In 2021, I began a project called Midnight City as a way to learn NFT systems, tokenization, and generative workflows using Python. Over the course of six months, I worked on it daily, drawing, coding, writing, and scripting to better understand how images, data, and narrative could operate together within a generative system.

The project resulted in a collection of 1,000 unique character cards, each tied to an evolving story and world. I created over 650 individual traits in Photoshop, ranging from facial features to clothing, and developed layered composition systems that could be exported and combined programmatically. Using Python, I generated over 50,000 distinct image variations, solving for issues around layering, compatibility, and visual coherence.

Beyond the technical process, the project was also an exploration of storytelling and world-building within a constrained system, where each variation still needed to feel intentional. The scale and intensity of the work led to a period of burnout that lasted close to a year, but it ultimately gave me a deeper understanding of generative systems, creative coding, and how narrative can be embedded within structured image production.

Within the project’s narrative, I introduced a class system. As a way to engage an audience more directly, I built a custom class quiz, which is still available [here] if you wish to explore further.