AI Chatbot for Vendure Storefronts

Support That Respects Your Architecture

You chose Vendure because you wanted control. A single, extensible commerce core exposes one GraphQL API, and you build whatever storefront your brand demands on top of it — Next.js, Nuxt, SvelteKit, Remix, Astro, or something entirely your own. The last thing a setup like that needs is a support tool that assumes a monolithic theme system and tries to inject itself into a backend it was never meant to touch. The Asyntai assistant respects the headless boundary: it lives in the frontend layer where your UI already runs, and it leaves your Vendure server, your Shop API, and your admin dashboard untouched.

That separation is the whole point of headless commerce, and it is why adding an assistant is so undramatic. You drop one script into the root layout of your storefront app, ship it through your normal CI pipeline, and the widget renders on every route alongside your own components. There is no plugin to register with the Nest-based core, no resolver to write, and no schema migration to run. The assistant is just another client-side script, sitting next to your analytics and your tag manager.

It Reads the Storefront You Already Shipped

In a headless world your product data lives in Vendure, but your shoppers experience it through the pages your frontend renders. Asyntai meets them there. Pointed at your live storefront URL, it crawls the rendered product detail pages, collection pages, and content routes exactly as a customer would see them — variant options, descriptions, pricing context, policy copy, and all. You do not have to expose a custom feed, replicate your catalog, or build a sync job against the Shop API. If it is on the page, the assistant can learn it.

For knowledge that never makes it onto a public route — internal sizing matrices, fabric or component specs, B2B pricing rules, fulfillment SLAs — you upload documents directly. Combine that with custom instructions and the assistant becomes a precise extension of your storefront rather than a generic bot. It answers from your material, in your voice, and defers honestly when something falls outside what it has been given.

Built for Teams, Not Just Solo Merchants

Vendure projects are usually run by developers, agencies, or in-house engineering teams, and that changes what good support tooling looks like. It needs to fit into version control, survive a redeploy, and never become the thing that blocks a release. Because the integration is a single tag in a layout file, it does exactly that: it is reviewable in a pull request, it travels with your codebase, and it carries zero coupling to your business logic. Swapping it, gating it to certain routes, or pulling it for a staging environment is a one-line change.

The same model scales cleanly when one Vendure backend powers several frontends or channels. Each storefront can run its own script with its own knowledge base and its own chat history, so a B2B portal and a consumer site backed by the same core can answer in completely different registers. You manage all of it from a single dashboard without ever redeploying the commerce core.

More Conversions Without More Headcount

A custom storefront earns you a fast, tailored shopping experience, but it does not magically answer the questions that stall a purchase. Does this ship to my region? Is the larger variant in stock? What is the return window? When those questions go unanswered at the moment of hesitation, carts get abandoned regardless of how polished the UI is. An always-on assistant closes that gap, resolving doubt in the same instant it appears and keeping shoppers moving toward checkout.

It also pays back in insight. The chat logs surface what real visitors keep asking — gaps in a product description, a missing spec, a confusing policy — and feed it straight back into your storefront content and Vendure catalog. Over time the assistant does double duty: it handles the front line of support 24/7, and it tells you precisely where to improve the experience next, all without adding a single person to your support rota.