ThreadHer: AI Agent Platform for Sustainable Fashion

• 

  ThreadHer Architecture Diagram

• 

  Homepage with upload interface and quick questions

• 

  Leather jacket uploaded for analysis

• 

  AI processing garment with Claude Vision

• 

  Analysis: Leather jacket, 50-100 kg CO₂, 17,000L water

• 

  Recommendations: Care tips, donation platforms, sustainable alternatives

ThreadHer: AI-Powered Fashion Sustainability Advisor

💡 Inspiration

I grew up without computer access. Today, I'm using AI to tackle fashion's environmental crisis. As a woman in tech who loves fashion, I recognized the gap between consumers wanting to be sustainable and having practical tools to do so. I built ThreadHer to empower women to make informed wardrobe choices while reducing fashion's 10% contribution to global carbon emissions (9.2*10^7 tons waste annually).

🛠️ How I Built It

AWS Serverless Stack: S3 (frontend + storage) → API Gateway → Lambda (Python) → Bedrock + Claude 3.5 Sonnet

Flow: Upload garment → Claude Vision analyzes → Returns CO2 (5-33 kg) + water (2,700-10,000 L) + personalized recommendations

💪 Challenges

• Secure image pipeline with pre-signed S3 URLs
• 15+ prompt iterations for consistent AI responses
  
• Async coordination with 3-5s response time
• Mastering Bedrock, Lambda, S3 simultaneously

📚 What I Learned

AWS serverless architecture, Claude Vision integration, prompt engineering. Impact modeling: with n users extending garment life by factor:

ΔW=n×36.7×(1−f1​) kg/year

Key insight: Technology empowers when it makes information actionable.

🎯 Impact

10,000 users at 25% lifespan extension = 73,400 kg waste reduction/year.

Next: Mobile app, wardrobe tracking, donation mapping.

From no computer to building AI for change. 🌿

Stack: Bedrock • Claude 3.5 • Lambda • S3 • API Gateway • Python • JavaScript

Built With

• amazon
• apigateway
• aws-lambda
• bedrock
• css
• dynamodb
• html
• javascript
• python
• react
• sql
• tailwind-css