Sweet As App: How AI Restaurant Discovery Works in Sydney

You want somewhere cosy for a date, a coffee after 4pm or a lunch spot that is not the same place you visited last week. Knowing what you feel like is easy enough. Turning that into a venue can mean bouncing between maps, saved Instagram posts and a group chat that has somehow become a committee. The Sweet As app is built around that gap between having an idea and knowing where to go.

The Sydney food discovery platform combines AI-powered search with shareable collections called Eatlists. Instead of beginning with a cuisine category, the idea is to describe the experience you want, then discover restaurants, cafés, bars and other venues that fit. It is a practical use of AI aimed at a very ordinary problem: choosing somewhere to eat without making the search the main event.

Speaking with TechDrivePlay, co-founder Tara Williams explains that the motivation was familiar. People have plenty of venues around them, but the effort of finding something new often sends them back to the same favourites. Sweet As is intended to make that discovery feel more personal, with natural-language searches and recommendations people can save, organise and share.

What is the Sweet As app?

Sweet As is a food and drink discovery app that Williams says is currently in public beta across Sydney. It is free for people to use, with a growing selection of venues across the CBD, Inner West, Eastern Suburbs, Northern Beaches and surrounding areas. Its focus is on helping someone decide where to go, whether they are a local looking beyond their regular spots or a visitor trying to make sense of an unfamiliar neighbourhood.

“We wanted to make discovering hospitality fun again,” Williams says. “Less scrolling, less overthinking, and more finding somewhere that actually matches what you’re in the mood for.”

That intention shapes the two main parts of the experience. Search starts with what someone wants from an outing, while Eatlists let people collect venues around an interest or occasion. A useful way to think about the app is as somewhere to turn a vague plan into a shortlist, then keep the places worth remembering for another day.

How AI restaurant discovery works

Williams wants people to search in the way they would describe a plan to a friend. Her example is “a cosy wine bar in Newtown for a first date”. There is more in that request than a venue type: it names a location, an atmosphere and an occasion. Sweet As is designed to look for venues matching that combination, rather than leaving the user to translate every preference into a separate filter.

A request for coffee after 4pm has a different practical requirement. The venue needs to suit the time as well as the craving. Williams says the app considers food, atmosphere, location and opening hours when looking for matches. If nothing fits closely, the aim is to suggest nearby alternatives in terms of the request, rather than leave someone with a dead end.

Her examples explain the intended search experience without requiring the reader to understand an AI model. A date-night search might help someone explore quieter or cosier venues; a coffee search might begin with when they actually need a café. The point is to give the search more useful context than a single category such as “Italian” or “bar”.

Sweet As’ published privacy policy adds some detail about the technology. It says AI analyses venue material such as menus, images, text and social posts to create searchable attributes, and uses preferences, location and behavioural signals for matching. That describes the platform’s approach; Williams’ examples show how the idea is meant to translate into an everyday search.

Eatlists turn saved venues into plans

Eatlists are collections of places someone loves or wants to try. Williams suggests building them around date nights, rainy days, favourite coffee spots or places to take visiting parents. Rather than having a single saved folder containing everything from a quick lunch to a special-occasion dinner, users can organise venues around the reason they might want to visit.

Think of Eatlists as playlists for places to eat and drink. A playlist makes individual tracks easier to return to as a collection; an Eatlist does that for venues. People can save places, explore other users’ lists and share their own, allowing discovery to continue through someone else’s interests and recommendations.

For a local, that could mean keeping a list of coffee spots away from the busiest streets. For someone visiting Sydney, it might mean collecting places around the neighbourhood where they are staying. These are examples of how to organise a list, rather than ready-made itineraries. The benefit is having a reason for each saved place, so the collection remains useful when the next “where should we go?” conversation begins.

Williams describes Eatlists as a way to move beyond another generic top-ten list. The appeal is the context behind a recommendation: a place someone saves for a quiet catch-up may suit a very different outing from the venue they recommend for a lively evening. That context gives people another way to explore, alongside the app’s search.

Where restaurant details come from

An atmosphere-led search still needs practical venue information behind it. Williams says Sweet As works directly with businesses to understand more than their menu, including the setting, experiences and details that bring people through the door. Venues can manage their own profiles, including menus, specials and other information, with support from the Sweet As team.

That gives businesses a role in how their information is maintained. It also helps explain why the app is interested in more than a cuisine label: two venues serving similar food might offer very different experiences. Opening hours and menus can change, so checking the venue’s current details before heading out remains a sensible final step, particularly when timing is the reason for the search.

How Sweet As fits alongside Google Maps and Instagram

Williams does not describe Sweet As as a replacement for every tool people already use. In her view, Instagram is useful for inspiration and Google Maps is useful when someone knows what they are looking for. Sweet As is aimed at the moment in between, when the person knows the kind of experience they want but has not chosen a venue.

“We’re for that moment when you don’t know where you want to go yet,” she says. “You just know how you want to feel, what you’re craving or what kind of experience you’re after.”

Her suggestion for a first search is simply to type what is on your mind, even if it feels unusually specific. That is a useful introduction to the concept: begin with the plan you are trying to make, rather than trying to guess the right keyword. A place found through a social post can still become part of that process, with an Eatlist providing somewhere to keep it for later.

Is Sweet As free, and how does it make money?

Williams says the app is free for diners and operates without paid ads, sponsored posts or commissions. Venues have a free Community option, alongside paid plans offering additional tools, insights and analytics about searches and discovery. She describes the business model as selling useful services to venues, rather than charging them for each customer referred.

“Because we don’t believe a venue should have to pay to be someone’s perfect match,” Williams says of that approach. Her stated intention is for discovery to follow what someone is looking for, rather than the size of a venue’s marketing budget. It is worth distinguishing the two parts of the model: paid venue tools exist, while Williams says recommendations are not sold as advertising placements.

Sydney availability, bookings and privacy

Williams describes Sweet As as a growing Sydney public beta, with the experience continuing to develop as venues and users join. The official Sweet As access page is the place to check current sign-up information. The website currently uses beta/TestFlight wording, so check its latest access instructions for your device.

Once someone finds a venue, Williams says its available links can take them to booking, ordering or further information. The app helps with discovery, then connects people with the venue’s own booking and ordering options.

On personalisation, the privacy policy says location is collected with consent, and describes the use of account information, preferences, searches and engagement data. It also says users can withdraw consent for location processing and request access, correction or deletion of personal information. The policy provides the fuller explanation of those choices and how to contact the business.

Tara’s Sydney suggestions, and what comes after Sydney

For a sense of the discoveries she hopes the app will encourage, Williams points to Ravenspur in Newtown and Soul Brew. She highlights Ravenspur for cocktails, an easygoing team and things happening at the venue, while describing Soul Brew as a quieter coffee or brunch option away from the crowds.

They also give the Eatlist concept a more concrete shape. Williams suggests themes such as “A little different” or “Coffee away from the crowds”, where the reason for choosing a place is part of the collection. The list becomes a starting point for an outing, rather than simply another set of restaurant names.

Melbourne is an obvious future possibility in Williams’ view, but she has not announced an expansion date. She says the immediate focus is getting Sydney right, growing the community and making something people want to return to. Other cities and features remain ambitions to share when they are ready.

The interesting idea here is small enough to recognise immediately: sometimes you know exactly what you want from a meal out, just not the name of the place that offers it. Sweet As is building its search around that description, with Eatlists helping people remember what they find. A cosy date, a late coffee or somewhere to take your parents is already a perfectly sensible place to start.