Home Assistant MCP logo

Home Assistant MCP

Connects coding agents to Home Assistant for live state, diagnostics, automations, dashboards, and controlled service calls.

Install Home Assistant MCP

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "home-assistant-mcp-2": {
      "command": "npx",
      "args": [
        "-y",
        "@vortitron/home-assistant-mcp"
      ],
      "env": {
        "HA_ALLOW_WRITE": "YOUR_VALUE_HERE",
        "HA_ALLOW_CONFIG_WRITE": "YOUR_VALUE_HERE"
      }
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Values are written into the snippet above so you can paste it as is. They never leave your browser.

We couldn’t confirm it starts correctly

The install command installed and started the server, which then stopped for configuration (such as an API key) that our sandbox does not have — expected for servers that connect to your account.

npx -y @vortitron/home-assistant-mcp

No response to initialize.

This is an experimental automated check and can have false negatives — missing environment variables, a slow cold install, etc. It doesn’t necessarily mean something’s wrong. Last checked .

Overview

home-assistant-mcp connects MCP-compatible coding agents directly to Home Assistant, with optional access to the ESPHome dashboard. It lets agents discover entities and devices, inspect live and historical state, render Jinja templates, review logs, trace automations, and view dashboard or camera data. Write operations can call Home Assistant services and modify automations, but are disabled by default and protected by explicit permissions. Reach for it when an agent needs to inspect or operate a real Home Assistant installation instead of relying on pasted entity IDs, YAML, or state values.

Use cases

  • Inspect entity state and attributes before changing an automation
  • Diagnose Home Assistant errors and failed automation runs
  • Call services to control devices from a coding agent
  • Render and test Jinja templates against live state
  • Review dashboard views, camera frames, and energy configuration

Key features

  • Live entity, device, area, and registry inspection
  • Historical state, logbook, and structured log queries
  • Jinja2 template rendering against current state
  • Automation validation, editing, and trace inspection
  • Service calls with returned service responses
  • Dashboard, camera, media, and state-watch tools

Tools80

What Home Assistant MCP lets your agent do. Expand a tool for its parameters.

Self-reported~2.6k tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked

Recent health check succeeded.

11/11 checks healthy over the last 32d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 15d ago via OSV.dev · @vortitron/home-assistant-mcp (npm)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

npm downloads1.8kPackage downloads in the last 30 days.
Last commit7d agoMost recent push to the default branch.
Install checkRunsOur sandbox started it; it then asked for an API key or config.
Tools exposed80Callable tools this server registers over MCP.
Directory activity13 viewsConfig copies, upvotes, and views on AllMCPs.

How Home Assistant MCP works

What home-assistant-mcp does

home-assistant-mcp gives MCP-compatible coding agents an interface to Home Assistant’s configuration, registries, live state, history, logs, automations, dashboards, and services. It can also connect to the ESPHome dashboard as an optional part of the Vome ecosystem. The server is useful for both investigation and controlled changes: an agent can identify the right entity, inspect its attributes, test a template, call a service, and check the result.

The read-oriented tools cover core configuration, areas, devices, helpers, entity metadata, automations, energy settings, logbook entries, and automation traces. Camera tools can return a current image or a reduced RGB grid, while dashboard snapshots combine view state, rendered templates, selected history, and camera frames in one response.

How it works

Agents use MCP tool calls to query the connected Home Assistant instance. Entity searches can be narrowed by domain, area, or free-text terms, and state requests can cover one or more entities. History and logbook queries support time and content filters; a start-only history query continues through the present rather than stopping at Home Assistant’s usual 24-hour boundary.

Templates are rendered against live state through Home Assistant’s Jinja environment. Diagnostic tools expose structured, deduplicated system errors, raw error-log entries, and Supervisor, Core, add-on, or host logs where applicable. When Home Assistant is restarting, the server can wait for it to return and reports restart status instead of exposing a raw gateway or WebSocket failure.

Setup and configuration

Write access is governed by an explicit safety policy. In direct mode, service calls and other write operations require HA_ALLOW_WRITE=true; configuration changes such as automation updates require HA_ALLOW_CONFIG_WRITE. Brokered mode uses VomeHome API-key scopes for per-instance ha:write and ha:config permissions, with client flags acting as optional local restrictions.

The server is described for Cursor, VS Code with Copilot, Claude Desktop, Claude Code, and other MCP-speaking clients. It is also presented as a standalone option for Home Assistant users, apart from the managed VomeHome service.

Tools and capabilities

  • Inspect Home Assistant configuration, entities, areas, devices, registries, helpers, and services.
  • Read current state, historical changes, logbook entries, energy preferences, logs, and automation traces.
  • Render Jinja2 templates against live Home Assistant state.
  • View dashboard snapshots, camera stills, camera pixel frames, and media files when the required Vome capabilities are enabled.
  • Wait for state changes through the Vome component instead of repeatedly polling.
  • Call Home Assistant services and receive service responses where the service supports them.
  • Create or update automations, including triggers, conditions, and actions, with configuration-write permission.
  • Validate configuration and inspect failed automation steps through trace details.

Limitations and notes

Write operations are not enabled by default. Permission requirements vary by operation: service calls, logging changes, and structured-log clearing use the write permission, while automation changes require configuration-write permission. File and helper operations have additional safeguards in the broader toolset, including checks against damaging protected Home Assistant files.

Some capabilities depend on VomeHome or matching Vome add-on and integration versions. Media access requires the ha:files capability, camera access through VomeHome requires Cameras to be enabled, and state watching relies on the Vome component. Supervisor-related log access applies to HAOS or Supervised installations. Camera and dashboard responses also depend on whether the MCP client can display or interpret image data.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Home Assistant MCP

home-assistant-mcp is an MCP server that connects coding agents to Home Assistant and optionally the ESPHome dashboard. Its main tools discover entities and devices, read live or historical state, render Jinja templates, inspect automations and logs, view dashboards and cameras, call services, and make permission-gated automation changes.

Install it through an MCP-compatible client configuration. The project identifies Cursor, VS Code with Copilot, Claude Desktop, and Claude Code as supported clients, and describes a one-command Claude Code setup for its side panes.

Yes, home-assistant-mcp works with Cursor. The server is intended for MCP-speaking clients, and Cursor is specifically named among its supported coding-agent clients.

Set `HA_ALLOW_WRITE=true` for direct-mode write operations, and set `HA_ALLOW_CONFIG_WRITE` for configuration changes such as automation updates. In brokered mode, VomeHome API-key scopes provide the per-instance `ha:write` and `ha:config` permissions.

Yes, it can inspect recent automation runs with `ha_list_traces` and retrieve step-by-step details with `ha_get_trace`. Trace responses identify the step where execution failed through `failed_at`.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewHome Assistant MCP AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/home-assistant-mcp-2?style=directory)](https://allmcps.com/mcp/home-assistant-mcp-2)
HTML Embed
<a href="https://allmcps.com/mcp/home-assistant-mcp-2"><img src="https://allmcps.com/api/badge/home-assistant-mcp-2?style=directory" alt="Home Assistant MCP on AllMCPs" /></a>