Deep-Grid

DeepGrid is built on the thesis that automated strategies provide superior liquidity and more consistent returns than manual trading. For this hackathon, I laid the foundation for a scalable, non-custodial trading bot platform on Sui.

The initial release delivers a fully functional Grid Bot built on DeepBook V3. It automates buying low and selling high, allowing users to profit from sideways market conditions without constant manual intervention. To ensure safety and isolation, I implemented a dedicated Balance Manager architecture that separates funds per strategy while enforcing trade-only permissions.

This submission represents phase one of a broader vision. I am actively building the underlying infrastructure to support additional strategies such as DCA, arbitrage, and copy trading. This project demonstrates the core system design, security model, and a complete, production-ready grid trading implementation.

I built this project to be scalable for future bot strategies. The frontend uses Next.js 15 and Tailwind CSS for a professional trading interface.

The key technical achievement for the hackathon is the "Smart Entry" system. I wrote custom logic that bundles a market swap and multiple grid orders into a single atomic transaction block. This allows users to deploy a bot with just one asset (like SUI) without doing manual calculations.

I also utilized DeepBook's Balance Managers to create distinct sub-accounts for every bot instance. This foundational work ensures we can easily add new strategy types like DCA or momentum bots after the hackathon.