AI Fitness Tracker lets you log meals and workouts using plain English. It parses foods, serving sizes, and macros to maintain daily totals and trends.
Important: The production/full version (not included in this site) is a Java Servlets + MySQL application with server-side authentication, persistent storage, and richer features (e.g., reminders, role-based views).
This portfolio includes a small, client-only demo that re-implements natural-language meal parsing in the browser so you can quickly try it out without a backend.
JavaServletsMySQLNLPOpenAI APINext.jsTypeScriptTailwindCharts
Features⌄
- Natural-language meal logging (e.g., "2 eggs and toast with butter, coffee")
- Macro extraction: calories, protein, carbs, fat with a lightweight rules engine
- Daily totals, editable log, and simple charts
- Auth and roles in the original server version (Java + MySQL)
Technical Implementation⌄
- Full App (not included): Java Servlets backend with a normalized MySQL schema for users, meals, and nutrition entries (SQL), plus auth and reminders
- Parsing: tokenization + quantity/unit detection + ingredient lookup with fallback heuristics (with optional GPT-assisted parsing)
- Demo (this site): Client-only React/TypeScript parser, localStorage persistence, and computed totals
- UI: Tailwind styling and accessible forms; charts placeholder ready for plug-in
Challenges & Solutions⌄
- Ambiguous inputs → Fallback database with portion defaults and user adjustments
- Unit variety → Normalize to grams/ml; map to standard servings where possible
- Porting from NetBeans/Tomcat → Browser-based demo removes server dependency for easy showcase
Technologies Used⌄
- Java, Servlets, MySQL (original full-stack)
- Next.js, React, TypeScript (demo reimplementation)
- Tailwind CSS for UI; Chart.js (optional) for trends
- OpenAI API (optional) for advanced parsing and suggestions
Try the Live Demo
This is a mini, client-only demo for the case study. The full app (not included) uses Java Servlets + MySQL with richer features.