A full-stack fitness tracking app with a personal AI coach, built with Next.js 14, TypeScript, PostgreSQL, and the OpenAI API.
| Layer | Technology |
|---|---|
| Frontend | Next.js 14 (App Router), TypeScript, Tailwind CSS |
| Backend | Next.js API Routes, Zod validation |
| Database | PostgreSQL via Prisma ORM |
| Auth | NextAuth.js (credentials + Google OAuth) |
| AI | OpenAI GPT-4o mini (streaming) |
| Deployment | Vercel + Supabase |
| Testing | Jest + ts-jest |
git clone https://github.com/yourusername/fitai.git
cd fitai
npm install- Go to supabase.com → Create new project
- Go to Project Settings → Database → Connection string → URI
- Copy the connection string
cp .env.example .env.localOpen .env.local and fill in:
DATABASE_URL="postgresql://postgres:[YOUR-PASSWORD]@db.[YOUR-REF].supabase.co:5432/postgres"
NEXTAUTH_SECRET="run: openssl rand -base64 32"
NEXTAUTH_URL="http://localhost:3000"
OPENAI_API_KEY="sk-..."Get your OpenAI API key from platform.openai.com
# Push schema to your database (creates all tables)
npm run db:push
# Seed with exercise library (27 exercises)
npm run db:seed
# Optional: open visual database browser
npm run db:studionpm run devOpen http://localhost:3000 → Create an account → Start logging workouts!
fitai/
├── prisma/
│ ├── schema.prisma # Database schema — all tables defined here
│ └── seed.ts # Seeds exercise library
│
├── src/
│ ├── app/
│ │ ├── api/
│ │ │ ├── auth/ # NextAuth + register endpoint
│ │ │ ├── workouts/ # Workout CRUD endpoints
│ │ │ └── ai/chat/ # AI coach streaming endpoint
│ │ ├── dashboard/ # Main dashboard page
│ │ ├── ai-coach/ # AI chat page
│ │ └── login/ # Login/register page
│ │
│ ├── components/ # Reusable UI components
│ ├── hooks/ # Custom React hooks (useWorkouts, useAiChat)
│ ├── lib/ # Prisma client, auth config
│ ├── types/ # TypeScript type definitions
│ └── __tests__/ # Jest unit tests
- Workout logging — Log exercises, sets, reps, weight, and RPE
- Progress tracking — Charts for strength progress and weekly volume
- Personal records — Automatically detects when you beat a PR
- AI coach — Chat with GPT-4o mini, which knows your full training history
- AI program generator — Describe your goal, get a full multi-week program
- Nutrition logging — Describe a meal in plain text, AI estimates macros
- Authentication — Email/password + Google OAuth
npm run dev # Start development server
npm run db:studio # Open Prisma Studio (visual DB browser)
npm run db:migrate # Create and apply a new migration
npm test # Run unit tests
npm run test:watch # Watch mode for tests
npm run lint # Lint the codebase# Install Vercel CLI
npm i -g vercel
# Deploy
vercel
# Set environment variables in Vercel dashboard:
# DATABASE_URL, NEXTAUTH_SECRET, NEXTAUTH_URL, OPENAI_API_KEY- TypeScript — End-to-end type safety from database to UI using Prisma's generated types
- PostgreSQL — Relational database design, complex JOIN queries via Prisma
- AI integration — Streaming API responses with OpenAI, prompt engineering with user context injection
- Next.js App Router — Server components, route handlers, server-side auth
- Testing — Unit testing pure functions and validation schemas with Jest
- Stripe subscription for premium AI credits
- Mobile-responsive polish
- Export workout data as CSV
- Body measurement tracking with photos
- REST API for future mobile app
MIT