Multi-Marketplace Opener & Scraper

Checking one product across markets used to mean opening a tab, typing the product ID, switching domains, and repeating — dozens of times a day. I collapsed that entire ritual into a single paste and one click. Then I added an Amazon scraper so the same popup could also pull a whole search page into a structured spreadsheet.

01 · The Problem

Cross-Marketplace Lookups Don't Scale by Hand

Managing global catalogs means constantly answering the same questions: Is this product live in the UAE? How does it look on the US store versus India? Is the matching listing up on Flipkart and Noon? Done manually, each answer is a chain of tedious steps:

• Domain juggling. The same product lives at a different URL on every Amazon marketplace (.com, .ae, .in, .co.uk …) and on Flipkart, Noon, Walmart, Target, and Home Depot. Opening one product everywhere meant typing or pasting into a dozen different search bars.
• Batch validation. Verifying a batch of 50 product IDs was 50 separate navigation actions — with no fast way to open them all and scan side by side.
• Competitive research. Pulling product data off an Amazon search page meant manual copy-paste or screenshots — no clean export, no sponsored-vs-organic split, no sales signal.
• Checkout friction. Testing listings across regions repeatedly needed valid local postal codes that nobody keeps memorized for 14 countries.

02 · What I Built

Three Tools in One Popup

The extension is organized into three tabs, each removing a distinct daily friction point:

03 · How It Works

Under the Hood

The opener is pure URL templating driven by the Chrome tabs API; the scraper injects a content script that reads Amazon's live DOM and serializes it to a workbook on the client. No servers, no data leaves the browser.

Engineering details worth calling out

• Resilient injection. The scraper first messages the page's content script; if it isn't loaded, it programmatically injects content.js via the scripting API and retries — so it works even on freshly opened tabs without a hard refresh.
• Retry-until-rendered extraction. Amazon hydrates results lazily, so the extractor retries up to five times (one-second intervals) until products actually appear, rather than failing on a half-loaded page.
• Sponsored vs organic, the hard way. Amazon hides ad labels behind rotating class names and "Trending now" carousels. The classifier checks multiple signals — feedback labels, component-type containers, and contextual carousel detection — to tag each row correctly.
• Sticky preferences. Every dropdown (comparison type, platform, region) persists to localStorage, so the popup always reopens in the exact configuration you last used.
• Manifest V3, scoped permissions. Host permissions are limited to the specific marketplace domains it touches; the service worker stays idle until invoked.

04 · Impact

Before vs. After

05 · What I Learned

Hard-Won Lessons

• Scraping real-world DOM is a moving target. Amazon's markup changes constantly. Querying by stable structural patterns and layering fallbacks — instead of one brittle selector — is what keeps the scraper alive across redesigns.
• Defaults are a feature. Persisting the last-used marketplace turned a five-interaction popup into a two-interaction one. The smallest UX detail drove the biggest adoption jump.
• Build the most-repeated action first. The single-marketplace opener shipped first because it was the most frequent task; everything else grew around that proven core.
• Keep it local. Doing the Excel generation client-side with SheetJS meant zero backend, zero data-privacy questions, and instant downloads.

Part of my Chrome Extension Productivity Suite — see the companion Brand PDP Group Opener →