Opening Screen
Stock out Window
Selling Window
Stock In Window
Console View

Inspiration

My uncle runs a kirana (neighborhood grocery) shop. During customer rushes he scribbles sales on a paper notepad, then spends late evenings after closing typing them into Excel. Often he only finds out an item is finished when a customer asks for it. I wanted a tool that works at counter speed and replaces the notepad.

What it does

Sopko is a mobile-first stock and sales tracker for small shops.

Stock out: tap an item, set the quantity, tap Sell. Stock drops immediately and today's total updates.
Oversell protection: the quantity can't go above what's in stock, and the app says "Only N left in stock".
Stock in: log arriving goods with quantity and cost, or add a brand-new item.
Low-stock and out-of-stock badges show what needs restocking at a glance.
One-tap CSV exports (sales, stock in, current stock) open cleanly in Excel or Google Sheets and double as a backup.
A sample data button loads a realistic kirana inventory for a quick demo.

It works with no server and no login. Data stays in the browser.

How we built it

I built it with the Devpost Learn skill pack in Antigravity. Instead of asking for code first, I let the AI interview me and produced scope.md, prd.md and spec.md before any code was written. The app uses React and Vite, a single storage module that keeps all data in local Storage, and custom CSS design tokens for the look. I built it in four working slices and tested each one in the browser before moving on.

Challenges we ran into

Setting up Node and PowerShell on a new Windows laptop (PATH and script-policy errors) took longer than expected.
Making the sell sheet stop at the available stock and explain why.
Fixing layout overflow on small phone screens (320 to 360px).

Accomplishments that we're proud of

A complete working loop in a few days: receive goods, sell items, see low-stock warnings, and export the data.
Keeping the scope small on purpose. I cut logins, a backend and barcode scanning so the core could be finished and tested.
Designing for a real person at a busy counter: big tap targets, clear status badges and one-handed use.

What we learned

Planning first changed what I built. Deciding what to cut kept the first version small enough to finish. I also learned to check each slice myself in the browser instead of trusting the output, which is how I caught layout and wording problems early.

What's next for Sopko

Voice input with a confirmation step so a shopkeeper can say "Maggi 2" instead of tapping, a 7-day sales chart, and top-seller and slow-moving item alerts.