FastChow.

The goal was simple.

The biggest focus of the rebuild was performance, and most of it came down to render work rather than network speed.

The home feed was janky at first. The images and the list library were not the cause; the app was repeating work it did not need to. Every card on screen was being rebuilt while you scrolled, because of how the list was wired together. Once each card only redrew when its own content changed, the scrolling smoothed out. That was most of the fix, before any library got swapped.

The rest is the same idea in different places. Screens show what they already have and refresh underneath, so going back somewhere is instant instead of a spinner. Cards reuse the same views as you scroll instead of building new ones. Tabs load when you first swipe to them and stay loaded, with the next one warmed up in the background, so swiping between search tabs does not stall.

The bag works this way too. It lives on your phone first. Tapping plus updates the screen straight away and the server catches up a moment later, in one batch, instead of a request per tap. You are never waiting on the network to find out whether your own tap counted, and rapid taps cannot fight with a stale reply.

The difference is most noticeable on lower-end Android devices, where a poorly optimised interface turns sluggish quickly. The objective was not to make the app technically faster. It was to make it feel faster.

The tab bar is the most-tapped thing in the app, so it should never be somewhere you wait. Home, Orders and Account switch the instant you press them, and they hold the place: what you had loaded and where you had scrolled to is still there when you come back.

You can hammer between all three and it keeps up. No flash of empty screen, no reloading something you were already looking at a second ago.

App updates are usually disruptive. You get a modal you cannot dismiss, or a trip to the store, and it lands in the middle of whatever you were doing.

Here, an update shows up as a banner above the feed. You can ignore it and keep going. If you take it, the new build downloads in the background while you carry on using the app, and the banner turns into a restart when it is ready. The update stays clear and easy to finish without taking the app away from you first.

It ships over the air, so a fix can reach people the same day instead of waiting behind a store review.

Search is one of the most important features in a food delivery app, and one of the easiest to get wrong. It is also the part of the current app people complain about most.

The empty state carries recent searches and a set of categories, so the common case is a tap instead of typing. Queries are debounced and cancelled as you type, so fast typing does not queue up stale requests. Results stay on screen while the next set loads instead of blanking out. Results come back grouped by store, and items can be added straight from the list.

The result is search that works the way it should on mobile: type, find, move on.

Favouriting is a simple interaction, but it has to feel reliable. In the current app it can be inconsistent, and you are left unsure whether the tap did anything.

Here the heart fills the moment you press it, and the request goes out behind that. If it fails, the heart goes back. Your favourites row on the home screen updates at the same time, so you can see the place you just saved without going looking for it. Small interaction, big difference in how the app feels.

Keeping up with an order is one of the biggest sources of friction in the whole product. Normally you end up navigating back and forth between screens to find out whether anything has changed.

The rebuilt order screen keeps itself current. Status, the rider on a map, and the full history of everything the order has been through, all in one place. When I am looking at my order, I should be able to trust that I am seeing the current state.

Funding a wallet at checkout has the same problem as the order screen. You add money, and then you are moving between screens, or closing and reopening the app, just to get the balance to catch up.

I redesigned it so the balance updates in the flow you are already in. You top up, you see it land on the checkout screen, you place the order.

Chowdeck has this already. I built my own version because the wallet felt incomplete without it, and because it is the thing that makes a balance make sense: top-ups, order charges, refunds, partial payments and gift card redemptions, grouped by month and filterable by balance type.

Choosing where your food goes is a small screen that gets used constantly, so it should be quick and quiet. Saved addresses are right there, adding a new one is a search field and a couple of taps, and switching between them does not reload the world around it.

The map that does exist, on an order that is already on its way, gets the same treatment. In the app I was using this is one of the clunkier areas: taps lag behind, panning stutters. Mine stays smooth.

I stayed as close to Chowdeck's look as I could without treating it as a spec: the greens, the accent colours from the category tiles, the chunky corner radii, and Plus Jakarta Sans throughout. Some of it is yours, some of it is my read on it. This page is set in the same colours and type.

Two and a half minutes of the app, unedited, running against the live API.

Everything the app knows how to ask for lives in one place: a separate package of 16 modules, one per area of the product, with every response checked as it arrives. If Chowdeck changes something on the side, it fails loudly in one spot instead of quietly breaking a screen.

The app on top of it is Expo and React Native, so iOS and Android come from the same codebase, and updates can go out over the air.

The biggest improvements here are not new features. They are the moments people repeat constantly: searching, scrolling, switching screens, favouriting, checking an order, funding a wallet, waiting for an update.

The idea was simple. Remove the waiting, remove the unnecessary navigation, and make every interaction feel immediate.