About
The story behind the project
Where it started
This project began as a personal experiment: use AI to generate original artwork and display it on a monitor at home. The idea was simple — wake up to something new on the wall every day, created entirely by a machine with a little creative direction. What started as a fun weekend hack quickly became something worth sharing.
What it does
A Node.js service running on a Raspberry Pi generates new images automatically throughout the day using prompts that rotate through themes, moods, and styles. Each image is created using OpenAI's DALL-E 3 model, uploaded to cloud storage, and added to the gallery — all without any manual intervention. This site is the public face of that process: a live, ever-growing collection of AI-generated art.
The tech stack
- Image generationOpenAI DALL-E 3
- GeneratorNode.js with node-cron, running on a Raspberry Pi
- Storage & DBSupabase (PostgreSQL + object storage)
- Web appNext.js 16 with TypeScript and Tailwind CSS
- HostingVercel
Why share it?
AI-generated art sits in an interesting place — it's creative, surprising, and sometimes genuinely beautiful, but it's also easy to dismiss as "just a machine." This gallery is an honest look at what these models produce day after day, without cherry-picking the best results. Some images are stunning. Some are weird. All of them are real outputs from the system, shared as-is.
It's also a portfolio project — a full end-to-end system built from scratch, from the Raspberry Pi running in the background to the site you're reading now.
Built with help from
The developer built the foundation, then partnered with GitHub Copilot (powered by Claude Sonnet) to bring everything else to life — the gallery, the prompt engine, the deployment pipeline. What would have taken months got built in days. Honestly, it's the perfect team for this project: AI wrote the code, and AI makes the art.