Phase 0 beta is live. Trade perps with AI agents — stay invisible.

Why GitSync

The structural problem with every existing on-chain perpetual venue — and how GitSync solves it.

Every major on-chain perpetuals venue today — Hyperliquid, dYdX, GMX, Synthetix — publishes your positions, orders, and strategies in plaintext. This creates a fundamental asymmetry between you and sophisticated extractors.

The Four Leakage Surfaces

Intent Leakage

Your desire to trade is visible before settlement. Extractors in the mempool see your order and front-run it before it confirms.

Position Leakage

Your open positions are public. Extractors target your liquidation price, hunt your stops, and squeeze funding rates against your direction.

Strategy Leakage

Recurring patterns in your trade history are observable. Profitable strategies get copied or front-run by observers who contributed nothing.

Execution Leakage

Your fill price, slippage, and partial-fill profile are visible. This is the canonical MEV surface — sandwich attacks, back-running, and adverse selection.

The AI Asymmetry

Beyond the informational asymmetry, a second gap has emerged. Sophisticated participants deploy ML models — price forecasters, funding-rate predictors, regime classifiers — that cost orders of magnitude more than what an individual trader can afford. GitSync collapses this asymmetry by making private AI agents a protocol primitive. Every user, from day one, has access to the same class of execution intelligence.

GitSync's answer: Privacy should be the default, not a feature. AI execution should be a first-class settlement primitive, not an off-chain afterthought. And proof obligations — PnL, solvency, fair matching — should be discharged by cryptography, not reputation.

Last updated just now