DEV Community

Cover image for πŸͺ ExoPlanet Companion β€” Discover Alien Worlds with AI πŸ‘½
Onwubualili Chinemerem
Onwubualili Chinemerem

Posted on

πŸͺ ExoPlanet Companion β€” Discover Alien Worlds with AI πŸ‘½

DEV Weekend Challenge: Earth Day

This is a submission for Weekend Challenge: Earth Day Edition

What I Built

Earth Day is a reminder of how precious our planet is β€” but it also made me wonder: what are the other planets out there like?
That question became ExoPlanet Companion, an AI-powered space exploration web app that lets anyone explore over 6,000 real confirmed exoplanets from NASA's database β€” and experience each one like never before.
Here's what the app does:
πŸ”­ Pulls live data from NASA's Exoplanet Archive β€” real planetary science, including radius, mass, orbital period, surface temperature, and distance from Earth
πŸ€– Uses open-source AI models via OpenRouter to transform that raw data into vivid, imaginative planet narratives β€” describing what the planet looks like, what the sky looks like from its surface, whether life could exist there, and one jaw-dropping fun fact
πŸ““ Maintains a persistent Explorer Journal powered by Backboard β€” every planet you visit gets saved to your personal journal, building a living log of your journey across the cosmos
🌠 Delivers an immersive dark space UI β€” animated starfield, glowing planet illustrations, and a cinematic exploration experience
The goal was simple: make the vastness of the universe feel personal, explorable, and emotional β€” in honor of the planet that started it all.

Demo

πŸš€ Live App: https://project-p0jrj-git-main-2chinemerems-projects.vercel.app/

πŸ“½οΈ Demo Walkthrough:

Code

https://github.com/2Chinemerem/exoplanet-companion

How I Built It

Architecture Overview

The application is built using a React frontend, a lightweight Express proxy server, and three core service layers responsible for handling NASA data, AI generation, and persistent memory.

At a high level, the system works by fetching real exoplanet data, enriching it with AI-generated simulations, and storing user interactions as persistent memory.

NASA Exoplanet Data Layer

The app integrates with NASA’s Exoplanet Archive, which provides access to thousands of confirmed exoplanets along with their physical and orbital characteristics.

I focused on extracting key attributes such as radius, mass, orbital period, equilibrium temperature, and stellar properties. These values form the scientific foundation for every planet simulation.

One key challenge was that NASA’s API does not support browser-based requests due to CORS restrictions. To solve this, I implemented a lightweight Express proxy server that securely relays requests from the frontend to NASA’s API.

AI Simulation Engine

To transform raw scientific data into meaningful experiences, I integrated open-source language models via OpenRouter.

Each planet’s data is passed into a structured prompt that guides the AI to generate a consistent, multi-part simulation β€” including appearance, atmospheric conditions, sky description, potential for life, and a unique insight.

A key design decision here was enforcing structured outputs. This ensures the AI responses are predictable and can be cleanly mapped to UI components, preventing inconsistencies and improving reliability.

Backboard β€” The Explorer Journal ⭐

Backboard plays a central role in making the experience stateful and personal.

When a user starts a session, a new memory thread is created. As users explore planets, each interaction is stored as a memory entry within that thread. This builds a persistent Explorer Journal that tracks every planet visited.

The use of automatic memory handling allows the system to intelligently retain meaningful exploration data without manual filtering, creating a scalable and context-aware memory layer.

This transforms the app from a simple viewer into a personalized exploration experience that evolves over time.

Key Technical Decisions

  1. Use of open-source AI models: Chosen for flexibility, cost-efficiency, and scalability without reliance on paid APIs
  2. Structured AI outputs: Ensured consistent UI rendering and minimized parsing errors
  3. Proxy-based architecture: Solved CORS limitations while keeping the system lightweight
  4. Persistent memory with Backboard: Enabled a continuous user journey instead of a stateless experience.

ExoPlanet Companion turns raw astronomical data into something human β€” a tool that makes space exploration not just informative, but deeply immersive and personal!

Prize Category

βœ… Best Use of Backboard

Top comments (0)