This is a submission for Weekend Challenge: Earth Day Edition
🌌 What I Built
Most climate apps show numbers.
This one talks back.
Earth Replies to Your Future is an immersive, cinematic web experience where users send a message to Earth… and Earth replies.
Not with data — but with emotion.
The goal was simple:
👉 turn awareness into feeling
Instead of dashboards and statistics, this project creates a personal conversation between the user and the planet.
After receiving Earth’s response, users make a choice — and the experience transforms into a visual future based on that decision.
🎥 Demo
👉 Live Experience: [https://earth-day-dev-weekend-challenge.vercel.app/]
👉 GitHub Repo: [https://github.com/pawarshivamd/earth-day-dev-weekend-challenge
What to expect:
- 🌍 AI-generated poetic responses (typed in real-time)
- 🎧 Earth speaking using voice synthesis
- 🌫️ Bad future → red tones, shake, smoke
- 🌿 Good future → glow, particles, calm atmosphere
(Tip: Use sound for full experience)
💻 Code
The full source code is available here:
👉 [Add GitHub Repo Link]
⚙️ How I Built It
🧠 AI Layer (Core Experience)
I used Google Gemini API (gemini-1.5-flash) as the emotional brain of the project.
Instead of basic prompts, I designed structured instructions:
- Assign a role → “You are Earth…”
- Control tone → poetic, reflective, human-like
- Limit length → short, meaningful responses
- Avoid robotic language
This ensures:
👉 every response feels alive
👉 every interaction is unique
🎨 Frontend Experience
- React + Vite → fast and modular UI
- Tailwind CSS → glassmorphism + responsive design
- Framer Motion → cinematic transitions between scenes
Instead of pages, I built scene-based flow to make the app feel like an interactive film.
🎧 Immersion Layer
- SpeechSynthesis API → Earth literally speaks
- Ambient background audio → triggered after user interaction
- Subtle particles + gradients → depth and atmosphere
⚡ Performance & UX Decisions
- Used Page Visibility API to pause heavy video/audio when tab inactive
- Avoided heavy video-first approach → used optimized visuals + motion
- Built a “Dice Roll” suggestion system for users who don’t know what to write
🧩 Key Challenges
1. Making AI feel human
Early outputs felt robotic.
✅ Solved using:
- strict prompt engineering
- role-based storytelling
- tone constraints
2. Balancing design vs performance
Heavy video backgrounds caused lag.
✅ Solution:
- lightweight visuals
- motion-based immersion instead of raw media
3. Creating emotion without data
No charts. No dashboards.
Focus was on:
- pacing
- storytelling
- user decisions
🏆 Prize Categories
✅ Best Use of Google Gemini
Gemini powers the entire experience:
- generates emotional responses
- adapts to user input
- creates a different journey every time
💭 Why This Matters
We don’t lack information about climate change.
We lack connection.
This project is an attempt to make people pause —
not to analyze…
…but to feel.
🕊️ Final Thought
“The future was never written. It was always yours to choose.”
🌍 If Earth could reply to you… what would you say?
Top comments (0)