Open media
Start with an idea for your home: a useful widget, a new automation, or a service that connects the two. With HomeGenie's AI assistants, you can describe what you want, try the generated result, and refine it through conversation or directly in the integrated editor.
You do not need to prepare a separate development environment to begin. HomeGenie brings the chat, program editor, and widget preview together, so you can move from a description to something you can test.
Build through conversation
Vibe Coding starts with describing how you want something to look and behave. The assistant can propose the code and styling, while you try the result and decide what to change next.
Gemini Genie brings automation programs and widgets into the same chat. Lailama offers these capabilities with local models, whose results depend on the model and the resources available on your computer. Through the capabilities you enable, an AI Genie can help you:
- Create and update programs in C#, Python, or JavaScript, and deploy them in HomeGenie.
- Coordinate tasks involving programs, devices, and services.
- Build custom widgets with HTML, CSS, and JavaScript.
- Revise code after compilation errors, using feedback supplied by HomeGenie to attempt a correction.
Generated code still needs testing. Compilation checks catch some problems; trying the program or widget shows whether it actually behaves as intended.
A workspace for trying ideas
The examples in this section are conversational case studies. They show the initial request, what the assistant produced, and how the result was refined—including changes made by hand.
- An integrated editor: Create and preview widgets without setting up a separate Node.js or Webpack toolchain. Configure your Genie first; individual projects may also need devices, services, or libraries.
- Live previews: Try generated HTML, CSS, and JavaScript in HomeGenie, then check how the widget behaves on your dashboard.
- Iterative design: Ask for changes such as “Make the buttons larger”, “Add keyboard controls”, or “Change the color scheme”. You can also edit the code yourself, combining conversation with direct adjustments.
New to HomeGenie AI? Follow the Cloud AI guide to configure Gemini Genie with an API key, or the Local AI guide to set up Lailama with a local model. Coding and multi-step tasks require a model capable of handling them.
How to use this section
Explore how each project connects its interface, state, and backend logic. You can follow the conversation, inspect the generated code, and use the result as a starting point for your own ideas.
The case studies include downloadable resources such as widget archives, automation programs, and Chat History JSON files. Check the resources listed in each article. Importing a chat lets you review the conversation and continue experimenting, but it does not install the project's widgets, programs, or other dependencies; import or configure those separately.
Some examples record earlier experiments with Automan and Widget Genie. Their conversations and downloads preserve that original work. For a new project, use the current Gemini Genie or Lailama capabilities described in the guides above; older examples may need adjustments to work with the current APIs.