About · 2026

Built to organize the uncertainty markets cannot price.

The questions that decide returns, macro regime shifts, AI labor displacement, fiscal stress, do not have a clean market price. Crene is a structured uncertainty system that decomposes them into falsifiable components, scores them daily across a four-model frontier ensemble, and discloses its calibration openly. The data is the moat.
The mission

Markets price what is liquid and near-dated. Everything structural is left to narrative. Crene exists to turn that narrative into structure: an anchor question, the component conditions underneath it, an explicit horizon, and a daily record of how the view moves and where the models disagree.

The aim is not a trade trigger. It is an honest, auditable input to institutional judgment on the questions that matter most and resolve furthest out.

Founder
Stephen Lee
Founder & CEO

Stephen spent seven years in institutional finance at Goldman Sachs and Credit Agricole before building Crene. Self-taught full-stack, he ships daily across the entire system, from the crawler to the four-model scoring engine to the frontend.

He built Crene because the questions that move institutional portfolios are exactly the ones with no clean market price, and no one was organizing them with falsifiable structure and disclosed calibration. Based in San Francisco and Seoul.

Stephen Lee, founder of Crene
How we hold ourselves accountable

We disclose calibration in tiers, including the misses. The external ensemble Brier is published, the macro product record is shown with its full resolved question set, and nothing is claimed that the record does not establish.

About Crene | Founder, mission, and approach