AutoHedge

Why we included this project

AutoHedge splits a quant desk's job into four agents, each owning one stage of a trade: forming a market thesis, running technical and statistical analysis, sizing the position, and placing the order. Built on the Swarms framework, the Director, Quant, Risk, and Execution agents pass structured JSON between stages, so you can trace exactly which analysis produced a decision. That transparency, plus the risk-first design where position sizing and validation gate execution, is what lifts it beyond a demo. Trading currently runs on Solana, with Coinbase and other exchanges on the roadmap, so the immediate audience is people already working in crypto. If you want to study how multi-agent pipelines get orchestrated and audited in a live financial workflow, this is a solid starting point.