Khatapata — Retail & Inventory Platform

The problem

A small shop in Nepal keeps its records on paper. Stock is counted by looking, sales are remembered, and credit — the khata — is a book of names and amounts that only the owner can read.

It works until it does not. Nothing reconciles, the same product sells at different prices depending on who is at the counter, and the question that matters most is the one the paper cannot answer: which of these products actually makes money. Software exists for all of this, and most of it is built for a business that already has a barcode scanner, a card terminal and someone whose job is data entry.

What the platform does

Point of sale first, because that is the part a shop uses every hour: checkout on any device, split payments, discounts, refunds, printed or digital receipts.

Around it, the things the counter creates a need for — real-time stock across locations with low-stock alerts and reordering, role-based access separating owners, managers, cashiers and staff, with shift tracking; a customisable online storefront and marketplace publishing from the same catalogue; and dashboards for revenue, best-sellers and staff performance.

The credit ledger is the part that makes it Nepali rather than generic. Recording who owes what, against which sale, and what has been paid back is the workflow the paper book was doing, and no imported POS has a place to put it.

What I owned

I worked on Khatapata as a full stack developer, across the Next.js and TypeScript front end and the Supabase data layer behind it.

That covers both halves of the problem this product has: the counter, where speed and getting out of the way are the whole job, and the back office, where the ledger, stock and reporting have to agree with each other after the fact.

Why it is built the way it is

The constraint that shapes everything is who is holding the device. This is a shop owner or a cashier, often on a phone, frequently mid-transaction with a customer waiting — not someone trained on a system.

So the product has to work at counter speed, on whatever hardware is already there, and it has to be adoptable one piece at a time. A shop that starts by replacing its paper khata and nothing else has to get value from that alone, or it will go back to the book by the end of the week.

Built on Next.js and TypeScript with Supabase and React Query — a stack chosen to keep one small team shipping across the counter, the back office and the storefront rather than maintaining three separate products.