How to Use the Home Assistant MCP in LangChain

Q: How do I connect LangChain to Home Assistant?

Install langchain-mcp-adapters and initialize a MultiServerMCPClient with your HTTP transport URL. Call client.gettools() and pass the resulting array directly into your ReAct agent.

Q: Can LangChain read my Home Assistant logs?

Yes. The getlogbookentries tool gives your agent direct access to recent system events. You can filter this feed by specific entity IDs and time ranges to isolate issues quickly.

Q: Does this MCP server support persistent context?

The connection is stateless by default. You need to use client.session() if you want the agent to remember device states across multiple interactions.

Build multi-step smart home reasoning pipelines by connecting this MCP Server to your LangChain agents.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Home Assistant MCP to LangChain

Create your Vinkius account to connect Home Assistant to LangChain and route execution through our secure gateway. The platform manages server hosting, runtime updates, and security layers. Configuration requires no manual server provisioning.

GDPR Free for Subscribers

Setup Home Assistant with LangChain

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Parse Smart Home State with LangChain

`list_entity_states` feeds your agent a complete snapshot of your connected devices. You get immediate access to every sensor, switch, and climate control metric currently active on your network. The agent reads this array and decides which components require attention before executing the next step in your chain. `get_entity_history` pulls the temporal data needed for complex reasoning. Your LangChain agent can review temperature fluctuations over the last 24 hours to determine if the HVAC system is short-cycling. It passes this historical context down the pipeline for further analysis.

Execute Complex Service Chains

`call_ha_service` is the primary mechanism for manipulating your physical environment through this MCP Server. Your ReAct agent determines the correct domain and service, then fires off a command like turning off all lights when the house is empty. Every action flows through LangSmith, giving you full visibility into execution latency and token usage. `fire_ha_event` triggers custom automations directly from your agent's decision loop. When an LLM detects a specific condition in your database, it pushes a custom event to your local network. The local controller picks up that event and runs the corresponding sequence instantly.

Render Jinja2 Templates Dynamically

`render_ha_template` lets your agent evaluate complex local logic without pulling raw data across the network. The system processes Jinja2 strings natively on your local hardware and returns the evaluated result. This keeps your LangChain memory footprint small while still allowing deep conditional checks. `get_api_status` acts as a mandatory health check before your agent commits to a long-running execution block. If the local network is down, the chain halts early. You avoid wasting API tokens on dead connections.

Setup guide

Set up Home Assistant MCP in LangChain

Prerequisites

Python 3.10+ installed
langchain-mcp-adapters + langgraph packages
Active Vinkius subscription with a valid endpoint token

1

Install dependencies
Run pip install langchain-mcp-adapters langgraph langchain-openai. The MCP adapters package converts MCP tools into native LangChain BaseTool objects.
2

Connect via HTTP transport
Use MultiServerMCPClient with "transport": "http" pointing to your Vinkius endpoint. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com.
3

Create a ReAct agent
Pass the discovered tools to create_react_agent() from LangGraph. The agent automatically routes Home Assistant tool calls through the MCP protocol.
4

Run with any LLM
Swap ChatOpenAI for ChatAnthropic, ChatGoogleGenerativeAI, or any LangChain-compatible model. The MCP tools work identically across all providers.

agent.py

from langchain_mcp_adapters.client import MultiServerMCPClient
from langgraph.prebuilt import create_react_agent
from langchain_openai import ChatOpenAI

async with MultiServerMCPClient({
    "home-assistant-mcp": {
        "transport": "http",
        "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp",
    }
}) as client:
    tools = client.get_tools()

    agent = create_react_agent(
        ChatOpenAI(model="gpt-4o"),
        tools,
    )
    result = await agent.ainvoke({
        "messages": "List recent Home Assistant transactions"
    })
    print(result["messages"][-1].content)

Get your connection token →

Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Home Assistant. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.

Why Choose Vinkius

Vinkius connects your tools to AI with real-time monitoring and automatic cost savings — all from one dashboard.

Connect Home Assistant now

Real-time monitoring

Live

visibility into every interaction

Connect your favorite tools to your AI and see exactly what's happening — every request, every response, in real time.

Built-in savings

60%

lower AI costs

Vinkius compresses data between your apps and your AI automatically. Lower bills every month — no configuration required.

Single dashboard

One

place for every integration

Every tool your AI connects to, managed from a single screen. One account, complete control.

Common questions about Home Assistant MCP in LangChain

Install `langchain-mcp-adapters` and initialize a `MultiServerMCPClient` with your HTTP transport URL. Call `client.get_tools()` and pass the resulting array directly into your ReAct agent.

Yes. The `get_logbook_entries` tool gives your agent direct access to recent system events. You can filter this feed by specific entity IDs and time ranges to isolate issues quickly.

Your agent might be passing the wrong domain or service parameters. Check your LangSmith traces to see exactly what JSON payload the LLM sent to `call_ha_service`. The domain must strictly match local entity types like light or climate.

The connection is stateless by default. You need to use `client.session()` if you want the agent to remember device states across multiple interactions.

Your local schedule remains strictly on your hardware unless an agent explicitly calls `get_calendar_events`. The MCP protocol executes locally, meaning your private appointment strings never leak to the broader internet. Authentication requires a single, scoped token that you control entirely.

Use it with your favorite AI tools

Connect this server to Cursor, Claude, VS Code, and more.

OpenAI Agents SDK sdk-python

Google ADK sdk-python

Pydantic AI sdk-python

Vercel AI SDK sdk-typescript