About Plinth
Almost everyone selling online is one person with a phone. Plinth is built for them.
We kept meeting the same person. She sells beautifully — skincare, or streetwear, or small-batch food. Her customers find her on Instagram and message her on WhatsApp. She answers every one herself, quotes the price, takes the order in a notebook, and packs the parcel. She is the storefront, the cashier, the courier desk and the support line.
When she looks for software, she finds tools built for a different company: a store builder that assumes a marketing team, a chat plug-in sold separately, a bot that follows rules and cannot close a sale. Three subscriptions, two dashboards, nothing speaking to anything else — and a percentage of every order taken off the top.
So the question we set out to answer was narrow and stubborn: what would commerce software look like if it assumed you were alone, sold in a chat thread, and could not afford to give away a slice of every sale?
Plinth is our answer. One system — a real storefront, a real checkout, one order book — and an AI operator that works the channel the sale actually happens in. Not a store with a chatbot bolted on. A store that can sell without you awake.
Store builders were designed for merchants with a marketer, a support desk and someone to reconcile orders. Most sellers are one person with a phone, answering every message themselves between packing parcels.
In India, Southeast Asia and the Gulf, the shop floor is a WhatsApp thread. That is where the price is asked, the size is checked and the sale is closed — and it is the one place commerce software has never really gone.
A percentage of every sale costs a large brand a rounding error and a small seller their margin. Commission is the most regressive way to price commerce software, and it is the industry default.
Each of these is a decision already made in the product, not a value on a wall.
We charge a subscription, never a share of your sales. Grow ten times and you pay us the same. It is the single clearest way to make our incentive your volume, not your dependence.
Our operator proposes typed actions that the runtime validates against your real catalog, your real stock and your real prices before anything is sent. It cannot invent a product, a price or an order. An answer it cannot ground, it does not give.
Four levels, switched per store — from the AI asking permission for every message to running a thread end to end. You move up as you learn to trust it, and drop back whenever you want.
Payment happens on a secure link. Card details and delivery addresses never enter a chat thread, and the AI has no access to either. Refunds are escalated to a human, always.
A question at 11:41 pm is answered in seconds, from live stock, at the price you set.
The website and the chat write to the same orders. Nothing is reconciled by hand at midnight.
Discounts, gift cards, bundles, abandoned-cart recovery, order timelines and tracking.
Premium themes, fully editable, every one a real working store rather than a template screenshot.
One subscription that replaces a store builder, a WhatsApp tool and a chatbot add-on.
We are at the beginning, and we would rather say so.
Plinth is early. The storefront, the checkout, the commerce features and the WhatsApp AI operator are built and running — you can open any of our storefronts and shop them right now. Self-serve billing is not built yet, so we set every store up with the merchant ourselves, over WhatsApp.
We are taking a first hundred merchants at half price, locked for life, in exchange for honest feedback while we finish. There is no waitlist theatre and no invented traction on this page: when we have numbers worth publishing, we will publish them here.
The ambition is not modest. Chat-led commerce is how a very large share of the world already buys, and almost none of it is served by software built for it. We intend to be that software — starting in India, Southeast Asia and the Gulf, where the behaviour is furthest ahead of the tooling.