Search APIs hand your LLM someone else's index: noisy snippets, wasted tokens, results that almost fit. Rented context. In this hands-on session you build context you own: a search engine over exactly the slice of the web you care about. Pick an e-commerce site, then vibe code the full pipeline: scrape its product pages, index them in a vector database, and query your new engine from an LLM. The payoff: context that is exactly what the model needs, far fewer tokens per answer, and faster responses. You leave with a working search engine over the store you picked, a blueprint that applies to any site you point it at, and a head start on the shift from renting context to owning it.
- A laptop
- An AI coding assistant you already use (Claude Code, Cursor, or similar), installed and signed in
- A Bright Data account, created before the session (link) 
- A niche in mind: the e-commerce store or product category you want your search engine to cover
- No coding, scraping, or vector database experience needed