Thrivable Logo
Profile
Product Identifier
Landing Page

Thrivable - AI-Powered Sustainability Platform

Inspiration

In today's world, environmental consciousness is more important than ever. We were inspired by the growing need for accessible tools that help individuals make sustainable choices in their daily lives. Many people want to reduce their environmental impact but struggle to find reliable information about products, materials, and their ecological footprint.

Thrivable was born from the vision of creating an intelligent platform that combines computer vision, AI analysis, and environmental data to empower users to make informed, eco-friendly decisions. We wanted to bridge the gap between environmental awareness and actionable insights, making sustainability accessible to everyone.

What it does

Thrivable is a comprehensive sustainability platform that helps users understand and reduce their environmental impact through intelligent image analysis and AI-powered insights.

Key Features:

Smart Image Analysis: Upload or capture photos of products/materials to get instant environmental impact assessments
Environmental Scoring: Receive detailed environmental scores (0-100) with CO2 footprint calculations
Sustainability Tips: Get personalized recommendations for reducing environmental impact
Community Leaderboard: Compete with others and track your sustainability progress
DIY Project Generator: Discover creative ways to repurpose items instead of discarding them
Real-time Analytics: Visualize your environmental impact over time

How it Works:

Capture: Take a photo or upload an image of any product
Analyze: Our AI analyzes the image and identifies components
Assess: Get detailed environmental impact scores and CO2 footprint data
Act: Receive tips and alternative solutions
Track: Monitor your progress with a points system and compete on the leaderboard

How we built it

Frontend Technologies:

React.js with Vite for fast development and optimal performance
Tailwind CSS for modern, responsive design
Framer Motion for smooth animations and user interactions

Backend Technologies:

Node.js with Express.js for robust server-side logic
Groq API for fast AI inference and image analysis
Tavily API for web scraping and data collection
Supabase for user authentication and data persistence
JWT for secure authentication

AI & Data Pipeline:

Computer Vision: Image recognition and material identification
Natural Language Processing: Environmental impact analysis
Web Scraping: Real-time data collection from environmental databases

Architecture:

Frontend (React) ↔ Backend (Node.js/Express) ↔ AI Services (Groq/Tavily)
                                      ↓
                              Database (Supabase)

Challenges we ran into

Data Pipeline Complexity

Our biggest challenge was building the robust data pipeline that transforms a simple image into an accurate environmental assessment. We engineered a multi-stage process that first uses Google Cloud Vision's Product Search and a custom OCR fallback to precisely identify the product. This triggers the Tavily API to crawl the web for sustainability data, which is then synthesized by the LLaMA 3 model via the Groq API to generate an actionable eco-score, tips, and a final carbon footprint analysis.

Web Scraping Limitations

We encountered various challenges with web scraping, particularly in finding and extracting the right environmental information. Many websites have:

Anti-bot measures that required careful request management
Dynamic content that required sophisticated scraping techniques
Inconsistent data formats that needed extensive parsing and normalization

Information Accuracy

Finding reliable, up-to-date environmental data proved challenging. We had to:

Verify data sources for accuracy and credibility
Handle conflicting information from different sources
Update outdated environmental metrics with current standards
Normalize data formats across various databases

Accomplishments that we're proud of

Technical Achievements:

Real-time Image Processing: Built a fast, responsive image analysis system
Modular Backend Design: We architected the backend using Express.js with a strong separation of concerns. Each core feature is encapsulated in its own dedicated module with distinct routes and controllers
Intelligent Web Scraping/Data Extraction: Developed a dynamic data extraction pipeline using that performs targeted web searches for environmental data, filters irrelevant sources, and extracts raw text content from the most promising URLs for AI analysis.

What we learned

AI Service Integration: Learned to effectively coordinate multiple AI APIs for complex analysis
Data Pipeline Design: Gained expertise in building robust data processing systems
Web Scraping Best Practices: Developed strategies for handling rate limits and dynamic content
Real-time Processing: Mastered techniques for optimizing performance in image analysis

What's next for Thrivable

Mobile App Development: Create native iOS and Android applications
Enhanced AI Models: Improve accuracy of environmental impact assessments
Social Features: Implement sharing and community features
Barcode Integration: Add product barcode scanning for instant environmental data

Technical Roadmap:

Performance Optimization: Implement caching and CDN for faster global access
Advanced AI: Integrate more sophisticated AI models for better accuracy
Data Expansion: Build partnerships for more comprehensive environmental databases
Blockchain Integration: Explore blockchain for transparent environmental impact tracking

Built with ❤️ for a sustainable future

Thrivable - Making sustainability accessible, one picture at a time.