Inspiration

I was inspired by a simple question:

Can an environment reveal something about us without ever asking who we are?

Most personality tests rely on direct questions, but people often answer based on how they see themselves—or how they wish to be seen. I wanted to explore a different approach: a living world that quietly observes behavior instead of collecting answers.

As a Product Designer working in educational technology, I’m constantly thinking about how digital experiences can support learning and self-reflection. The Garden Remembers is my attempt to apply that mindset to a more emotional and experimental experience.

Nature became the perfect metaphor. Just as a real garden changes over time in response to care and interaction, this garden remembers every step the player takes.


What it does

The Garden Remembers is an interactive web experience where players explore a mysterious garden filled with ambiguous situations and subtle choices.

Rather than tracking right or wrong answers, the experience observes how players move through the world—how they explore, hesitate, return, take risks, or respond to unexpected moments.

The current journey includes a lost wallet, a paper threshold, a quiet shelter, and a social circle. The later two share one open field, so players can approach either situation, pass between them, redirect their attention, or return.

The garden changes in response to those interactions. Pigment travels through the landscape, planting and materials shift, and earlier choices echo later in the journey. Instead of telling players who they are, the experience encourages them to reflect on how they behave.

There are no points, diagnoses, moral labels, or “correct” endings.


How I built it

I’m a Product Designer, not a software engineer, and I had virtually no experience building an interactive web experience like this before.

What I did have was curiosity, determination, and a willingness to learn by building.

Using Codex as my development partner, I transformed an early HTML prototype into an immersive first-person WebGL experience. Codex helped me bridge the gap between design and implementation, allowing me to iterate on architecture, interactions, animation, gameplay, accessibility, and testing while I focused on the creative vision and user experience.

The project is built with HTML, CSS, JavaScript modules, Three.js, WebGL, and Web Audio. I separated movement, encounters, consequences, memory, world composition, and release quality into testable systems, and the current automated suite contains 143 passing checks.

GPT-5.6 helped me reason across the whole experience: comparing interaction branches for equal depth, finding input and state edge cases, and translating creative rules into regression tests. The creative direction remained human-led, and the model is not used at runtime or given player data.

This project wouldn’t exist without AI—not because it replaced creativity, but because it empowered me to build something I couldn’t have built alone.


Challenges I ran into

The biggest challenge was designing a system that feels personal without feeling invasive.

I wanted players to feel understood—not judged.

Finding the balance between meaningful behavioral observation and respecting the player’s own interpretation became the core design challenge. Every path needed to feel equally considered, without framing one response as better than another.

From a technical perspective, I also had to learn as I built. Since I don’t come from a software engineering background, almost every technical decision was new territory. Codex became both a collaborator and a teacher throughout the process.


Accomplishments that I’m proud of

  • Transforming the idea of a psychological assessment into an immersive interactive experience.
  • Building a functional experience despite having no professional development background.
  • Using AI as a creative collaborator rather than only as a coding assistant.
  • Designing an experience that encourages reflection instead of classification.
  • Combining my experience in educational technology with interactive storytelling in a completely new way.
  • Building automated tests that protect both technical behavior and the project’s non-judgmental design principles.

What I learned

I learned that behavior often says more than answers.

Two people can make the same decision for completely different reasons, and those differences emerge through movement, hesitation, exploration, and timing—not just through the final choice.

I also learned that memory feels more convincing when it returns later and slightly transformed. An immediate reaction shows that a system registered an action; an echo encountered somewhere else makes the world feel as if it retained something.

On a personal level, this project completely changed how I think about AI-assisted development. As a designer with very limited programming experience, I realized that AI can dramatically lower the barrier between having an idea and bringing it to life.


What’s next for The Garden Remembers

This is only the beginning.

I want The Garden Remembers to evolve into a truly adaptive living world—one that remembers players across multiple visits, introduces new situations dynamically, and creates increasingly personal experiences over time.

My long-term vision is to explore how AI can support reflection, learning, and emotional growth—not by telling people who they are, but by helping them notice patterns in their own behavior.

The Garden doesn’t tell you who you are. It helps you notice who you’ve been.

Built With

  • gpt-5.6
Share this project:

Updates