PL EN
// Case study · mobile app · AI

ZenKoszyk: an AI shopping assistant ecosystem

ZenKoszyk turns ordinary receipts into a tool for saving money and shopping healthier. We built the whole ecosystem: an iOS and Android app, a website with a help centre, a community forum and a backend that reads store flyers every day, matches deals to what you buy and shows where to buy cheaper.

zenkoszyk.pl ↗
2app stores: App Store and Google Play
5retail flyer parsers
100personalised offers per user
15 mincycle for checking new flyers
// What we did

App, website, forum and an AI backend

One team designed and built every part — from the phone screens to the server that reads flyers and receipts.

// 01

iOS and Android app

Receipt scanning from the camera, gallery or PDF, purchase history, spending statistics by category and day, a barcode and label scanner. Subscriptions via the App Store and Google Play, sign-in with Google and Apple.

// 02

AI receipt recognition

A two-step process: the AI model first transcribes the receipt exactly, then turns abbreviated till names into full product names and adds a category, photo and description. Keeping the steps apart stops it inventing products.

// 03

Flyer parsing

Every 15 minutes the backend checks retail flyers — including ALDI, Kaufland, Carrefour and Dino — extracts the deals, keeps price history and deactivates expired offers. The administrator adds new stores in the panel.

// 04

Personalised deals

The recommendation system compares purchase history with the deals database and picks up to 100 offers per user — taking into account favourite stores, diet (e.g. vegan, gluten-free, lactose-free) and excluded products.

// 05

“Where is it cheaper?”

For every product on a receipt, the app looks for the same item in other stores — on Allegro and Google Shopping — and shows where it costs less.

// 06

An AI agent that knows your shopping

A chat that analyses the user’s receipts and answers questions about their shopping: how much they spent, on what, what to buy cheaper and how to change the basket. It follows the chosen goal: saving money, fewer calories, less sugar or healthier food.

// Website and forum

zenkoszyk.pl and the community

The product website with app downloads, a forum for conscious consumers and a help centre.

// App

ZenKoszyk on the phone

App Store screenshots: receipt scanning, Zen goals, price comparison, spending by category, diet-matched deals and the AI assistant.

// Backend

A server that works for the user

A Python API (FastAPI) with PostgreSQL and vector search, a background job queue and a React admin panel.

  1. 01Admin panel: users, payments, plans and receipt limits, AI model keys, email templates and statistics.
  2. 02Stores and flyers managed in the panel: manual parsing runs, logs and the list of deals.
  3. 03Payments: App Store and Google Play in the app, Stripe on the web.
  4. 04Vector search (pgvector) links products from receipts with deals from flyers.
  5. 05Background jobs: receipt recognition, flyer parsing and product data enrichment.

Want your own AI-powered app? Websites and apps →

// Let’s talk

Got an idea for an AI app?

Free online consultation. We will help you plan the app, backend and AI agent — and tell you what it costs.

info@axisway.com+48 791 793 584ul. Szewska 8, 50-122 Wrocław