CustomTkinter AI UI Generator

This Flask-based AI application allows you to create beautiful CustomTkinter desktop UIs just by describing them in natural language. Powered by Google's Gemini API, it instantly generates Python code, provides a live preview, and lets you interactively refine your design through conversation.

Features

• Natural Language UI Generation: Describe your UI in plain English (e.g., "make a dark-themed login screen with a password field")
• Live Preview: Instantly see your generated UI in action
• Interactive Refinement: Chat with the AI to tweak your design ("make the button red" or "move it to the right")
• Custom Instructions: Add specific guidelines to direct how the AI builds your UI
• Code Export: Download the complete Python code to use as a foundation for your project

Setup

1. Clone this repository
2. Install the required dependencies:

   pip install -r requirements.txt
   

3. Get a Gemini API key from Google AI Studio
4. Add your API key to the .env file:

   GEMINI_API_KEY=your_api_key_here
   

5. Run the application:

   python app.py
   

6. Open your browser and navigate to http://127.0.0.1:5000

Usage

1. Describe Your UI: Enter a description of the UI you want to create in the input field.
2. Generate UI: Click the "Generate UI" button to create your CustomTkinter interface.
3. Refine Your Design: Use the chat to request changes or improvements to your UI.
4. Add Custom Instructions: Set specific guidelines for how the AI should generate UIs.
5. Export Code: Download the Python code to use in your own projects.

Requirements

• Python 3.7+
• Flask
• Google Generative AI Python SDK
• CustomTkinter
• Pillow
• python-dotenv

Example Prompts

• "Create a dark mode app with a sidebar navigation and main content area"
• "Generate a login form with username and password fields, remember me checkbox, and login button"
• "Make a file browser interface with a tree view and preview pane"
• "Design a settings page with toggles for dark mode, notifications, and sound"

Limitations

• The AI may occasionally generate code that needs minor adjustments
• Complex UI layouts might require iterative refinements
• The preview functionality depends on your local environment setup

License

MIT