ML-driven sports betting product: predictions, Kelly-based bet sizing, live odds, and a real subscription surface. Live at aiadvantagesports.com.
Most "AI betting" projects stop at a notebook. This one ships.
AI Advantage turns model output into a decision a user can actually act on:
- Predict — model-driven picks across NBA, NFL, and MLB
- Size — Kelly-based stake recommendations that protect bankroll
- Time — live odds and line-movement views to catch value before the market adjusts
- Monetize — premium tier with real Stripe checkout, not a fake paywall
It's the product layer of a larger sports-analytics stack — the modeling lives in adjacent repos, this is where it meets a user.
| Area | What it does |
|---|---|
| Game analyzer | Enter a matchup, get an analysis and a recommendation |
| Live odds | Track lines and movement across the slate |
| Kelly sizing | Translate edge + bankroll into a stake |
| Multi-sport | NBA, NFL, MLB workflows |
| Premium | Stripe subscription + one-time checkout |
| Newsletter | Notion-backed capture wired into Substack |
React · TypeScript · Vite · Tailwind CSS · shadcn/ui
Serverless API routes handle Stripe checkout and newsletter capture so previews and production stay fully functional.
git clone https://github.com/ianalloway/ai-advantage.git
cd ai-advantage
npm install
npm run devOpen http://localhost:5173.
netlify.toml— Netlify SPA deployvercel.json— SPA rewrites for Vercel preview deploysapi/create-checkout-session.ts+api/checkout-session.ts— server-side Stripe checkout
STRIPE_SECRET_KEY=***
STRIPE_PREMIUM_PRICE_ID=price_...
STRIPE_ONE_TIME_PRICE_ID=price_...
PUBLIC_APP_URL=https://your-domainIf unset, the app falls back to static Stripe Payment Links via VITE_STRIPE_CHECKOUT_URL and VITE_STRIPE_ONE_TIME_CHECKOUT_URL.
The homepage form posts to /api/newsletter-subscribe, which creates a Notion entry, emails a notification, and redirects into the Substack subscribe flow.
NOTION_API_KEY=***
NOTION_PARENT_PAGE_ID=...
RESEND_API_KEY=re_...
RESEND_FROM_EMAIL="AI Advantage <onboarding@yourdomain.com>"
NOTIFY_EMAIL=ian@allowayllc.com
SUBSTACK_PUBLICATION_URL=https://allowayai.substack.com| Repo | Role |
|---|---|
| sports-betting-ml | Models — logistic regression, XGBoost, ensembles |
| nba-ratings | Ratings + win-probability library |
| kelly-js | Kelly / odds / bankroll math (TS) |
Ian Alloway — Portfolio · LinkedIn · Writing
Proprietary. All rights reserved.

