80%-Win AI Betting App

I’m building a two–phase solution that lets everyday punters open an app, pay a small subscription, and instantly see the highest-edge bets of the day—together with the bookmaker that offers the best line. Before we put anything in users’ hands, the underlying model must first prove an 80 % hit rate on historical and live data.

Phase 1 – Model & Tracking
I need an AI that simultaneously ingests football (soccer), basketball, horse-racing, and any other feeds you recommend, then surfaces whatever market offers the “best potential bet.” You are free to combine game-winner, player props, totals—whatever the numbers say is most exploitable. During this phase the system must log every recommendation in a transparent database so we can publish the rolling strike rate; only when it sits at 80 % or better will we move on.

Phase 2 – Subscription App
Once the threshold is met, wrap the model in a clean mobile (or responsive web) interface that:
• shows today’s top picks ranked by edge,
• names the sportsbook offering the softest odds, and
• gates access behind a monthly in-app purchase or Stripe subscription.

Core tech is up to you—Python, R, TensorFlow, PyTorch, scikit-learn, even proprietary tools—as long as the final API is well-documented and the UI is fast and intuitive.

Timeline
I’d like a fully functional MVP within a month, with interim demos so we can course-correct early.

Deliverables
• Trained model with reproducible training pipeline
• Live tracking dashboard proving the 80 % win rate
• Production-ready app (code + UI/UX) connected to real-time odds feeds
• Deployment scripts and brief hand-off session

If you’ve tackled predictive sports models or betting-edge apps before, tell me how you measured accuracy and handled odds provider integration—I’m eager to get this in front of users quickly.