🌙 DreamEcho: Visualize Your Dreams, Own Your Subconscious
DreamEcho is a web-based platform that transforms dream descriptions into symbolic 3D models using AI. It enables users to interact with and preserve their dreams as NFT-like digital assets — even bind them to NFC chips for physical touchpoints.
💡 Inspiration
Everyone dreams, but we forget 90% of dreams by morning. What if we could capture dreams as assets?
DreamEcho is inspired by the idea that dreams are not just stories — they're emotional experiences, personal insights, and metaphoric journeys. We want to give them visual, interactive, and lasting form.
⚙️ What It Does
- Accepts dream descriptions (text input)
- Uses DeepSeek API to extract key symbols, emotions, characters
- Sends results to Tripo API for AI-based 3D object generation
- Loads and renders the
.objmodel in real time via Three.js - Dream model can be:
- Explored interactively (orbit, zoom, pan)
- Interpreted with emotional/symbolic context
- Minted as an NFT (next step)
- Linked with NFC tags for physical interaction (e.g. via card)
🛠️ How We Built It
- Frontend:
HTML5,CSS3,JavaScript ES6,Three.js - 3D Model Loader:
OBJLoader + OrbitControls - AI Integration:
DeepSeek API(dream analysis) +Tripo API(model generation) - Hosting:
GitHub Pages - NFC linking: via
nfc tag URLand smartphone triggers - Local server for testing:
Python3 -m http.server
🧱 Challenges
- Accurate symbolic extraction from freeform dreams
- Optimization of 3D model rendering for mobile
- Balancing between performance and aesthetic detail
- Integrating diverse APIs with smooth UX
- Handling model cache & error fallback in browser
🎓 What We Learned
- How to parse natural language symbols from dreams
- How to load large 3D assets responsively
- Web deployment tricks with GitHub Pages
- The role of storytelling and visual semiotics in emotional design
🚀 What's Next
- NFT minting workflow with WalletConnect
- Dream journal system with multi-model views
- Multi-user gallery and social dream board
- More emotional metadata (valence/arousal tagging)
🔗 Try It Out
🌐 Live Demo
📁 Source Code on GitHub
🎥 Video Demo
👤 Created By
Wu JiaJun (Shenzhen Technology University)
Email: epwujiajun@icloud.com
GitHub: wujiajunhahah
Log in or sign up for Devpost to join the conversation.