Side project · 2026

Groundwork

An AI-native property discovery app that retrieves and interprets development application (DA) data to provide home buyers with a clearer picture of any property.

Tools + Tech

Claude CodeClaude Code
Claude APIClaude API
PostGISPostGIS
ReactReact
GW—001 · Structural Field
Issued 06 · 2026 — Drawn EW
Scale 1:200
Rev NY
010 m
Yaw 324°
Elev 06°
Early brand design explorations. Scroll and drag to explore.
Context

Groundwork is an AI-native property discovery app for Australian home buyers.

It pulls development applications from local government websites and interprets them in plain English, so that buyers can quickly and easily understand what's being built around a property they are considering.

Problem

The primary property apps in Australia are built to serve agents. I wondered: what would a property app designed to serve buyers actually look like?

A key challenge when buying property is understanding what is going on around a prospective property, and what impact it might have. Is there an apartment building going up across the road that will block your morning sunlight and change the character of the neighbourhood?

This information is freely available — but it's fragmented and hard to understand. Buried in government portals, spread across council websites, and written in architecture and legal speak that the average person can't parse.

Solution

Groundwork puts the right information in front of buyers at the moment they need it most — while they're actively searching.

It works two-fold:

  • Aggregates DA data and breaks it down in plain English — what's being built, how close it is, and what it means for your sunlight, views, privacy, and neighbourhood character
  • Makes this information searchable and accessible at the point of decision, not buried in a government portal nobody knows to check
Implementation

I started with first-principles research — mapping the current buyer journey, auditing existing tools, and stress-testing product-market fit. From there, I have been working with Claude Code to bring Groundwork to life.

Scrapes local government websites
To find relevant development application (DA) data
Stores information in database
Tagged with relevant metadata
Leverages PostGIS
To handle proximity queries — finding every DA within 500m of a given address
Interprets DA data with AI
Through a Claude API call and specialised prompt that synthesises DA information and impacts
Provides search functionality
So that users can search property addresses and understand implications of nearby DAs