Homekit is an open bridge that gives AI agents direct physical control over Apple Home devices via the Model Context Protocol (MCP).

Stars

69

7-day growth

No data

Forks

2

Open issues

0

License

MIT

Last updated

2026-07-27

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It is the first open bridge between Apple's HomeKit framework and the AI agent layer, enabling agents to control home devices without needing human-oriented tap-to-toggle interfaces.

Who it is for

  • AI developers building agentic systems
  • Apple Home users seeking enhanced automation
  • Developers using MCP-compatible clients (Claude, Cursor, Windsurf)
  • Home automation enthusiasts

Use cases

  • Agent-assisted home control via natural language commands (e.g., 'turn off the kitchen lights before the meeting')
  • Creating and activating scenes dynamically using AI agents
  • Automating routines and triggering existing automations from an agent
  • Managing multiple Apple Homes and accessory states programmatically

Strengths

  • Compatible with the Model Context Protocol, enabling use with multiple popular AI clients
  • Provides eight structured tools for listing, getting, setting accessories, scenes, and automations
  • Offers both a CLI and an MCP server, with JSON output for easy integration
  • Open-source under MIT license with active CI and CodeQL security analysis

Considerations

  • Requires macOS 13 Ventura or later and a configured Apple Home on that Mac
  • Initial setup requires the Homekit macOS App for authorization
  • Limited to the Apple Home ecosystem; does not support non-HomeKit smart home devices

README quick start

the bridge between AI Agents & Reality.

Homekit gives any AI agent direct, physical control over Apple Home. Flip lights. Lock doors. Read sensors. Through a single open interface.

Awaken · Understand · Witness · Install · Begin · Website · Docs · Changelog


✦ Awaken

Apple Home is powerful. But it's locked.

Locked behind apps. Behind Siri. Behind tap-to-toggle UI designed for humans, not agents. Every AI model in 2025 can reason, plan, and act — but the moment you ask it to "turn off the kitchen lights before the meeting", it hits a wall.

Homekit tears that wall down.

It exposes your entire Apple Home — every light, lock, thermostat, sensor, scene, and automation — as structured tools that any AI agent can call. The Model Context Protocol becomes the language. Your home becomes the body.

This is the first open bridge between Apple's HomeKit framework and the agent layer.


✦ Understand

Architecture

┌─────────────────────────────────────────────────────────┐
│                    AI Agent Layer                        │
│         Claude · ChatGPT · Cursor · Windsurf            │
│              Any MCP-compatible client                   │
└────────────────────┬────────────────────────────────────┘
                     │  Model Context Protocol (MCP)
                     │  8 structured tools
                     ▼
┌─────────────────────────────────────────────────────────┐
│                  homekit-mcp server                      │
│           stdio transport · JSON-RPC 2.0                 │
└────────────────────┬────────────────────────────────────┘
                     │  Unix socket / IPC
                     ▼
┌─────────────────────────────────────────────────────────┐
│              Homekit macOS App                           │
│         Native HomeKit framework bridge                  │
│     Authorization · Accessory state · Scene control     │
└────────────────────┬────────────────────────────────────┘
                     │  Apple HomeKit Framework
                     ▼
┌─────────────────────────────────────────────────────────┐
│                   Apple Home                             │
│   Lights · Locks · Thermostats · Sensors · Cameras      │
│          Scenes · Automations · Multi-Home               │
└──────────────────────────────────────────

Description

Homekit gives any AI agent direct, physical control over Apple Home. Flip lights. Lock doors. Read sensors. Through a single open interface.

Related repositories

Similar projects matched by category, topics, and programming language.

0xwilliamortiz
Featured
0xwilliamortiz GitHub avatar

openclaude-improved

OpenClaude is an open-source CLI coding agent that runs on any platform and supports a wide range of LLM providers, offering the same tools and workflows as Claude Code.

AI & Machine LearningLarge Language Models
577
OneInterface
Featured
OneInterface GitHub avatar

stormy-cookbook

Stormy Cookbook provides copy-pasteable recipes for using the Stormy Social Data API and MCP server, which offers a unified REST API and MCP tools to search profiles, posts, and emails across TikTok, YouTube, Instagram, LinkedIn, X, and Reddit through a single HTTP contract.

AI & Machine LearningLarge Language Models
72
makecindy
Featured
makecindy GitHub avatar

cindy

Cindy is an open-source AI agent that runs locally on your machine, integrates multiple AI harnesses and models, and provides memory, skills, and automation to perform real work in your projects and apps.

AI & Machine LearningLarge Language Models
958