Scenario: Mechanical Keyboard Under $80

User prompt: "Buy me a mechanical keyboard under $80, brown switches, TKL layout"

Workflow

1. Parse intent: product=mechanical keyboard, switches=brown tactile, layout=TKL, budget=$80
2. Search keyboard vendors and marketplaces
3. Spec-match: hard filter on switch type and tenkeyless layout - specs extracted from listings, not guessed
4. Score price-quality across vendors with review aggregation
5. Curate top 3 spec-true options
6. Approval checkpoint
7. Payment, shipping, tracking

Why generic AI search is insufficient

Generic search returns keyboards that look right but ship with the wrong switches. Firestarter verifies specs from listing data and executes the purchase end to end.

Try this with the API