MaVita

Proactive AI nutrition to prevent chronic diseases, combat illness, and extend healthspan.

Comment

Inspiration

The seeds of MaVita were sown in a place of profound personal loss. Having lost both my sister and, most recently, my aunt to cancer, I became obsessed with a single question: How can we move from a reactive healthcare system to a proactive, preventive lifestyle? Diet is our most frequent interaction with our biology, yet most nutrition apps stop at calorie counting. I wanted to build something deeper, an "Oncology-first" nutrition assistant that understands how every bite of food influences metabolic pathways, DNA repair, and chronic inflammation. MaVita is my effort to transform grief into a tool that empowers others to safeguard their health through personalized nutrition.

What it does

  • MaVita is a "Bio-Engine" for longevity. It uses Gemini 3 to transform a simple meal photo into a comprehensive biological report.
  • Multimodal Meal Analysis: It recognizes culturally diverse dishes and identifies their specific glycemic impact.
  • Metabolic Pathway Mapping: The app reasons through biochemistry to show how nutrients support functions like Collagen Synthesis or Mitochondrial Biogenesis.
  • CGM Prediction Intelligence: By syncing with Continuous Glucose Monitors, it predicts glucose spikes before they happen and offers "Smart Pairing Protocols" to blunt them.
  • Preventive Priority Map: Users can toggle focus goals such as Cancer Prevention, Diabetic Health, or Radiant Glow (skin health), and the AI tailors all recommendations to that specific vulnerability.

How I built it

  • The application is powered by the Gemini 3 API, utilizing its next-generation reasoning and multimodal capabilities.
  • Vision & Multimodality: I used Gemini 3’s vision capabilities to identify ingredients in complex, mixed-ingredient meals where traditional databases often fail.
  • Reasoning Engine: I implemented a "Chain-of-Analysis" prompt structure. Gemini 3 doesn't just look up data; it reasons about the interaction between specific phytonutrients (like Anthocyanins or Ellagic Acid) and the user's specific health goals.
  • Low-Latency Feedback: Using Gemini 3’s reduced latency, the "Bio-Feed" provides near-instant analysis, making the experience feel like a live conversation with a nutritionist.

What I learned

Building this project taught me that AI is the missing link in personalized medicine. I learned that we are no longer limited by static databases; with Gemini 3, we can create "Living Logic" that adapts to a user's DNA, wearable data, and daily habits in real-time.

What's next for MaVita

I plan to integrate deeper DNA Blueprint analysis, allowing the Bio-Engine to suggest meals that specifically compensate for a user's genetic predispositions. My goal is to put a world-class preventive oncology nutritionist in the pocket of everyone, regardless of their background or location.

Built With

Share this project:

Updates