DriveMatch
Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

Buying a car is overwhelming: there are thousands of models, trims, and options; prices and running costs are hard to compare; owner opinions are scattered across forums and social media, and dealerships hold most of the information advantage. DriveMatch solves this with an agentic, conversational assistant that first gets to know the driver (how they live, drive, spend, and what they value), then turns that into a personal set of selection guidelines used to research and rank real vehicles with live, cited evidence on pricing, reliability, owner sentiment, and common problems. From there, it finds nearby dealers that have the chosen car, estimates how flexible their pricing is, drafts the first message to the dealership with the usual blocking questions already answered, and equips the buyer with tactics to avoid overpaying. ModMarket supplements this by applying the same sentiment-grounded approach to planning modifications for a car the user already owns. It even lets the user identify modifications from images to make it even easier to select the parts they want.