Multi-agent AI · Live StockX data
DRIP
An AI sneaker stylist that reasons about fit and taste instead of just returning search results, with live resale prices per shoe.
The idea
DRIP builds outfits around your sneaker collection, like Nike SNKRS meeting a personal stylist. A multi-agent LLM does the styling: a Stylist agent drafts a look and a Critic agent refines it, so the advice is honest and weather- and material-aware. It will tell you not to wear suede in the rain rather than hype every pick.
Real data underneath
Live StockX resale prices, buy links and product images stream in per shoe through KicksDB, and search hits the entire real catalog rather than a static list. A camera scanner extracts dominant colours with k-means and auto-captures when you hold a shoe steady, then an AI writes the fit verdict.
Under the surface it runs a content-based recommendation engine with taste vectors and MMR diversity, a weather-aware outfit engine grounded in colour theory, a community feed, a saved-fit lookbook, a voice stylist and a multi-currency shop. The core browsing and deterministic styling flow works key-free; provider keys enable live AI and market data. It installs as a mobile-first PWA.
Highlights
- Stylist and CriticOne agent drafts the look, a second refines it into honest, wearable advice.
- Live resale pricesStockX prices, buy links and images stream per shoe through KicksDB.
- Camera colour scank-means extracts a shoe's colours and auto-captures, then the AI writes a verdict.
- Key-free coreCore browsing and deterministic styling work without keys; provider keys enable live AI and market data.