AI vision on every promising listing
Bargain Brief is a personal buyer's agent for households who'd rather buy used. You describe a want in one sentence — e.g. "queen bed frame, solid wood, under $300" — and before anything reaches your inbox, AI vision inspects the listing photos: grain and veneer cues, joinery hints, brand marks, charger/battery presence, obvious damage. Titles lie. Photos tell on them.
Judgment you can act on — or silence
Vision output becomes structured attributes, then a score against your want. Digests stay sparse (0–3 finds) with why it fits, what to check before you buy, and a fair-value band. Empty weeks are intentional — reputation is what we don't send. MVP: furniture & tools on eBay used & refurbished.
Why this is the moat
Deal apps race on refresh speed. We race on recognition — photo-first extraction plus scoring — so the product compounds as a judgment corpus, not a scraper farm. Pipeline: fetch → prefilter → vision-backed extract → score → digest. Live at bargainbrief.app.