CarCast

⭐ AI used-car price forecasts. See 60 days ahead.

Cars

SaaS

Artificial Intelligence

Visit WebsiteSee on Product HuntTwitter ⧉

Hi PH 👋 — I'm Rick, building CarCast. Quick context: I spent the last year watching car prices swing wildly — used-truck prices up 12% in three months, then a hard correction. Every dealer I talked to was making big inventory bets using yesterday's data (Manheim Market Report, KBB) — backward-looking averages that tell you what cars sold for last week, not where prices are heading. CarCast is the opposite. We feed 52+ weeks of dealer-listing data, the Manheim Used Vehicle Value Index, and the BLS used-car CPI into a proprietary AI forecasting stack and produce probabilistic 60-day forecasts for 300+ vehicle segments. Each forecast comes with P10/P50/P90 confidence bands and a Rising / Stable / Softening trend classification. What it's good for: • Independent used-car dealers timing auction buys • Wholesalers spotting segment-level shifts before the market does • Retail buyers deciding whether to buy now or wait • Fleet teams forecasting end-of-cycle residuals Stack note for the curious: 200M-parameter foundation model for time-series, FastAPI + Postgres backend on Fly, Next.js front-end on Vercel, Modal GPU inference. Forecasts re-run every Sunday morning. Free plan covers 5 segments and 30-day forecasts — enough to evaluate. Pro ($149/mo) unlocks all 300+ segments and the 60-day horizon. Agency ($499/mo) adds API access for dealers integrating into a DMS. Happy to answer anything — what's broken in your current pricing tools, what segments you wish we covered, where the model is wrong. Roast me, I can take it. — Rick