It always started the same way. Six o'clock, the day's energy depleted, someone asking where do you want to eat. Then twenty minutes of opening Yelp, opening Google Maps, scrolling, comparing star ratings, opening a fourth-party menu PDF, closing it, considering takeout, abandoning takeout, ending up in the same neighborhood pizza place we always end up in.
The tools were the problem. Every restaurant app had quietly turned itself into a feed. Endless options, filterable along eleven dimensions, sortable by recency, popularity, distance, price, dietary restriction — and worse, every option was fine. None of them were obviously wrong. So you'd just keep scrolling, looking for the option that would let you commit. It never came.
The "feed" is the bug
A feed is the right interface for things you don't have to act on — news, photos, jokes from friends. It's the wrong interface for a decision that has to end in get in the car. A feed defers commitment. It tells you, implicitly, that there's always something better one more scroll down. That's exactly what you don't need at six pm on a Tuesday.
What we wanted was the opposite shape: a single button that hands you one suggestion and walks away. Not here are 47 nearby restaurants, sorted. Just: go here. If you hate it, hit the button again.
The hardest part wasn't the picking
Apple's MapKit will happily hand you a list of restaurants within a radius. That's an afternoon. The hard part was deciding what to leave out. Every instinct as a developer is to add a filter, a knob, a customization. We resisted almost all of them.
The free version of Tonight's Table has exactly four controls: where you are, how far you'll go (up to 10 miles), whether to hit GPS, and the big button. That's it. No cuisine picker, no price range, no opening-hours toggle, no map view. We watched ourselves want to add each of these and forced ourselves to ask: does this make a six-pm-Tuesday decision faster, or slower?
If a feature gives you something to think about, it's the wrong feature for this app.
The visited log was the one exception. Over a few weeks of dogfooding it became obvious: the random pick wasn't useful if it kept landing on the same five places. So we built a way to mark places as "been there, done that," and a toggle to skip them next time. That's it. Even that lives behind Pro, partly because it was real work to build, mostly because everything else stays free.
What Pro is for
Pro is $0.99/month or $4.99 once. It unlocks every filter — the cuisine style picker (when you really do want pizza), the wider search radius (free is locked at 10 miles; Pro goes up to 45), and the "skip the places I've already been" toggle. It exists for two reasons: to support indie development, and to nudge people toward the free tier first.
We do not run a server. We have no analytics SDK. We don't know who you are, where you searched, or what you ended up eating. We genuinely cannot tell you which features people use the most — that's a problem we're happy to have.
Tonight, just go
The product idea isn't novel. People have been writing "I can't decide where to eat" Reddit posts since Reddit existed. What we hope is novel is the discipline — refusing to dress up a one-button app as a feed, refusing to track anyone, refusing to add the cuisine filter to the free tier just to drive engagement metrics.
Stop scrolling. Start eating. That's the entire pitch.