AI diet coach · Flutter + LLM
TrainerAI
An AI fitness coach app: answer six questions and get a personalised daily diet plan with calories and macros for every meal.
Built and shared as an Android APK
- When
- Apr – Aug 2025
- My part
- Solo: app, landing page and backend
- Onboarding
- 6 questions
- Plan
- Per-meal macros
- Repos
- App · site · API
- Commits
- 42
Chapter 1 of 5 ·
A trainer in your pocket that writes your meals
Most diet apps make you log food. This one asks how you live and what you eat, then writes the plan for you: breakfast to dinner, with calories, protein, carbs, fats, fibre and water for the day.
$ git log feat: Implement diet plan generation feature with API integration
Chapter 2 of 5
Six questions, no forms
How often you exercise, whether you track steps, how many you take, and whether you eat keto, vegan, paleo or anything else. Big tap targets, one question per screen, and you can always go back.
Chapter 3 of 5
A language model that answers in meals
The answers go to a large language model through Groq, which returns a structured plan the app can render: each meal with its time, items and macros. Regenerate if you don't like it, or start over.
Chapter 4 of 5 ·
A product needs a front door
A Next.js landing page with a download section, FAQ, dark and light modes, an animated hero and a published virus-scan report for the APK, so people could trust a download from outside the Play Store.
Chapter 5 of 5 ·
Accounts that follow you
An Express and MongoDB backend with Google sign-in and JWT sessions, packaged with Docker, then plans saved to Firestore so they follow you between phones.
$ git log Initialize backend for JWT authentication with Google OAuth, including user signup and login functionality, database connection, and middleware for token verification.
Next cartridge
Zoheaca
A healthcare platform for India: find doctors and labs nearby, book appointments and tests, and keep medical records in one place.