Bytez

Inspiration

I’m the person in the relationship who can usually “read” a menu and pick the one dish that hits. My girlfriend… not always. She’ll order something that sounds great, and it shows up like, and she suddenly wants mine.

My team realized the problem isn’t finding a restaurant—it’s finding what to order at that restaurant. Most apps tell you where to go. Bytez is built around the question people actually ask at the table: “What dish should I get?”

What it does

Bytez is a dish-first food review app for nearby restaurants.

Discover nearby restaurants on a map + scrollable list
Create dish-level reviews with a rating, tags, and optional photos
Search within a restaurant by dish
- Get dish suggestions as you type
- On an exact match, see a Dish Summary (avg rating, top tags, latest photo) above the individual reviews
Save (favorite) reviews to revisit later
Profiles with avatar upload + a simple appearance setting

How we built it

iOS (SwiftUI) for the app and modern UI (cards, motion, brand-consistent theming)
MapKit for nearby discovery + restaurant search (MKLocalSearch, map pins, restaurant detail pages)
Supabase
- Auth for sign up/sign in (email/password + Apple + Google)
- Postgres tables for profiles, restaurants, reviews, and saved_reviews
- Storage for review photos + avatars

Challenges we ran into

Supabase relationship ambiguity when joining reviews ↔ profiles required explicit foreign-key selection in queries.
Email confirmation flow made sign-ups look “broken” until we added a dedicated pending-confirmation UX.
Map search deduping + rate limits: MKLocalSearch can return overlaps across queries, so we deduped results and throttled requests.
Demo reliability: we kept a consistent fallback map center (College Station), because iOS Simulator can't use your actual true location

Accomplishments that we're proud of

The Dish Summary experience: search a dish inside a restaurant and instantly get a “quick verdict” before reading every review.
A real end-to-end product slice: discover → review w/ photos → view everywhere → favorite → profile.
A modern, living UI that feels intentional and shippable, not a hacky prototype.
Practical polish like avatar caching and a fullscreen image viewer with zoom/pan.

What we learned

People don’t want “more reviews”—they want faster confidence. Dish-level aggregation beats endless scrolling.
Backend details directly affect UX: auth edge cases and database joins can make an app feel seamless or broken.
Map apps are as much product design as engineering: what loads instantly, what’s “nearby,” and what needs graceful fallback.

What's next for Bytez

Smarter dish matching (aliases/synonyms, not just exact titles)
A “What should I order?” ranking per restaurant based on consistency + review volume
Filters by tag (spicy, healthy, late-night) and stronger dish discovery UI
Shareable dish cards for friends before you go
Fully live location-first onboarding and scaling beyond demo mode