The problem it started as
Gayatri Khatri started Diyashna The Grace as a small ladies' clothing collection — sarees, kurtas and everyday wear, sold out of one shop the way most boutiques start. The business side ran on the same quiet math every small shop runs on: what's on the shelf, what's in the back room, and what's been promised to a customer online. That math lived in three places that didn't talk to each other — a notebook at the counter, a spreadsheet updated when someone remembered, and whatever the last online listing happened to say. A dress could sell in the shop and still show "in stock" online an hour later.
She didn't set out to build a product. She set out to stop that one mistake in her own shop — build a billing counter, a stock list and a point of sale that actually agreed with each other, for Diyashna The Grace and nowhere else.
How it was actually built
A page, not a platform
Her first version was a single offline page kept open at the counter of her own shop — nothing more than a fast way to log a sale and subtract it from a running count, built to survive Diyashna The Grace's patchy internet rather than depend on it.
standalone prototypeLiving on the till, not just in a browser tab
The prototype became an installable app on the shop's own devices — added to the home screen like any other app, so her staff opened it the same way every morning instead of hunting for a bookmark.
installable appTwo records, so one bad connection couldn't lose a sale
Stock and sales data started syncing to two places at once — a fast live layer for the shop floor, and a permanent, spreadsheet-style record as a backup that didn't depend on any single service staying up. If one lagged, the other still had the truth.
dual syncFrom her shop's tool to everyone's platform
Once it worked for Diyashna The Grace, she and Lokendra rebuilt it from the ground up into RetailOS — a proper point-of-sale and back-office system covering products, inventory, customers, suppliers, reports and finance, with eSewa and Fonepay built in alongside cash, and room for more than one seller to share the same stock count.
production rebuildThe fix was never "build software for other people." It was "make my own shop's shelf, till and listing agree with each other" — RetailOS is just what that fix looks like once it's finished, and once it's shared.
The engineer who built it
Gayatri could see exactly what was wrong on her own shop floor — she just couldn't write the code to fix it. That part fell to Lokendra Shah, a computer engineer still early in his career at the time, who took each problem she described from the counter and turned it into something that actually ran: the offline page, then the installable app, then the dual-sync layer, then the full rebuild into RetailOS. Every phase above was her diagnosis and his engineering — her imagination of what the shop needed, made real one build at a time as his own skills grew alongside the product.
That division of labor became the template for how the two of them still work: Gayatri sets the direction from what sellers actually need, and Lokendra — now CTO — builds the engine, DGC Forge, that turns each new idea into a shipped product across the whole ecosystem.
Where the story leads
That's the same principle the rest of the DGC ecosystem is built on. Once RetailOS kept one shop's stock count honest, DGC Bazaar could let that shop's listings appear in a shared marketplace without a second spreadsheet. DGC Concierge AI could answer a shopper's question about stock without guessing. Every product that came after RetailOS inherited its one rule: there is exactly one place stock is true, and everything else reads from it.
Diyashna The Grace still runs on it today — seller number one, on the same ledger every other seller now shares.