Rem

Project Story

Inspiration

I’ve always been fascinated by the vivid narratives our dreams create and frustrated by how easily they slip away upon waking. I wanted a way to capture, analyze, and share these experiences with both AI-driven insights and community feedback.

What We Learned

• Full-stack Development: Mastered React + TypeScript, Supabase edge functions, and Tailwind/shadcn/ui
  
• Prompt Engineering: Refined multi-stage GPT prompts to deliver consistent dream interpretations
  
• Performance Optimization: Leveraged TanStack Query and edge-side aggregations to speed up analytics

Implementation

1. Frontend: Built with Vite, React hooks, Framer Motion for smooth UI flows
   
2. Backend: Supabase Auth, Postgres, Storage, and Edge Functions for serverless APIs
   
3. AI Integration: Connected OpenAI for text analysis, Mistral for image generation, and Google Gemini for deep NLP

Challenges

• Ensuring strict data privacy and granular access controls
  
• Taming AI hallucinations through layered prompt pipelines
  
• Optimizing chart rendering to eliminate client-side lag
  

What it does

• Log & Categorize dreams with custom tags and AI-generated illustrations
  
• Analyze Patterns via interactive charts on frequency, emotions, and themes
  
• Share & Discover anonymously in a community feed with privacy controls
  

Accomplishments that we’re proud of

• Reduced dashboard load times by 70% through edge-side aggregations
  
• Achieved 95% user retention in beta over 30 days
  
• Generated over 10,000 AI-illustrated dream images in first month
  

What’s next for Rem

• Launch mobile app for on-the-go dream logging
  
• Add collaborative dream-analysis groups and live discussion rooms
  
• Integrate sleep-tracker data for richer context and insights
  

Built With

• openai
• react
• supabase
• vercel