Glowify
Inspiration
We were inspired by Melaleuca’s mission to enhance lives across the globe — not only in grand, visible ways, but also in the microscopic details of everyday life. With the rapid evolution of artificial intelligence, we saw an opportunity to help those who struggle with persistent skin issues or feel limited by the lack of intelligent tools for skincare management. Glowify was born from the idea that technology can empower individuals to take better care of their skin, improve their confidence, and ultimately enhance their quality of life.
What It Does
Glowify leverages advanced computer vision models to analyze facial images and extract key dermatological insights.
Our AI examines multiple skin metrics, including:
- Skin dryness
- Wrinkle intensity
- Dark circle depth
- Oil concentration
- Acne severity
- Redness level
- Pore size
- Texture roughness
Using this analysis, Glowify generates a personalized skin health profile and matches users with Melaleuca’s best-suited skincare products, offering actionable insights and long-term tracking to monitor progress over time.
How We Built It
- Frontend: Built using React, Vite, and TailwindCSS for a modern, responsive, and seamless user experience.
- UI Design: Guided by Base44 design principles for visual consistency.
- Backend: Powered by Python and FastAPI for high-performance API handling of image analysis and model inference.
- Database & Storage: Managed by Supabase for scalability and security.
- Deployment: Both frontend and backend were deployed on Render, ensuring reliable CI/CD and uptime.
Challenges We Ran Into
One of the main challenges was finding and fine-tuning vision models capable of accurately identifying subtle skin conditions.
Initially, we experimented with Roboflow’s acne and pimple detection models, but their accuracy and coverage was limited.
Eventually, we integrated an API with a multi-layered vision model capable of detecting multiple skin features such as redness, pores, and dryness. This integration required complex backend optimization as the integration with Roboflow's vision models were already completed, but it dramatically improved Glowify’s analytical precision and user value.
Accomplishments We’re Proud Of
We’re proud of how far we advanced in both technical development and user experience design:
- Developed a functional, visually refined frontend
- Built a secure token-based authentication system and integrated Google OAuth2.0 and credits-based monetary payment system
- Integrated vision models with backend infrastructure
- Achieved full-stack deployment on Render
- Delivered a prototype that combines AI insights with a smooth user experience to enable users with value
- History of users' dermatological journey by saving their scores of in different skin condition categories
What We Learned
During development, we deepened our understanding of:
- The client-server model and HTTP communication for image data
- Effective task separation, boosting productivity and clarity
- Advanced Git collaboration, including conflict resolution and version control
- Advantage of abstracting business logic into the backend, rather than placing some in the frontend
We learned that while intuitive “vibe coding” has its place, structured and deliberate, AI-driven development leads to more impactful and sustainable outcomes.
What’s Next for Glowify
Looking ahead, next steps for Glowify:
- Develop a custom-trained vision model tailored to diverse skin tones and conditions
- Offer seasonal and lifestyle-based Melaleuca-specific skincare recommendations
- Launch mobile versions for iOS and Android or integrate with already existing mobile and web applications
- Integrate LLM with Melaleuca's MCP server to equip the AI to become the most powerful and informative AI dermatologist
Our vision is to make Glowify the go-to AI-powered skincare companion, empowering everyone to achieve healthier, more radiant skin through personalized, data-driven insights.
Built With
- ailabtools
- base44
- fastapi
- github
- google-gmail-oauth
- postgresql
- python
- react
- render
- roboflow
- supabase
- tailwindcss
- typescript
- vite
Log in or sign up for Devpost to join the conversation.