Rubbish Day

Constraint

Each council exposes collection schedules through different endpoints, formats and failure modes. There is no universal waste-collection API.

Decision

Build source-specific integrations per council, then normalise everything into a single, address-first answer.

Proof

1,000+ monthly sessions across NZ and AU, no account required, runs for about $10 a month.

---

The problem

I live in Titirangi. Auckland Council runs a fine service. But finding out which bins go out when required navigating to a council page that, charitably, felt clunky.

Dropdowns. A separate calendar view. Colour-coded zones that did not match the physical bins. I would get halfway through and give up. The red bin would stay inside. My neighbours would judge me silently.

I wanted a more certain answer — not a reminder based on memory, but a data-driven way to know what was actually due at a specific address, accounting for interruptions like public holidays.

The build

I found a semi-documented endpoint that Auckland Council’s own site was already using — not a public API, no developer documentation, but it returned JSON and mapped zone IDs to collection dates. I had a weekend. I built a prototype in an afternoon.

For about a year it ran quietly inside Home Assistant. When Auckland Council’s integration broke and I kept waiting for a patch that did not arrive, I learned n8n and rebuilt it myself. When colleagues started asking about it often enough, I rebuilt it again — this time as a public web app.

The hard part is not the interface. It is the source landscape. There is no neat universal API underneath council waste data. Each provider has its own habits, endpoints, formats and failure modes. Every integration had to be worked out separately. That is still true as coverage expands across Aotearoa and into Australia.

The result

Rubbish Day now handles over 1,000 sessions a month. It does one thing: tell you what goes out, and when. The whole thing runs for about $10 a month. The council has not complained about the API usage, which I take as tacit approval.

The narrowness is deliberate. Useful software does not need to feel large. The best sign a product is well-scoped is when you struggle to think of anything meaningful to add.