AutoRadar
Who it's for
Used car buyers in Serbia
The pain
Searching across multiple car listing sites is tedious. Prices lack context, listings go stale, and there's no easy way to compare deals.
The solution
AutoRadar aggregates listings from major used car platforms, enriches them with price analysis and market context using AI, and surfaces the best deals in a single dashboard.
The Problem
Buying a used car in Serbia means checking 3–4 different websites daily, mentally tracking prices, and trying to figure out if a listing is a good deal or overpriced. There’s no centralized way to compare.
What I’m Building
AutoRadar is a smart aggregator that:
- Scrapes listings from major Serbian car marketplaces on a schedule
- Normalizes data into a consistent format (make, model, year, mileage, price, location)
- Analyzes pricing using historical data and AI to score each listing
- Alerts buyers when a listing matches their criteria and is priced well
Technical Approach
The backend is a Java/Spring service with scheduled scraping jobs and Spring AI for the intelligence layer. Listings are stored in MongoDB. The frontend is a React SPA with Tailwind, focused on fast filtering and clear data presentation.
Current Status
The scraping pipeline is running and collecting data. Basic search works. Next up is the AI scoring system and user alerts.