Instead of configuring APIs, tokens, or developer connectors, users now link their Homey account to ChatGPT with a single click and immediately start giving plain-language instructions . The shift matters because it targets a much wider audience: households that already own smart devices but never wanted to learn flow logic, zone management, or automation editors.

What you can actually do with the Homey ChatGPT app

Once connected, you can manage nearly every part of your Homey setup through ordinary text prompts. Athom and early coverage have confirmed these capabilities:

• Check device status and control connected hardware — ask whether a specific light is on, or tell ChatGPT to turn off all downstairs plugs .
• Switch devices on and off — control individual lights, switches, and other connected gear without opening the Homey mobile app .
• Create and update Flows and Advanced Flows — describe the automation you want in plain English and let ChatGPT build it. Both standard Flows and more complex Advanced Flows are supported .
• Set and trigger Moods — activate preconfigured scenes or adjust several devices at once with a single sentence .
• Rename devices and reorganize zones — fix naming mistakes or move a device from "Living Room" to "Kitchen" without digging through settings menus .
• Customize your entire Homey configuration — make broader setup changes by describing the result you want, rather than navigating through the web app .

All of this runs inside ChatGPT using natural language. Athom designed the integration so that users do not need any prior technical knowledge or manual configuration beyond the initial account link .

Which Homey setups are supported

The native ChatGPT app works across the full current Homey ecosystem. Athom explicitly lists support for :

• Homey Cloud — the free cloud-based service that doesn't require any hub hardware
• Homey Pro — the flagship smart home hub with local processing and support for Zigbee, Z-Wave, and other local protocols
• Homey Pro Mini — the compact version of Homey Pro
• Homey Self-Hosted Server — for users who want to run Homey on their own hardware

Homey platforms connect to devices through multiple protocols including Matter, LAN APIs, cloud APIs, Zigbee, Z-Wave, and Bluetooth, depending on which hardware or service tier you use . Any device already visible in your Homey setup becomes controllable through the native ChatGPT app once your Homey account is linked .

How the new ChatGPT app differs from the older MCP Server

Before the native ChatGPT app, advanced Homey-AI integration went through the Homey MCP Server. MCP — the Model Context Protocol — is a standard that lets AI agents interact with external services securely .

The earlier route required users to manually configure a connector in ChatGPT's developer settings, supply API tokens, and understand the difference between local and cloud access . While powerful, it was clearly aimed at developers and early adopters comfortable with command lines, environment variables, and token authentication.

Athom itself acknowledged the contrast when it launched the native app. The new integration builds on the same underlying MCP infrastructure but eliminates the manual setup step entirely .

In practical terms, the difference comes down to audience:

• The native ChatGPT app is the consumer-friendly, one-click option for everyday Homey users who want to talk to their home without reading documentation .
• The MCP Server remains the right tool for developers, power users, and anyone building custom AI workflows involving Claude Desktop, Cursor, Windsurf, VS Code Copilot, JetBrains AI, or other MCP-compatible platforms .

This two-track approach mirrors what's happening across the broader smart home and AI platform landscape. MCP server registries now list over 9,400 servers, and major coding environments have adopted the protocol as a standard way to connect AI to external tools . Athom is keeping Homey present in both worlds without forcing regular users to learn the infrastructure.

AI platform support beyond ChatGPT

For people who prefer Claude, Mistral, or custom developer tools, the Homey MCP Server remains available . Several community and official MCP implementations for Homey exist:

• Athom's own Homey MCP Server, announced in late 2025, lets ChatGPT, Claude, and other MCP-compatible AI assistants query device state, trigger Flows, Advanced Flows, and activate Moods .
• Community-built MCP servers enable Claude Desktop to control Homey Pro devices over the local API using natural language, with setup requiring Python and a Homey Bearer token .
• Broader MCP directories list Homey implementations that expose devices, zones, and logic through the protocol for use in agent workflows and LLM applications .

These options are not competitors to the new ChatGPT app; they serve a different user who wants programmable, composable AI integrations rather than a zero-config conversation interface.

Where this fits in Athom's long-term strategy

Athom's original Kickstarter for Homey carried the tagline "Talk to your home" and described the product as speech-controlled home automation for lights, music, climate, and TV . The native ChatGPT app is a direct continuation of that vision, but updated for a world where large language models can do far more than recognize preset voice commands.

The integration turns ChatGPT into a primary interface for Homey — not just a supplementary skill — and lets users configure and control their home the same way they'd ask a knowledgeable roommate to handle something. Athom's bet is that conversational AI, stripped of technical setup steps, will feel more natural to most households than traditional app-based automation editors .

The company's parallel investment in MCP infrastructure suggests they're also preparing for an ecosystem where multiple AI assistants can talk to Homey. The ChatGPT app is the first mass-market product; the MCP Server stack is the underlying platform play.

