This application uses AI to analyze photographs of buildings and identify the presence of Christopher Alexander's 15 fundamental patterns. It provides a web-based interface for capturing or uploading images, which are then processed using advanced AI models to recognize architectural patterns.
The purpose of this project is to make architectural pattern recognition accessible to everyone. Whether you're an architect, a student, or simply someone interested in the built environment, this tool can help you understand and appreciate the fundamental patterns that contribute to wholesome and life-affirming spaces.
- Live camera feed for real-time building capture
- Photo upload functionality
- AI-powered analysis of 15 fundamental architectural patterns
- Detailed results display with explanations
- Frontend: HTML5, vanilla JavaScript, Tailwind CSS
- Backend: Node.js-based API
- Hosting: Vercel
- AI/ML: OpenAI API
- Storage: Vercel Blob Storage
This project is designed to be hosted on Vercel. To deploy your own instance:
- Fork this repository
- Sign up for a Vercel account if you haven't already
- Create a new project in Vercel and link it to your forked repository
- Set up the following environment variable in your Vercel project settings:
OPENAI_API_KEY: Your OpenAI API key
To run this project locally:
- Clone the repository
- Install dependencies with
npm install - Create a
.envfile in the root directory and add your OpenAI API key: OPENAI_API_KEY=your_api_key_here - Run the development server with
npm run dev
Please fork this if you wish to contribute. I dont intend to add new features to this currently.
- None planned. Frankly I am very buusy and this is a prototype.
This project is open source and available under the MIT License.
This project is inspired by the work of Christopher Alexander and his book series "The Nature of Order", as well as a quote from Nikos Salingaros: "At the time of writing, using large language models to evaluate the 15 properties hidden in an image is not very accurate."
Note: This entire project, including this readme, the definitions of the patterns and all the code was written by Anthropic Claude 3.5 sonnet, so thanks to Anthropic for making such a great model, and to cursor for the IDE.
For more information or to get in touch, you can contact me on Twitter @dannyraede or visit my website dannyraede.com.