AI · multi-agent In development

LLM Trader

What happens when local AI models argue over a trade?

LLM Trader trading console dashboard with wallet balances, auto-trade engine status, token universe, and recent trades.

About this project

I built LLM Trader to explore how local language models make decisions about a market. They can suggest SOL/USDC trades, but a separate set of rules checks the price, size and risk before anything can happen.

Most of the work happens in paper mode. I can follow a proposal from the first suggestion to the final decision, see where the models disagree, and change the experiment without moving real money.

Screenshots

Risk and policy page showing operational halt, slippage caps, snapshot age limits, preflight rulebook, and RPC provider redundancy.
System health page with services status, auto-trade engine state, and per-provider health for Jupiter and RPC endpoints.
Wallet page with hot wallet balances, browser wallet connection, fund and withdraw flows.
Positions and PnL page listing open trades and realized profit / loss.