Follow these steps to spin up the entire Full Stack application, complete with Live Data ingestion and the Machine Learning prediction pipeline.

1. Create a free project at supabase.com.
2. Go to SQL Editor and run the full SQL block from prompt.md to create the tables (games, predictions, player_cards, media_takes, etc.).
3. Copy your Project URL and service_role key from Settings → API.

1. Create a free Redis database at upstash.com.
2. Copy the UPSTASH_REDIS_REST_URL and UPSTASH_REDIS_REST_TOKEN.

Navigate to the backend directory and set up your environment:

cd backend
cp .env.example .env

Fill in the following values in your new .env file:

• GEMINI_API_KEY: Your Google Gemini API Key.
• SUPABASE_URL: Your Supabase Project URL.
• SUPABASE_SERVICE_KEY: Your Supabase Service Role Key.
• UPSTASH_REDIS_REST_URL: Your Upstash REST URL.
• UPSTASH_REDIS_REST_TOKEN: Your Upstash REST Token.

Navigate to the frontend directory:

cd frontend
cp .env.local.example .env.local

Fill in the following values:

• NEXT_PUBLIC_API_URL: http://localhost:8000 (for local development)
• NEXT_PUBLIC_SUPABASE_URL: Your Supabase Project URL.
• NEXT_PUBLIC_SUPABASE_ANON_KEY: Your Supabase Anon Key.

To power the AI and ML models, you need to install the dependencies, fetch the real historical data, and train the XGBoost models. 1. Install Python Dependencies

cd backend
pip install -r requirements.txt

2. Bootstrap Database Tables Fetch all current NBA players and the last 3 seasons of historical game data necessary to train the AI:

python -m app.scripts.bootstrap_players
python -m app.scripts.bootstrap_historical

3. Train the ML Pipeline Run the XGBoost training script. This will extract the 15 requested feature sets from the historical database and generate the model artifacts (saved in app/ml/artifacts/):

python -m app.ml.train

4. Run the API Server Start the FastAPI background scheduler and live endpoints:

uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload

Now that the backend is streaming live predictions, spin up the React interface:

cd frontend
npm install
npm run dev

1. Push your repository to GitHub.
2. Create a new Railway project → connect repo → set root to backend/.
3. Load all backend environment variables (GEMINI_API_KEY, etc.) into Railway.
4. Railway will auto-detect the provided railway.toml and run uvicorn.

1. Create a Vercel project → connect repo → set root to frontend/.
2. Load all frontend environment variables. Ensure NEXT_PUBLIC_API_URL points to your active Railway domain.
3. Deploy!