Pricing intelligence for U.S. importers — enter product costs in plain English and get landed cost, MSRP, margins, and retailer-specific buyer analysis in seconds.
You describe your product in plain text: manufacturing cost, shipping, origin country, quantity, and target retailer. Groq's LLM extracts structured product data from that input, then a local HTS category map resolves the applicable tariff rate and any country-specific surcharges. A second Groq call uses those inputs to calculate landed cost, derive wholesale and MSRP from target margins, and generate buyer intelligence tailored to the specific retailer. Everything runs server-side in a Vercel serverless function — the API key never touches the browser.
- React 19 + TypeScript + Vite
- Groq (
llama-3.3-70b-versatile) for extraction and pricing analysis - Vercel serverless functions (
api/analyze.ts) - Vitest + Testing Library for unit and integration tests
git clone https://github.com/alcj3/GlobalQuote.git
cd GlobalQuote
npm installCreate .env.local in the project root:
GROQ_API_KEY=your_key_here
Then start the dev server:
npm run devThe app runs at http://localhost:5173. The /api/analyze endpoint is served by Vite's proxy in development.
npm run test71 tests across 7 files covering extraction parsing, tariff lookup, pricing analysis, retailer margin logic, and the HTTP handler.
- Tariff rates reflect the HTS schedule as of early 2025. Rates introduced or modified after April 2025 — including additional Section 301 tranches and reciprocal tariff actions — are not reflected.
- Country surcharge coverage is limited to the major sourcing origins (China, Vietnam). Other countries fall back to the MFN base rate with no surcharge applied.
- The HTS category map covers broad product categories, not 10-digit HTS subheadings. Rates are directionally accurate for estimation but should not be used for customs filings.