Smart Eat
A nutrition and health tracker that builds a daily norm from a medical questionnaire, splits it across meals, and reads every meal, glucose and blood pressure reading against it.
9
questionnaire steps before the first meal
5
meals a day at most, each with its own norm
5
glucose reading types, fasting to after exercise
4
zones on the glucose chart, hypo to diabetes
- Client
- A nutrition startup
- Industry
- Health and nutrition
- Platforms
- iOS and Android from one React Native codebase
- Our part
- Team lead, product, app and backend
- Timeline
- 2024–2025, released
- Screens
- English build, demo account, stock dish photos
The brief.
Context
Smart Eat is a nutrition app for people managing their weight or a diet-sensitive condition, including diabetes. It learns the user's body, health and habits first, turns that into a daily norm, and then tracks food, blood glucose and blood pressure in one place. We led the team and built the app from the client's first draft to release.
Challenge
A calorie counter was not enough. The norm had to account for illnesses, allergens, diet system and religion, and be split across the user's own meal schedule. Diabetics needed bread units next to carbohydrates and glucose shown against clinical ranges. All of it had to stay quick to log several times a day.
Approach
A nine-step questionnaire feeds the backend, which returns a daily norm split into meals. Logging a meal goes from meal to place to dishes and portions, with search and a smart sort that fits portions to what is left of the meal's norm. The health diary charts glucose and pressure, and an assistant answers questions in a chat.
It gets to know you first.
Before the first meal the app asks nine short questions: gender, name and birthday, height and weight, goal and activity, illnesses and allergens, diet system and religion, and how many meals a day. Pickers are rulers and wheels rather than text fields, and every step can be edited later from the profile.
One norm, split across your meals.
The user decides how many meals they eat and what share of the day each one gets. Every meal then carries its own calories, protein, fat, carbohydrates and bread units, and the home screen shows how much of the day's norm is already eaten.
- Bread units sit next to carbohydrates everywhere, because that is how people with diabetes count.
Recipes that fit the norm.
Home Kitchen offers a recipe of the day, community recipes by category and the user's own. Each recipe card shows weight, calories, protein, fat, carbohydrates and bread units with the ingredients. Users can add their own dish with a photo, building it from searchable ingredients with known nutrition.
Glucose and pressure, in context.
Glucose readings are logged by type, from fasting to after exercise, and charted by day, week or month over coloured zones: hypoglycemia, normal, prediabetes and diabetes. Blood pressure has its own diary and charts, and a reading can be entered straight from a reminder notification.
Ask about your own numbers.
The assistant tab is a chat with suggested questions. The user asks in plain words, the question goes to the backend, and the reply arrives in the conversation through Firebase Realtime Database.
From questionnaire to logged meal.
The app is React Native with MobX stores and talks to a REST API that requires a Firebase ID token on every call. Dishes and ingredients are searched through Typesense, the assistant chat runs over Firebase Realtime Database, and Firebase Cloud Messaging carries reminders.
Sign-in
Phone number and SMS code through Firebase Auth.
Questionnaire
Body, health, diet and meal schedule are saved to the profile.
Daily norm
The backend returns the day's norm split into the user's meals.
Meal
Dishes found through Typesense; a smart sort fits portions to the meal's remaining norm.
Diary
Meals, glucose and pressure land in one history with day, week and month charts.
What we built.
Our part
- Leading the team from the client's first draft
- Product flows and interface
- iOS and Android app in React Native
- Daily norm split across the user's meals
- Meal logging with dish search and portion sorting
- Recipes with own-dish creation and ingredient search
- Glucose and blood pressure diaries with charts
- AI assistant chat and push reminders
Stack
- App
- React Native 0.74
- TypeScript
- MobX
- React Navigation
- i18next
- Data and search
- REST API
- Typesense
- Firebase Realtime Database
- Platform
- Firebase Auth
- Firebase Cloud Messaging
- Charts and UI
- react-native-gifted-charts
- d3-shape
- Bottom sheets
- Wheel pickers
Building a health product?
Let's make it easy to use every day.
We turn medical and nutrition rules into apps people actually open several times a day. Tell us about yours.
