REDPATH

RedPath - Best Price Finder v3.6

One search, every store. RedPath is a Chrome extension that sweeps 27 retailers across the web in parallel, reads prices straight out of their search results, and ranks every offer cheapest-first - with the single lowest price spotlighted at the top.

How it finds prices - the extraction engine

scraper.js runs a tiered extraction waterfall. Every store runs through up to four independent strategies, so a store only fails if all four miss:

  1. Site parser: eBay, Amazon, Walmart, Best Buy, and Newegg have parsers tuned to their exact markup (the fast path).
  2. Structured data: schema.org Product/ItemList data as JSON-LD and microdata, the machine-readable layer most stores publish for Google.
  3. Embedded state mining: modern storefronts are React/Next.js apps that ship their search results as JSON inside the page (__NEXT_DATA__, Redux state…). The engine deep-walks every JSON blob hunting for product-shaped objects by key signature (name/price/url aliases, nested price objects).
  4. Automatic record detection: find repeated sibling card structures in the raw DOM, pick the block that best matches a product grid, then pull title/price/link/image from each card (skipping strike-through “was” prices).

If everything misses or the store blocks robots, RedPath still gives you a one-click manual search link. Mercari, Google Shopping, and Bing Shopping are manual links from the start (fully client-rendered).

Scraping public pages is best-effort - retailers can and do block it. The reliable alternative is a store’s official API, and RedPath supports the eBay Browse API out of the box:

Once connected, eBay searches go through the official API - no scraping, no blocking, cleaner data - and results are ranked by the same Specific/General engine as everything else. Your key is stored only on your device (chrome.storage.local) and is never part of the shipped code or sent anywhere but eBay. Remove it anytime to fall back to public-page reading.

This is the model for adding more official sources later (Best Buy and Walmart both offer free keys); each store adapter is small and self-contained.

Match modes:

Condition filter (Any / New / Refurbished / Used): each listing is classified by Scraper.detectCondition: from eBay’s authoritative API condition field when available, otherwise from title keywords (“Renewed”, “Refurbished”, “Open Box”, “Pre-Owned”, “Used”, …). Anything with no signal defaults to New (most retail listings). Switching the filter re-ranks the current results instantly without re-searching.

Match filtering - the keyword + category engine. After each sweep RedPath decides which listings are genuinely the product using a local relevance engine (scraper.js), with no AI, no API key, and no server: it runs entirely in the popup and works identically on Chrome, Brave, Edge, and Firefox. Beyond plain token matching it:

It is conservative: a real product with a long, spec-stuffed title is always kept, and Specific mode keeps the accuracy high by requiring every query word to appear.

Ranking defenses (so the “best deal” is actually the product):

Install (Load Unpacked)

  1. Open Chrome and go to chrome://extensions
  2. Toggle Developer mode ON (top-right corner)
  3. Click Load unpacked and select this folder (Red Path)
    • already installed? Hit the ↻ reload icon on the RedPath card instead
  4. Chrome will note the extension can “read data on all websites” - that’s the permission that lets it check prices on any store across the web
  5. Pin RedPath from the puzzle-piece menu

Important: live price scanning only works when RedPath runs as an installed extension. Opened as a normal web page, the browser’s security rules block it from reaching other sites - the page will tell you when that’s the case.

How to use

Security & privacy posture (Chrome Web Store readiness)

Hardening applied for publication - this is also the honest list of what the extension does and doesn’t do:

What no extension can promise: client-side code is always readable by the person who installs it (CWS forbids obfuscation), and searches necessarily reach retailers from the user’s own browser/IP. “Impossible to hack” doesn’t exist in software - the strategy above is minimizing what an attacker could ever get: there are no secrets in the extension, no user data beyond one locally-stored search term, and no privileged surface beyond reading retailer search pages.

RedPath can tag every outbound store link with your affiliate IDs so you earn a commission when someone buys - at no extra cost to them. As the developer, you set the IDs once in the CONFIG block at the top of affiliates.js; the same IDs apply to every user (they are public tracking IDs, not secrets, and never change the price a shopper pays):

CONFIG field Get it from Covers
ebayCampaignId partnernetwork.ebay.com (free) eBay
skimlinksId sovrn.com/commerce ~48,000 merchants (Walmart, Target, Best Buy, …) - one ID
amazonTag affiliate-program.amazon.com Amazon (see below)
templates (JSON) Impact / CJ / Rakuten dashboards any store, per-store {url} template

For each link RedPath uses, in order: your custom template for that store -> a native tag (Amazon tag, eBay EPN params) -> a universal Skimlinks wrap -> otherwise the link is left untouched. Blank fields simply don’t earn; nothing breaks.

You must comply with two things (both handled/flagged):

Realistically, affiliate income needs real user volume - commissions are a few percent per sale - so treat this as “monetize once you have users,” not instant profit.

Performance

Notes & limitations

Files

UI stack (npm)

Package Used for
@fontsource/inter Body text, headings, buttons
@fontsource/jetbrains-mono Prices and ranks
@fontsource/playfair-display Display serif accents
animejs Subtle entrance animations

The REDPATH wordmark is an image (src/assets/wm_title.png), not live text, so no display font is bundled for it.

src/vendor/ is pre-built and checked in, so Load Unpacked works with zero build steps. Only if you update packages: npm install then npm run vendor.

Licence

Apache License 2.0 - see legal/LICENSE.

Author

Karen Contreras Cubas - karenc.cubas@gmail.com

Privacy policy: legal/privacy-policy.md.