I built Calibrated because I was tired of pundits making confident predictions with zero accountability. Economists, investors, and politicians say things constantly that turn out to be completely wrong, and nobody keeps score. The idea was simple: match every prediction to a real prediction market (Polymarket or Kalshi), record the market price at the time of the claim, and auto-score it when the market resolves using Brier scoring. No cherry-picking, no manual intervention, no bias. Built the entire pipeline in about 2 weeks using Claude for prediction extraction, pmxt for market data, and Supabase as the backend. The pipeline runs automatically, finds predictions from 70+ tracked public figures, and scores them against crowd consensus. Happy to answer any questions about how the scoring works or what's next for the product!

Calibrated

Who's actually worth listening to?

Who's actually worth listening to?