About Wazza

G'day. Matt here, the bloke building Wazza.

A mate of mine and I were chatting at the pub in early 2026 about punting. Dave's a pretty experienced punter — the kind of bloke I'd always call on a Saturday to ask what races and horses he likes.

Meanwhile I'd been watching TikToks about how most people aren't using 1% of what Claude or ChatGPT can actually do. So I said to Dave: I'd love to build something, but I've got bugger all good ideas. We kept coming back to the idea of a personal wagering agent — not a generic one, but one that actually knows you. What you're good at. What you're not.

That was it. Wazza was born. I got a Claude Max account and started vibe coding version 1.

That was a cluster f*ck. Claude spun up all kinds of third-party accounts and tried scraping fixtures and odds. A disaster on every level.

So back to the drawing board. Watched a tonne of videos on best practices. Learned the Software Development Lifecycle I'd used my whole career in tech. Applied it properly.

Version 2: properly planned and executed. Hosted on dedicated infrastructure in a quality data centre. Commercial data feeds — not realtime, but ~5 minutes delayed is good enough for most of us.

Wazza's machine learning models train on 12 months of Australian racing data and retrain every day. Picks consider odds, weather, track, recent form, and what's trending on the day — e.g. if Gold Coast picks are hitting at >80%, Wazza knows about it.

There's a dashboard if you like that sort of thing. But the real power is chatting to Wazza. When you sign in he's at your service.

Go ahead, register, and give it a go. Don't be afraid to share feedback — Wazza's new and he's got a bit of growing up to do.