theNasty puts marine parts inventory into the conversation. Customers describe the repair — ChatGPT finds the parts that fit, prices the cart, and checks out without ever opening a storefront.
What your customers see — theNasty's inline card, rendered by ChatGPT.
Demo · suppliers, parts, and prices shown are sample data.
Native to ChatGPT
theNasty follows the host's own display patterns and theme — light or dark — so buying a part feels like part of the answer, not an ad wedged into it.
A compact parts list right under the answer, with one clear action.
Compare the same part across suppliers — price, stock, and ship date side by side.
Expand to a focused order review when it's time to pay, then drop back into the thread.
Sample suppliers and prices, for illustration.
How it works
ChatGPT apps are built on the Model Context Protocol. Connect theNasty's MCP server and sign in — the same server and tools also work in Claude, Cursor, and any other MCP client.
They @mention theNasty, or ChatGPT brings it in when a repair question needs parts. No store to find, no app to install.
Fitment-matched parts across every wired-in supplier, with stock, lead time, and anything already on the customer's shelf skipped.
price_quote returns a server-computed cart, checkout hands off to Stripe, and the receipt lands back in the conversation.
Built for commerce
Every result carries a fitment confidence — exact, required, or optional — so the model recommends the right kit, not the closest keyword.
Boats, owned parts, prior orders. ChatGPT knows what's already on the shelf, so it never sells a second case of gear lube.
Prices are always recomputed from the catalog, payment is Stripe-hosted, and orders are confirmed only by a verified webhook.
Make or distribute parts? Drop-ship through theNasty and show up in ChatGPT the next time someone's boat won't start. You get a purchase order the moment the customer pays.