🔗 https://www.supplement-smarter.com/
SupplementSmarter is a price-tracking + comparison app for supplements (starting with protein and creatine) built as a production-style full-stack project
It combines a product browsing UI with a backend pipeline that scrapes retailer listings, normalises product data, and serves ranked results (value-first) plus price history and all-time low summaries
- 4 retailers (SprintFit, NoWhey, Xplosiv, NZProtein)
- Daily scrape via cron (~03:00 NZT) -> Postgres -> Rebuild read tables
- 24,683 raw listing snapshots in
public.scraped_products - Canonical identity:
(category, brand_normalised, product_normalised, weight_grams)(+ flavour normalisation) - Continues on single-retailer timeout/block; rebuilds from available sources and refreshes fully on next successful scrape
Supplement shopping is messy:
- retailer listings vary in naming/weights/flavours
- “cheap” isn’t always “best value”
- it’s hard to know if a price is actually good without history
SupplementSmarter addresses this by:
- scraping multiple NZ retailers into a consistent format
- normalising + filtering into category-specific tables
- ranking by value score (then price)
- showing offers, price history, and all-time low / current low for a product
- Protein + Creatine category pages, ranked by value
- Search suggestions across categories
- Product pages with:
- current offers (per retailer)
- price history timeline
- all-time low + current low summaries
- Daily scrape + normalisation pipeline
- Automated tests (unit + API + scraper checks)
- Architecture:
docs/architecture.md - Schema (PDF):
docs/schema/schema.pdf - Deploy:
docs/deploy.md - Normalisation & matching:
docs/normalisation_matching.md - Value score:
docs/value_score.md
- Retailers: 4 (SprintFit, NoWhey, Xplosiv, NZProtein)
- Raw listing snapshots: 24,683 rows in
public.scraped_products - In-stock at scrape-time: 94.7%
- Scheduled ingestion: daily cron (~03:00 NZT)
Frontend: Vite + React + TypeScript, Tailwind
Backend: Node.js + Express (TypeScript via tsx), PostgreSQL
Pipeline: Playwright/Cheerio scrapers + SQL normalisation + read-optimised tables
GET /healthGET /api/protein(+/suggest,/:slug)GET /api/creatine(+/suggest,/:slug)GET /api/supplements/suggest?q=...
From server/:
npm run test # run everything
npm run test:watch # dev watch mode
npm run test:unit # unit tests
npm run test:api # API integration tests
npm run test:scrape # scraper tests
npm run test:ci # what CI runs (unit + api)MIT License