How I Built ShopAI — AI Price Comparison Across Every Store (Solo Maker Story)

typescript dev.to

The Problem

Last month I was shopping for a new mechanical keyboard. I spent 3 hours opening 8 different tabs — Amazon, AliExpress, JD, Taobao, Best Buy, eBay, Newegg, Walmart — comparing the exact same model across every store. The price difference was 40%.

I thought: why isn't there a single place where I can search once and see prices from ALL stores?

What I Built

ShopAI — AI-powered price comparison across every major online store. Type what you're looking for, and ShopAI:

  • Scrapes prices from 6+ major e-commerce platforms simultaneously
  • Uses AI to intelligently match products across different stores (same keyboard, different listings with different descriptions)
  • Shows you the cheapest option, the fastest shipping, and the best overall value
  • Works for any product — electronics, books, fashion, home goods

The Tech Stack

  • Next.js + React — Frontend with server-side rendering for SEO
  • AI Product Matching — Custom model for cross-store product deduplication (this was by far the hardest part)
  • Web Scraping Pipeline — Rotating proxies + rate limiting to handle multi-store queries
  • Edge Functions — Price queries resolve in under 2 seconds

Hardest Problems Solved

1. Product Matching is AI-Hard

The same keyboard on Amazon is "Keychron K2 Wireless Mechanical Keyboard (Hot-Swappable)" — on AliExpress it's "Keychron K2 75% BT5.1 RGB Swappable" and on JD it's just "Keychron K2机械键盘". Same product, three completely different descriptions. I had to build a custom matching algorithm using embeddings + fuzzy string matching.

2. Anti-Bot Protection

Every store has different anti-scraping measures. The pipeline handles Cloudflare challenges, CAPTCHAs, rate limiting patterns, and dynamic content loading differently per store.

3. Speed vs Completeness

Users expect results in 2 seconds, but scraping 6+ stores takes time. I built a streaming architecture where results appear as they come in, with the best matches surfacing first.

Results So Far

  • Launched 2 weeks ago
  • 500+ products searched daily
  • Average price savings per search: 22%
  • Users from 20+ countries (mainly US, China, India)

What's Next

  • Adding more stores (Costco, Target, ASOS, Zalando)
  • Price history tracking and drop alerts
  • Mobile app with barcode scanning
  • Browser extension for one-click price comparison

Lessons for Solo Makers

  1. Ship fast, iterate faster — First version was just 2 stores and a basic text search. I shipped it in 3 days.
  2. AI is the differentiator — Without the product matching AI, this would just be another comparison site. The AI makes it feel magical.
  3. Pricing transparency is a superpower — Users love knowing they're actually getting the best deal. The trust builds quickly.

Thanks for reading! I'd love feedback from the dev community — especially on the scraping architecture and the AI matching. What would you do differently?

Try it out: shopai.fun

Follow my solo maker journey: @yjj061014

Source: dev.to

arrow_back Back to Tutorials