Ovi - A Pregnancy Nutrition Tracker App

Overview

This app supports pregnant individuals with personalized nutrition guidance from conception to birth. It provides weekly-calibrated calorie and nutrient targets, meal suggestions, AI-powered food analysis, and calming health insights without overwhelming the user.

Rather than acting like a diet app, it functions as a nutritional companion—adapting daily, helping users understand what their body and baby need, and making daily food choices easier, safer, and more intentional.

Inspiration & Related Work

This app was inspired by a close friend of ours Anjana Danda, a dedicated medical assistant who is currently pursuing a future as a Physician Assistant (PA). Through her work in healthcare, she noticed numerous challenges related to women’s diets and overall care during pregnancy. Driven by a passion to improve this area of maternal health, she researched extensively and helped lay the foundation for this project, guiding us to create a solution that truly addresses the unique needs of pregnant individuals.

This app also builds on the technical foundation of NeuroSpeak, an AI speech-therapy assistant that won first place at Emory’s AI Hackathon 2025. Lessons in accessibility, empathetic UX, and responsible AI scaffolding directly influenced this project.

Product Goals

Personalization

Adjust targets based on daily of pregnancy, health conditions, and user goals.

Ease of Use

Quick logging via food images, barcodes, or smart search.

Trustworthy Insights

Guidance derived from United States Department of Agriculture(USDA), Food Safety and Inspection Service (FSIS) and National Institute of Food and Agriculture (NIFA) dietary standards.

Safety First

AI answers real-time food safety questions (e.g., sushi, caffeine, cheeses).

Calm Design

Gentle insights, no guilt-based tracking or weight-focused metrics.

Core Features

Weekly Nutritional Guidance

Automatically recalibrates calorie and micronutrient needs each week.
Visual indicators show whether targets are met without red alerts or stress triggers.

Food Logging (Three Modes)

Photo Logging — AI meal parsing using Gemini Vision.
Barcode Scanning — For packaged foods.
Smart Search — Autocomplete powered by Spoonacular.

AI Q&A Assistant

Ask questions like:

Can I eat brie cheese right now?

Receives week-specific, safety-aware responses. Supports follow-ups such as:

What about baked brie?

Nutrient Breakdown

Real-time charts for vitamins, minerals, and macronutrients.
Highlights pregnancy-specific targets such as folate, iron, choline, magnesium, and calcium.

Smart Meal & Grocery Suggestions

Recommendations based on nutrient gaps for the current day or week.
Filters include: vegetarian, gestational diabetes, allergies, and food aversions.

Journal & Mood Tracker

Track cravings, symptoms, sleep, and energy.
Identify trends through correlation with nutrient intake.

Technology Stack

Backend

Firebase Firestore — Structured user + pregnancy data.
Firebase Auth — Secure authentication.
Firebase Functions — Weekly updates and automated insights.
Gemini Vision API — Photo-based meal analysis.
Gemini Pro — Chat, food-safety reasoning, recommendations.
Spoonacular API — Food database and nutritional facts.

Frontend

React Native (Expo) — Cross-platform application.
Tailwind (NativeWind) — Styling system.
React Navigation — App flow routing.
Lottie — Animated onboarding and week transitions.

AI Use Cases

Task	Model/API	Description
Photo meal analysis	Gemini Vision	Detects food items and portions
Food safety assistant	Gemini Pro	Answers “Can I eat this?”
Nutrient-gap suggestions	Gemini Pro	Suggests meals based on micronutrient needs
Journal prompts	Gemini Pro	Generates mood/symptom reflections

Competitive Advantages

Week-by-week personalization (not just trimester blocks)
AI photo-based meal logging
Real-time food safety chat
Calm, non-punitive design
Privacy-first: no third-party data reselling

Challenges We Ran Into

Week-specific nutrient calculations: USDA recommendations are trimester-based, but we implemented smooth week-level interpolation to avoid abrupt nutrient target changes.
Food safety reasoning: Soft cheeses, raw fish, caffeine, deli meats—rules vary by context. Ensuring AI responses were medical-grade, cautious, and non-hallucinated required intense testing and prompt engineering.
Maintaining a calm, reassuring UX: Many pregnancy apps trigger anxiety with harsh red indicators. We intentionally designed gentle UI patterns and supportive wording.

What We Learned

Responsible AI in health contexts: How to prevent hallucinations, give safe alternatives, and use disclaimers correctly.
The science of pregnancy nutrition: Understanding weekly changes in folate, choline, DHA, iron, calcium, magnesium, and more.
Combining vision models with nutrition databases: Converting food images into structured, medical-grade nutrient insights.
The importance of a calm maternal-health UX: Users respond positively when stress, guilt, and alarms are removed.
Coordinating multi-API workflows: Orchestrating Gemini Vision, Gemini Pro, Spoonacular, and Firebase Functions seamlessly.

What's Next for Ovi:

Medical Provider Dashboard
A clinician portal (with user consent) for reviewing trends and giving personalized advice.
Symptom to Nutrient Linking
Explain cravings, nausea, headaches, energy dips, and sleep changes through nutrient correlations.
Baby Development + Nutrition Insights
Weekly insights explaining fetal development and the nutrients your body needs for that growth.
Offline Logging Mode
Caching food logs and syncing automatically when back online.
Language & Cultural Food Support
Use hydration, heart-rate, and sleep data to give health-aware nutritional guidance.