Lunar Oasis

Lunar Oasis visualizes real messaging activity as a living pixel-art ecosystem on the moon.
With falling leaves
With raining effect

Inspiration

The Moon has always felt impossibly far away—something we look up at, dream about, but never touch. When we saw the theme "Fly Me to the Moon," we didn't want to just admire it from a distance. We wanted to leave something there.

That's where Lunar Oasis comes from. We thought: what if our everyday conversations could actually grow something on that distant, silent surface? Not just empty messages that vanish, but real, living traces of our connections. So we built a place where every chat you have plants a seed on the Moon—and over time, those seeds turn into a garden. It's our way of making the Moon feel a little closer, and turning "someday" into something you can participate in right now.

What it does

Lunar Oasis is a messaging-driven AI agent that lives inside iMessage. Users chat with it as they normally would. As the conversation deepens, the agent plants and grows virtual plants on a lunar surface shown on a web dashboard. When a conversation reaches a milestone, a special tree is unlocked and becomes a shared memory for the group. It turns ordinary chats into a quiet, collective act of building something on the Moon.

How we built it

We built Lunar Oasis using Photon Spectrum to seamlessly connect our AI agent to iMessage and other real messaging platforms. The backend (Node.js/TypeScript) listens for incoming messages via Photon, updates the conversation state (message count, water level), and triggers plant growth logic. We used a lightweight database to store conversation states and plant data. The frontend web dashboard displays the lunar surface in real time, polling the backend API for updates.

Challenges we ran into

Our biggest challenge was aligning the backend with the frontend's actual behavior. We discovered that the initial backend requirements document was more aspirational than what the frontend code actually implemented. We had to pivot and build strictly to the real on-screen logic—no complex LLM analysis, just a clean state model. Another challenge was integrating Photon's messaging infrastructure, but its terminal provider allowed us to test our agent logic locally before connecting to real iMessage.

Accomplishments that we're proud of

We're most proud that we actually made "building something on the Moon" work. You send a message in iMessage, and a plant really grows on the web dashboard. We connected chat to the lunar surface: as users talk normally, our Photon Agent quietly catches every message and syncs the water level, message count, and plant growth to the frontend in real time. We also fully integrated Photon Spectrum, testing the agent logic locally with its terminal provider before switching to real iMessage.

What we learned

We learned how to build a real-time state synchronization system between a messaging platform and a web UI. We also learned the importance of communication within a team—aligning on what the frontend actually does versus what the documentation says. Most importantly, we learned how to use Photon Spectrum to bridge the gap between AI agents and the platforms people already use.

What's next for Lunar Oasis

We want to expand the ecosystem with more plant types and interactive elements, like letting users "water" plants from the web dashboard. We also plan to support more messaging platforms and add persistent data storage for long-term group memories.