G1
GOLDMINE.

G1 Goldmine has been live since 2015. Ten years of production code. A paying customer base of thoroughbred breeders across Australia, Europe, and North America. A database of every Group 1 stakes winner in the world, mapped against pedigree patterns going back generations.

It's the kind of SaaS most agencies wouldn't touch. Not because the product is bad — because rewriting a live revenue-generating platform is genuinely the hardest kind of engineering work there is. You can't ship breaking changes. You can't lose data. You can't take the system down for a weekend to migrate it. Real breeders are using it to make real decisions about real matings — every day.

In 2025, G1 Racesoft asked us to take it on.

We'd already shipped Stallion Match for them in 2022 and stayed in that codebase as their long-term technical partner. G1 Goldmine was the older, bigger, more critical platform — the one their revenue runs on. Handing it over wasn't a small decision.

Learning the domain — again, deeper

Three years on Stallion Match gave us a strong foundation in thoroughbred breeding. G1 Goldmine goes deeper. Pedigree analysis. Bloodline tracking across generations. Stakes-winner correlation. The 4x4 Impact Analysis system that lets breeders see what happens when specific sires and dams are combined.

The platform isn't just data — it's a model of how breeding decisions actually get made at the top of the sport. We're spending the time to understand the model, not just the codebase.

Modernizing the application layer

The product was originally built on .NET with a Joomla CMS layer and a React frontend bolted on later. Ten years of architectural decisions, each made for good reasons at the time, layered on top of each other.

We're rebuilding the application layer on NestJS and NextJS — modern, maintainable, and aligned with how we work across our other long-term partnerships. The frontend gets a proper component system. The backend gets a clean API surface. The deployment moves to AWS so the infrastructure can grow with the product.

Keeping MSSQL — deliberately

Not everything gets replaced. The data layer stays on MSSQL.

Ten years of breeding data, pedigree relationships, and stakes-winner correlations live in that database. The schema works. The queries are tuned. Migrating to a different database engine would introduce risk we don't need to take, for benefits the product doesn't actually need.

Modernization isn't replacement. It's knowing what to keep and what to rebuild.

Modernizing without disruption

This is the hardest part of the work. Real customers are paying to use G1 Goldmine right now. They expect their reports to load, their pedigree searches to return results, their subscriptions to renew.

So we're modernizing while shipping. New architecture goes in behind feature flags. Migrations happen incrementally. The user doesn't see the rebuild — they see a platform that keeps working and gradually gets faster, cleaner, better.

Building the AI layer

Alongside the modernization, we're actively building AI-driven analytics into the platform. Pedigree pattern recognition that goes beyond what the current rule-based system can find. Suggested matings that surface combinations a breeder might not have considered. Insight generation across the global stakes-winner database that would take a human analyst weeks to produce.

The AI doesn't replace the breeder's judgment — it surfaces the patterns the breeder uses to make better calls. Same principle as the rest of how we work.

The work is in active progress. The platform continues to serve its existing customer base while the underlying architecture gets modernized in the background. AI capabilities are being shipped incrementally — not as a separate product, but as a layer that compounds with the rest of the platform's capabilities.

This is the kind of work that takes years, not months. We're planning to be in the codebase for the long version.

Most agencies do greenfield work. New products, new codebases, the comfortable kind of building where nothing is on fire if you break something.

Modernizing a 10-year-old production SaaS is the opposite kind of work. It needs a team that understands the domain deeply enough to know what's load-bearing. That respects the decisions the original engineers made, even when those decisions need to change. That has the discipline to ship modernization incrementally instead of as a big-bang rewrite. And that knows the difference between modernizing for its own sake and modernizing because the business actually needs it.

G1 Racesoft trusted us with their flagship platform after seeing what we did with their newer one. That's the version of long-term partnership we believe in — clients who hand us more of their business, not less, the longer they work with us.