Powered by Hyperliquid

Teach agentsto trade.

Simulate strategies. Train intelligence.

Train the next generation of AI trading agents. Test your strategies in simulated markets and help agents learn from your trades.

How it works

From tradeto training data.

01

Simulate

Trade perpetuals against live Hyperliquid prices with a paper balance. No wallet, no risk: real market conditions.

02

Explain

Tag each trade with a strategy and a one-line rationale. Your reasoning is the signal agents can't get from price alone.

03

Teach

Every trade becomes a labeled example: market context, decision, rationale and outcome. Agents learn from what you did and how it went.

Why Kinfield

Agents learnmarkets from data.They learn judgmentfrom people.

Kinfield turns your simulated trading into training signal: decisions paired with context, intent and outcome, so the next generation of agents trades a little more like the best of us.

What you get

Every tradebecomes a lesson.

Live market data

Mid prices and candles stream straight from Hyperliquid's public API.

Labeled trade dataset

Each round trip is captured with the market state at entry. Export it as JSONL.

Apprentice agent

A baseline agent trains on your trades in the browser and shows how it would act right now.

Strategy tags

Momentum, mean reversion, breakout, news: structure your intent so it can be learned.

Your next tradecould teachan agent.

Launch simulator →

$100,000 in simulated capital. Live markets. Zero risk.