Instead of configuring APIs, tokens, or developer connectors, users now link their Homey account to ChatGPT with a single click and immediately start giving plain-language instructions . The shift matters because it targets a much wider audience: households that already own smart devices but never wanted to learn flow logic, zone management, or automation editors.

What you can actually do with the Homey ChatGPT app

Once connected, you can manage nearly every part of your Homey setup through ordinary text prompts. Athom and early coverage have confirmed these capabilities:

• Check device status and control connected hardware — ask whether a specific light is on, or tell ChatGPT to turn off all downstairs plugs .
• Switch devices on and off — control individual lights, switches, and other connected gear without opening the Homey mobile app .
• Create and update Flows and Advanced Flows — describe the automation you want in plain English and let ChatGPT build it. Both standard Flows and more complex Advanced Flows are supported .
• Set and trigger Moods — activate preconfigured scenes or adjust several devices at once with a single sentence .
• Rename devices and reorganize zones — fix naming mistakes or move a device from "Living Room" to "Kitchen" without digging through settings menus .
• Customize your entire Homey configuration — make broader setup changes by describing the result you want, rather than navigating through the web app .

All of this runs inside ChatGPT using natural language. Athom designed the integration so that users do not need any prior technical knowledge or manual configuration beyond the initial account link .

Which Homey setups are supported

The native ChatGPT app works across the full current Homey ecosystem. Athom explicitly lists support for :

• Homey Cloud — the free cloud-based service that doesn't require any hub hardware
• Homey Pro — the flagship smart home hub with local processing and support for Zigbee, Z-Wave, and other local protocols
• Homey Pro Mini — the compact version of Homey Pro
• Homey Self-Hosted Server — for users who want to run Homey on their own hardware

Homey platforms connect to devices through multiple protocols including Matter, LAN APIs, cloud APIs, Zigbee, Z-Wave, and Bluetooth, depending on which hardware or service tier you use . Any device already visible in your Homey setup becomes controllable through the native ChatGPT app once your Homey account is linked .

How the new ChatGPT app differs from the older MCP Server

Before the native ChatGPT app, advanced Homey-AI integration went through the Homey MCP Server. MCP — the Model Context Protocol — is a standard that lets AI agents interact with external services securely .

The earlier route required users to manually configure a connector in ChatGPT's developer settings, supply API tokens, and understand the difference between local and cloud access . While powerful, it was clearly aimed at developers and early adopters comfortable with command lines, environment variables, and token authentication.

Athom itself acknowledged the contrast when it launched the native app. The new integration builds on the same underlying MCP infrastructure but eliminates the manual setup step entirely .

In practical terms, the difference comes down to audience:

• The native ChatGPT app is the consumer-friendly, one-click option for everyday Homey users who want to talk to their home without reading documentation .
• The MCP Server remains the right tool for developers, power users, and anyone building custom AI workflows involving Claude Desktop, Cursor, Windsurf, VS Code Copilot, JetBrains AI, or other MCP-compatible platforms .

This two-track approach mirrors what's happening across the broader smart home and AI platform landscape. MCP server registries now list over 9,400 servers, and major coding environments have adopted the protocol as a standard way to connect AI to external tools . Athom is keeping Homey present in both worlds without forcing regular users to learn the infrastructure.

AI platform support beyond ChatGPT

For people who prefer Claude, Mistral, or custom developer tools, the Homey MCP Server remains available . Several community and official MCP implementations for Homey exist:

• Athom's own Homey MCP Server, announced in late 2025, lets ChatGPT, Claude, and other MCP-compatible AI assistants query device state, trigger Flows, Advanced Flows, and activate Moods .
• Community-built MCP servers enable Claude Desktop to control Homey Pro devices over the local API using natural language, with setup requiring Python and a Homey Bearer token .
• Broader MCP directories list Homey implementations that expose devices, zones, and logic through the protocol for use in agent workflows and LLM applications .

These options are not competitors to the new ChatGPT app; they serve a different user who wants programmable, composable AI integrations rather than a zero-config conversation interface.

Where this fits in Athom's long-term strategy

Athom's original Kickstarter for Homey carried the tagline "Talk to your home" and described the product as speech-controlled home automation for lights, music, climate, and TV . The native ChatGPT app is a direct continuation of that vision, but updated for a world where large language models can do far more than recognize preset voice commands.

The integration turns ChatGPT into a primary interface for Homey — not just a supplementary skill — and lets users configure and control their home the same way they'd ask a knowledgeable roommate to handle something. Athom's bet is that conversational AI, stripped of technical setup steps, will feel more natural to most households than traditional app-based automation editors .

The company's parallel investment in MCP infrastructure suggests they're also preparing for an ecosystem where multiple AI assistants can talk to Homey. The ChatGPT app is the first mass-market product; the MCP Server stack is the underlying platform play.