- All parsing happens client-side. Your order data never hits our servers unless you explicitly share. You can upload your Amazon CSV export or use a bookmarklet that extracts from the order history page.
- Stack: React + Vite on Cloudflare Pages, Hono.js workers on Cloudflare Workers, D1 for the few things we do store (shared links, anonymous benchmarks).
- AI is used in two places: Claude Haiku for product categorization and personality generation, OpenAI with web search for the savings report (finds cheaper alternatives to your repeat purchases).
- Monetization: Free tier shows basic stats. $5 unlocks AI categorization, personality analysis, shareable cards, and the savings report.
The bookmarklet was the hardest part - Amazon's DOM structure varies wildly between desktop/mobile, account age, and regions. Currently on v22.
Happy to answer questions about the architecture or share more about specific implementation details.
Hey HN! Built this over a weekend because Amazon doesn't give you any year-end insights into your shopping habits. Mostly inspired by this: https://x.com/anshnanda/status/1994641538354393312 :P
Key technical decisions:
- All parsing happens client-side. Your order data never hits our servers unless you explicitly share. You can upload your Amazon CSV export or use a bookmarklet that extracts from the order history page.
- Stack: React + Vite on Cloudflare Pages, Hono.js workers on Cloudflare Workers, D1 for the few things we do store (shared links, anonymous benchmarks).
- AI is used in two places: Claude Haiku for product categorization and personality generation, OpenAI with web search for the savings report (finds cheaper alternatives to your repeat purchases).
- Monetization: Free tier shows basic stats. $5 unlocks AI categorization, personality analysis, shareable cards, and the savings report.
The bookmarklet was the hardest part - Amazon's DOM structure varies wildly between desktop/mobile, account age, and regions. Currently on v22.
Happy to answer questions about the architecture or share more about specific implementation details.