Skip to content
← All projects

16 · In-house · App development · AI integration

Shopping app with AI assistant – built with Replit

A multilingual progressive web app for a WooCommerce shop: with an AI shop assistant, push notifications for offers and order status, and checkout via WooCommerce. Built with the AI development tool Replit around ten times faster than the conventional way, by our own estimate.

  • In-house
  • PWA
  • React
  • TypeScript
  • Tailwind CSS
  • Express
  • Replit
  • OpenRouter
  • DeepSeek V3
  • Firebase Cloud Messaging
  • WooCommerce REST API
  • WordPress REST API
  • i18next
≈ 10×
faster to build than the conventional way (our estimate)
2 languages
German and English, including the AI assistant
iOS & Android
installable as a PWA, with push notifications

01 · Problem

An online shop should also be available as an app on the smartphone – with search, wishlist, cart and checkout, with push notifications for special offers and order status, and with an AI assistant that answers questions like “I am looking for product xy in size S between 50 and 100 euros” or “How much is shipping?”. The existing WooCommerce shop was to remain the data source, without maintaining products twice. And we wanted to find out how far an AI development tool like Replit goes on a real project of this size.

02 · Solution

A progressive web app with React, TypeScript and Tailwind CSS, developed with Replit. Products, categories, payment methods and blog posts come via the WooCommerce and WordPress REST APIs; a dedicated server with Express and Node.js keeps the credentials secret and calculates tax and shipping live via WooCommerce. The AI assistant – a chatbot right in the app – runs via OpenRouter with the language model DeepSeek V3: it recognises product questions, searches the shop live by category, price and offers and shows the results as clickable product cards in the chat. A webhook from WooCommerce triggers push notifications via Firebase Cloud Messaging as soon as a price drops or an order status changes. On top come fault-tolerant search, filters, wishlist, recently viewed products, reviews, light and dark mode and German and English; the fonts are hosted on our own server instead of an external font service.

03 · Result

An app that can be installed on iOS and Android and pulls all content directly from the shop – with checkout and payment methods from WooCommerce, order overview and confirmation email. Replit implemented the interfaces to WooCommerce, Firebase and OpenRouter correctly at the first attempt and kept checking its own code; by our estimate development was around ten times faster than the conventional way. We also saw the limits: an outage of the platform lasting several hours, in which no data was lost, and individual tasks for which Replit took longer than a manual solution. All in all: recommended – also as a tool for prototypes that are to be tested with clients quickly.

04 · Insights

  • App home screen with product search and filters
  • Home screen in dark mode
  • AI shop assistant in the chat
  • The assistant shows current offers as product cards
  • Answer to a question about shipping costs (English)
  • Search with filters for category, price and offers
  • Product detail with sale price
  • Shopping cart
  • Order summary at checkout
  • Order completed
  • Push notification for a special offer
  • Push notification for a changed order status
  • Switching between German and English
  • WordPress blog inside the app

Next project

Your project here?

Tell us about your problem. The first conversation is of course free of charge and without obligation.

Request a project →

Newsletter

News from the vona workshop.

AI tools we actually use, our weekly AI review & insights from our projects – short, practical and without spam.