Maharsh Khatri

Cortex - Hedge funds for Polymarket, Kalshi, and Hyperliquid

by
Describe a strategy in English. Cortex compiles it to a typed DSL, backtests it on 30 days of real Hyperliquid data, and runs it as an autonomous agent with an 11-rule risk engine on every order. Paper-settled against live prices.

Add a comment

Replies

Best
Maharsh Khatri
Hey Product Hunt — I'm Maharsh, one of the two people behind Cortex. This came out of a frustration I couldn't shake: I had trading ideas but no honest way to test them. Today your options are to code a strategy yourself in Pine Script or Python, or hand your money to an opaque copy-trading bot and hope. There's no middle ground where a normal person can describe an idea in plain words, see how it actually would have performed on real data, and watch it run with real risk controls. So we built that middle ground. You describe a strategy in English — say, "short SOL when funding flips negative, trail a 1% stop, move to breakeven after it's 1% in profit." Cortex compiles it into a typed strategy you can read rule-by-rule, backtests it on real Hyperliquid data in seconds, and deploys it as an autonomous agent that runs 24/7 with an 11-rule risk engine checking every order. It's all paper-settled against live prices, and every agent posts to a public leaderboard — so performance is transparent instead of screenshot-based. The build was its own story. My original co-founder stepped away early and I shipped the first MVP solo. Then Nikul joined — he's 17 and deferring CMU — and rewrote our backtesting core as a 1M-event/second order-book simulator in Rust, which is why the backtests are both fast and honest. We're early, and we'd genuinely love your feedback — especially on the strategy compiler and which markets you'd want us to support next. Type in a strategy and tell us where it breaks.