ClearCart
Helps you figure out if an Amazon product is actually worth buying. It reads the reviews with AI, scores the product out of 100, shows the price history, and ends with a clear verdict: Recommended, Solid, Caution, or Avoid.
What it does
- Reads hundreds of reviews and pulls out the themes and red flags so you don't have to scroll through them yourself.
- Scores trust from review authenticity, sentiment, product quality, and trend.
- Shows price history so you can tell a real deal from a fake one.
- Price alerts that email you when a product drops below your target.
Where it runs
- A Chrome extension that works right on the Amazon product page.
- An iOS app: share an Amazon link from Safari or the Amazon app and it analyzes it on the spot.
- clearcart.app on the web.
How it's built
It's a monorepo with three packages: the Chrome extension, a Next.js web app and API, and a shared TypeScript package. Products are scraped client-side, by the extension on the page or by a headless WKWebView in the iOS app. Price history comes from Keepa, and an LLM writes the analysis with schema-validated output. The final trust score comes from four weighted scorers, and when the model's verdict disagrees with the math, the math wins. The iOS app is SwiftUI with zero third-party dependencies.