Cross-Platform Racing Betting Bot

I need a single application that runs smoothly on iPhone, iPad, and Windows and can log in to my TAB account, detect the instant fixed-odds boards go live, and fire the bets automatically. The program must handle all three codes—harness racing, gallops racing, and greyhounds—without me having to switch screens or data sources.

Odds Assessment & AI engine
Alongside the live prices, the app should generate its own “true” price for every runner by crunching horse form, driver / rider performance, gear changes, plus any other variables you believe lift accuracy. The model can be built with Python (scikit-learn, TensorFlow, or similar), but it must expose the result in the front-end instantly so the staking logic can compare AI price to TAB price before firing.

Interface essentials
• Grid view for every race with columns for TAB price, AI price, and a field where I can overwrite with my own price if I choose.
• Running market percentage tallied at the bottom of each race.
• Real-time profit & loss ledger that updates once bets are matched and again when results are posted.

Speed & stability targets
• Price polling frequency tight enough to capture the first fixed-odds line (sub-second ideal).
• Secure credential storage and robust error handling so a single API hiccup never places an unintended bet.
• Consistent UX across iOS and Windows—Flutter or React Native is fine, but I’m open to separate native builds if performance demands it.

Deliverables
1. Compiled apps for iPhone, iPad, and Windows.
2. Source code and model files.
3. Setup guide plus a short Loom walkthrough showing the bot firing on a live market.
4. Basic test suite proving odds capture, AI calculation, bet placement, and P&L tracking.

I’ll supply TAB credentials and any historical data you require for model training once we start. Let me know how you’d architect the cross-platform stack, which machine-learning framework you favour, and the timeline you’d need to reach a working beta.