The problem HedgeOS solves

Most users understand the outcome they want, such as protecting their ETH from a large price drop while staying within a fixed budget, but they do not understand the technical options parameters needed to achieve it. They must choose the correct PUT or CALL, BUY or SELL direction, strike price, expiry, premium, contract size, and live order, which makes on-chain options difficult and risky for non-expert users. HedgeOS solves this by allowing users to describe their risk goal in natural language, then converting that goal into structured financial intent and matching it with suitable live Thetanuts options.

Challenges we ran into

Thetanuts options.

Challenges We Ran Into: The biggest challenge was making AI useful without allowing it to silently make unsafe financial decisions. We had to separate AI interpretation from deterministic verification, preserve the user’s original intent when browsing alternative contracts, correctly handle PUT/CALL and BUY/SELL mappings, support structures such as PHYSICAL_PUT, maintain correct USDC decimal conversions, and prevent portfolio values such as 2 ETH from being misinterpreted. We also had to handle live market changes, revalidate the selected Thetanuts order before authorization, fix false ORDER_CHANGED detections, and design the system so that HedgeOS can prepare and verify the exact transaction while keeping final financial authorization outside the AI boundary.