Skip to content

Our products

We build for clients. We also build for ourselves.

Two products, designed and operated in-house. They are where our opinions about engineering come from — and the reason we can be specific rather than theoretical about what shipping costs.

KisanGyan

Live on Google Play

kisangyan.in

Mitra, the AI inside KisanGyan, speaks the farmer's language, remembers their land, and advises across seasons — not just today.

KisanGyan is our own product, live on Google Play. It puts an AI agronomist named Mitra in the pocket of farmers across India — and Mitra speaks, rather than expecting a farmer to type.

Most agricultural advisory answers one question: what should I do today? Mitra answers the harder ones. What should I grow this season, given my soil and the monsoon forecast? What should follow it next year? How do I rebuild soil that has been declining for three seasons? It remembers the farmer's name, their family, and the history of their land.

Nivista

In beta

nivista.in

Photograph an empty room and see it furnished in a regional Indian style — vastu-aware, budget-aware, and generated in seconds.

Nivista is our second product, currently in beta. Point your phone at an empty or half-furnished room, choose a style and a budget, and get back a photorealistic interior a few seconds later.

The reason it exists is that global virtual staging tools produce homes that no Indian family recognises. They do not understand a pooja space, they do not know that a Kerala home and a Jaipur home want different things, and they have never heard of vastu. Nivista treats all of that as core product logic rather than a cosmetic filter.

Why it matters to you

Running our own software changes the advice we give.

Everything below is something we learned by paying for it ourselves, not by reading about it.

We know what launch actually involves

Play Store data-safety forms, staged rollouts, forced-upgrade flows, review rejections. Our estimates include the fortnight that most proposals quietly omit.

We know what things cost to run

Inference cost per user, cloud bills at low volume, the price of a bad notification policy. We can model your running cost because we model our own.

We know which shortcuts are safe

Some corners can be cut for years without consequence. Others cost a rewrite. We have taken both kinds and can tell them apart.

Want something like this built for you?

The same team that built and operates these products is available for client work.