Eatensity
Project brief
3,300+ Indian and global foods with Nutri-Score grading and a daily calorie tracker. A full-stack Turborepo monorepo spanning web, mobile (coming soon), and a B2B API (coming soon).
Eatensity started with a question that felt oddly unanswered: why is it so hard to find reliable nutritional data for Indian food? The global databases skew Western. The Indian ones are scattered across PDFs and academic papers. So I built the bridge.
The platform indexes 3,300+ foods from IFCT 2017, Open Food Facts, and USDA — unified into a single searchable interface with Nutri-Score grading adapted for Indian dietary patterns. The calorie tracker lets you search and log meals quickly without hunting through ingredient lists.
Architecturally, it's a Turborepo monorepo: Next.js 15 with React 19 for the web app on Vercel, React Native with Expo for the Android app on Google Play, a Hono edge API, and Supabase handling Postgres, Auth, and Realtime subscriptions.
The B2B REST API is an intentional next step — the food data layer is useful beyond my own app. Third-party wellness platforms and fitness apps will be able to query it directly.
Approach
Data first. I built and validated the nutritional database before touching any UI. Every food entry is normalized across three sources with conflict resolution rules.
Outcome
Live on web. Play Store and B2B API are coming soon.
More work in a similar range.
figuretools
Static collection of calculators and converters. No framework, no accounts, no ads. Built with Eleventy, builds in under a second.
Kanso
CSV import, rule-based auto-categorization, and anomaly detection using mean + 2σ. Built with Spring Boot and Thymeleaf.