Dhruv Seth — PortfolioSheet 01/01—X 0000 · Y 00000Rev b4dbfbaScale 1:1
dhruvseth.in
  • Home
  • Projects
dhruvseth.in

© 2026 Dhruv Seth · India · Remote

HomeProjectsEmailLinkedIn ↗
Index/Projects/Diet X
Case study2026

Diet XDiet X

Nutrition tracking built for Indian food — not just calories.

Role
Mobile & Backend Engineer
Status
In progress
Stack
Flutter logoFlutterDart logoDartRRiverpodGgo_routerNode.js logoNode.jsHHiveMongoDB logoMongoDB
Links
Private
01 · Client02 · Services03 · DataFlutterRiverpodgo_routerNode.jsHiveMongoDB
Fig. 01 — System overviewSwipe to explore →Written in Dart
Summary

Diet X helps Indian users answer three questions every day: what did I eat, am I meeting my nutrition goals, and what should I eat next? It logs real Indian meals in realistic servings, tracks macros and micronutrients against a personal plan, flags intake gaps, and plans the week ahead — without ever pretending to be a doctor.

Measured
Foods in the database
1,419
Food categories
20
Nutrients tracked
22
Contents
  1. 01The problem
  2. 02How it works
  3. 03The calorie engine
  4. 04The food database
  5. 05Features
  6. 06Architecture
  7. 07Boundaries
  8. 08Status

Diet X

Diet X is a Flutter nutrition app made for Indian users. It goes beyond calorie counting to answer three daily questions: What did I eat today? Am I meeting my nutrition goals? What should I eat next?

The problem

Most calorie apps are built around Western packaged food. Logging roti, dal, poha or sambar means guessing at a "generic" entry, in servings nobody actually eats, and the app only cares about calories anyway. Indian diets need Indian foods, Indian servings, and the full nutrient picture — iron, calcium, B12 and the rest — not just a number.

How it works

01
Plan
Onboarding answers become daily targets
02
Log
Indian foods, realistic servings, per meal
03
Track
Calories, macros and micronutrients vs. targets
04
Spot gaps
"You are low on iron today"
05
Suggest
Foods that close the gap
Fig. — One day in Diet X
  1. Onboarding builds a personal plan. After a splash screen and onboarding slides, the app asks for your goal (lose, gain, build muscle or maintain), gender, age, height, current and target weight, activity level, diet preference (veg, vegan, non-veg, Jain, keto or custom) and water goal — then generates daily calorie, protein, carb, fat and water targets.
  2. Log what you actually ate. Search Indian foods — roti, dal, paneer, poha, dosa, idli, sambar, rajma, chole, khichdi, biryani, sabzi and more — pick a real serving (bowl, cup, piece, grams or ml), and log it to breakfast, lunch, dinner or snacks.
  3. See consumed vs. required. Every nutrient shows progress against its target.
  4. Close the gaps. Low intake is flagged with foods that help — spinach, dal, paneer, almonds and similar.

The calorie engine

Targets come from a documented, deterministic calculator — no black box:

text
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
BMR (Mifflin-St Jeor)
  male    = 10·kg + 6.25·cm − 5·age + 5
  female  = 10·kg + 6.25·cm − 5·age − 161

TDEE      = BMR × activity (1.20 sedentary … 1.90 extra active)

Target    = TDEE − 500   lose weight     (floored at 1200 kcal)
            TDEE + 400   gain weight
            TDEE + 200   build muscle
            TDEE         maintain

Protein   = 1.8 g per kg body weight
Fat       = 25% of calories
Carbs     = whatever calories remain   (floor: 50 g)

Mifflin-St Jeor was chosen because it's the most accurate resting-metabolism predictor that doesn't need body-fat percentage, which onboarding doesn't ask for. BMI is shown for context only — it never drives the target, because it ignores muscle mass and frame size.

The food database

The nutrition data is its own project: 1,419 foods across 20 categories, every value per 100 g or 100 ml, sourced from IFCT 2017 (India's Food Composition Tables, NIN Hyderabad), USDA FoodData Central and verified manufacturer labels. Categories include North Indian curries, South Indian, rice and wheat dishes, pulses and legumes, street food and sweets. A Node.js seeder loads it into MongoDB.

Every entry has the same shape:

tty — One entry from north_indian_curries.jsonLive
$ jq '.[] | select(.name == "Dal Tadka")' north_indian_curries.json
{
"name": "Dal Tadka",
"servingSize": 100, "unit": "g",
"calories": 118,
"macros": { "protein": 6.2, "carbs": 14.8, "fat": 3.8, "fiber": 4.5 },
"minerals": { "iron_mg": 2, "calcium_mg": 38, "potassium_mg": 348 },
"source": "IFCT 2017 / NIN Hyderabad"
}
Fig. — One entry from north_indian_curries.json

Features

  • Personalized plan — goal, body metrics, activity, diet preference and water goal become daily targets.
  • Indian food logging — realistic servings, four meal slots, edit or delete any entry.
  • Full nutrient tracking — calories, macros and fiber; vitamins A, B1, B2, B3, B6, B12, C, D, E and K; calcium, iron, magnesium, zinc, potassium, sodium and phosphorus.
  • Deficiency detection — intake gaps flagged with food suggestions to fill them.
  • Custom meals — combine foods into a saved meal with total nutrition, then reuse it.
  • Weekly meal planner — a 7-day Indian meal chart matched to your goals and diet type; edit or regenerate it.
  • Symptom-based suggestions — for acidity, bloating, stomach ache or low energy, light foods like khichdi, banana or curd rice.
  • Water and progress — daily water intake, weight history, calorie and nutrient trends, daily and weekly reports with charts.
  • Profile and settings — edit goals, notifications, and light or dark theme.

Navigation is five bottom tabs: Home, Add Food, Planner, Progress and Profile. The design is calm and premium — rounded cards, subtle glass, smooth animations, clear progress indicators, in light and dark.

Architecture

The app is Flutter, organised by feature — onboarding, plan, food logging, meals, nutrients, weekly planner, symptoms, water tracking, progress, profile and settings. State is managed with Riverpod, navigation with go_router, and data is stored on the device with Hive so the app works offline. The core models are User, Food, DailyLog, Meal, WeeklyPlan and WaterLog.

Boundaries

Diet X never diagnoses a disease or replaces a doctor or dietitian. Gaps are framed as intake gaps ("low on iron today"), not medical conditions, and symptom suggestions are dietary ideas only. The calorie engine refuses to go below 1200 kcal a day — anything lower needs medical supervision.

Status

In progress. The sign-up, log-in and forgot-password screens are built; email and Google authentication and cloud sync are planned on top of the current offline-first storage.

Built with

  • Dhruv Seth
    Dhruv Seth
    Mobile & Backend Engineer
    LinkedIn ↗
  • DG
    Deepankar Garg
    Co-builder
    LinkedIn ↗
Next sheet→
NexusNexus

A product & engineering workspace — from a single task to the release it ships in.