Curateria is a Melbourne-first dining guide built on AI that reads the Google reviews; the Curateria editorial team picks and publishes the lists.
Every “best Italian in Melbourne” list reads the same: ten restaurants, half of them paid, the other half copy-pasted from a 2019 listicle. Someone's algorithm decided generic was safer than specific, and now nobody can tell you whether the pasta is actually good. Just that the room is “vibrant.”
Meanwhile, the people who actually know (the ones eating six nights a week, comparing kitchens, watching menus shorten and prices creep) write in scattered places: long Reddit threads, restrained Google reviews, Instagram captions nobody scrapes properly. The signal is there. It's just not being read.
So we built what we wanted to read. An AI that reads the Google reviews across Melbourne, turns them into a single score per restaurant, and hands the output to a small editorial team who turns it into a digest, over a hundred curated lists, and a search that finally understands what “Korean in Richmond, Tuesday lunch, under $40” actually means.
Our model reads the Google reviews across Melbourne, with fake five-star reviews filtered out before anything is scored.
Google reviews, fake five-stars filteredWe take the average of a restaurant's Google star ratings, drop the reviews our model flags as fake, and apply a volume weighting that can only pull a score down when there aren't many reviews yet. One number, nothing bolted on.
Read the methodology →The score is a starting point, not the answer. Curateria Editorial · an AI corpus reader plus a Melbourne-based founder · argues with the model and writes the lists you actually want to read.
Curated lists · 1 cityOr browse our curated lists, written by humans, scored by AI.