DealsOn

DealsOn is an AI Chrome extension that automatically finds matching certified refurbished tech deals on eBay & Back Market while you shop Amazon, saving up to 50%!

Comment

Inspiration

What it does

How we built it

markdown

DealsOn.tech — AI-Powered Refurbished Tech Deal Finder

💡 Inspiration

Every year, millions of consumers overpay for brand-new laptops, smartphones, and gadgets simply because certified refurbished alternatives are scattered across dozens of different marketplaces. At the same time, millions of perfectly good electronics end up as e-waste.

I wanted to solve both problems with a single tool: a smart, zero-friction browser extension that acts as a real-time bargain hunter. By bringing certified refurbished deals directly to consumers right at the moment of purchase, DealsOn helps shoppers keep money in their pockets while promoting circular electronics consumption.

🛠️ How We Built It

DealsOn is built as a lightweight Manifest V3 Chrome Extension powered by Google's Gemini API and modern web standards.

  1. DOM Content Script & Extraction: When a user views a product page on major retailers like Amazon or Best Buy, the content script extracts raw, unstructured title and spec data.
  2. AI Parsing via Gemini API: Raw product titles are often filled with marketing noise. We pass the data to Gemini, which extracts the core hardware attributes—identifying the exact base model, storage capacity, color, and hardware specs.
  3. Real-Time Spec Matching: DealsOn queries partner networks (eBay Refurbished, Back Market, Swappa, and Newegg) using structured parameters to ensure exact model alignment.

4. Side Panel UI: Built with responsive vanilla CSS and HTML5, the extension presents matching certified pre-owned options in a clean side-panel UI with 1-click buy links and official warranty information.

⚠️ Challenges We Faced

  • Noisy E-Commerce Titles: E-commerce titles vary wildly across retailers (e.g., "Apple iPhone 14 Pro 128GB Deep Purple Unlocked Good Condition" vs. "iPhone 14 Pro 128 Purple"). Standard regex pattern matching frequently failed. Integrating Gemini was a game-changer, allowing us to parse noisy data into clean JSON schemas with high accuracy.
  • Affiliate & API Limits: Rate-limiting and affiliate deep-link parameter construction required building dynamic URL encoders and fallback query structures to prevent broken links.

* Manifest V3 Migration: Transitioning to Manifest V3 required re-architecting background service workers and side-panel communications to keep memory footprint minimal.

📚 What We Learned

  • LLMs as Data Normalizers: We learned that Gemini shines not just at conversational tasks, but as an ultra-fast, deterministic data normalizer for web scraping pipelines.
  • Mathematical Savings Curve: We analyzed secondary electronics pricing data and quantified the average consumer savings model: $$\text{Savings Percentage} = \left( 1 - \frac{P_{\text{refurb}}}{P_{\text{new}}} \right) \times 100\%$$ For tech generations between $1$ to $3$ years old (such as iPhone 13 through iPhone 16), our empirical findings show average savings of: $$30\% \le \text{Savings} \le 50\%$$ This data validated that highlighting certified refurbished tech provides substantial financial incentive for users while maintaining product longevity. --- ## 🚀 What's Next for DealsOn
  • Automated Price Drop Alerts: Allowing users to track specific refurbished models and receive notifications when prices drop below a custom threshold.
  • Expanded Retailer Support: Adding support for additional certified refurbished networks globally.
  • Mobile Browser Extension: Porting DealsOn to Safari iOS and Firefox Android for on-the-go deal finding.
Share this project:

Updates