The GST rule your receipt scanner doesn't know exists

The GST rule your receipt scanner doesn't know exists

A restaurant receipt and a client dinner look identical to an OCR tool. Under Australian GST rules, they aren’t: entertainment expenses have their input tax credit (ITC) denied by default — GST was charged, but the business can’t claim it back — and nothing printed on the receipt tells you that.

Most receipt scanners stop at “find the vendor and the total.” Ours classifies each line against the actual rules. A ribeye and a glass of shiraz at a steakhouse come back entertainment_denied, with the rationale “entertainment ITC denied, no exception context.” An in-flight meal on an interstate business flight comes back taxable — because entertainment incidental to travel is a recognised exception that flips it back to claimable. Same-looking meal, different rule, extracted correctly either way.

Every line carries the tax_code, a confidence score, and that one-line rationale naming the rule — so you (or your accountant) can check the call instead of trusting a black box. ABN checksum validation is built in too.

Try Receipt Extraction →

This reflects publicly documented Australian tax law and is provided as software output, not tax advice.

Related Posts

A wrong ABN and a smudged one look the same to OCR. We made sure ours doesn't guess.

A wrong ABN and a smudged one look the same to OCR. We made sure ours doesn't guess.

Every Australian Business Number carries a checksum — a modulus-89 formula baked into the digits themselves, so a wrong ABN is usually mathematically detectable, not just "looks a bit off." Which rais

read more
Build vs. buy: should your bookkeeping tool parse receipts with a raw LLM call?

Build vs. buy: should your bookkeeping tool parse receipts with a raw LLM call?

Here's the honest version of the pitch first: if all you need is vendor, date, and total off a receipt, you should probably build it yourself. A single vision-model call — GPT-4o-mini, Gemini Flash —

read more
How accurate is "just ask the model to categorize it"? We measured — 1–2 out of 12.

How accurate is "just ask the model to categorize it"? We measured — 1–2 out of 12.

If you're mapping a product catalog into Google's Product Taxonomy, the obvious first move is to hand the title and description to a capable model and ask for the category. It reads well in a demo. So

read more
AI product classifiers hand you category IDs that don't exist. Here's how we stopped ours.

AI product classifiers hand you category IDs that don't exist. Here's how we stopped ours.

Ask a capable AI model to classify a product into Google's Product Taxonomy and it will do something unsettling: return an ID that looks completely valid — right format, plausible category name — and

read more