtime. Common issues:
- Cannot connect to the Modbus device: verify network reachability, firewall rules, and correct Modbus port.
- MCP client cannot discover tools: ensure the server is running and accessible at the MCP endpoint; check logs for MCP handshake messages.
- In early development iterations, API surface may change; monitor the repository for updates and pin versions when deploying to production. Configuration tips:
- Never expose Modbus endpoints directly to untrusted networks; prefer a secured network or VPN between your MCP clients and the gateway.
- Use container orchestration (Docker Compose, Kubernetes) to manage restarts and health checks for the MCP server.
Related MCP Servers
trpc-agent-go
trpc-agent-go is a powerful Go framework for building intelligent agent systems using large language models (LLMs) and tools.
station
Station is our open-source runtime that lets teams deploy agents on their own infrastructure with full control.
tiger-cli
Tiger CLI is the command-line interface for Tiger Cloud. It includes an MCP server for helping coding agents write production-level Postgres code.
gopls
MCP server for golang projects development: Expand AI Code Agent ability boundary to have a semantic understanding and determinisic information for golang projects.
kubernetes
A Model Context Protocol (MCP) server for the Kubernetes API.
gcp-cost
💰 An MCP server that enables AI assistants to estimate Google Cloud costs, powered by Cloud Billing Catalog API and built with Genkit for Go