Invite only · AI-agent quant trading

Quant trading, run by AI agents. Built for retail.

Autonomous agents research, backtest, and execute systematic strategies on your brokerage account — the kind of trading hedge funds run, sized for a retail account.

Trading involves risk of loss. Past performance does not guarantee future results.

Research
Agents read filings, prices, and news 24/7.
Backtest
Every signal is replayed against decades of data.
Execute
Orders sized and routed inside your risk envelope.
Monitor
Live Sharpe, drawdown, and exposure recomputed daily.
Agent paper account, last 90 days
$2,047
+$84 this month

How it works

Four steps from "connect" to "agents trading on your behalf."

1. Connect your broker

Read-only first. Your money never leaves your brokerage account.

2. Agents propose strategies

Quant strategies matched to your account size, goals, and risk tolerance.

3. Paper-trade for 60 days

Every strategy proves itself on a paper account before touching real capital.

4. Graduate to live, supervised

Live trading with hard daily loss limits and an audit log of every action.

For the curious

The quant, demystified

Four equations behind the strategies the agents run — and what the agent actually does with each one on your account.

The math

Mean reversion (z-score)

z = (price − μ₂₀) / σ₂₀
  • μ₂₀20-day average price
  • σ₂₀20-day standard deviation
What the agent does

The agent scans ~5,000 tickers nightly, flags any with |z| > 2, then checks volume and sector regime before sizing a contrarian entry.

The math

Momentum (12-1 month)

M = (P[t−21] / P[t−252]) − 1
  • P[t−21]price 1 month ago
  • P[t−252]price 12 months ago
What the agent does

The agent ranks the universe weekly by 12-1 momentum, longs the top decile, skips earnings weeks, and rebalances monthly.

The math

Position sizing (fractional Kelly)

f* = (b·p − q) / b → f = 0.25 · f*
  • pestimated win rate
  • q1 − p
  • baverage win / average loss
What the agent does

The agent estimates p and b from your strategy's last 200 trades, sizes each entry at quarter-Kelly, and hard-caps at your daily loss limit.

The math

Risk (annualized Sharpe)

S = ( (R_p − R_f) / σ_p ) · √252
  • R_pportfolio return
  • R_frisk-free rate
  • σ_preturn volatility
What the agent does

The agent recomputes Sharpe daily per strategy. If a live strategy's rolling Sharpe drops below half its backtest, the agent auto-pauses it and notifies you.

You don't need to know the math. The agent runs it. We show it so you can verify what's happening on your account.

Institutional discipline. Retail-sized.

Systematic, backtested, supervised — and conservative by design.

AI agents that don't sleep

Agents continuously scan markets, score setups, and place orders within the risk envelope you set.

Quant strategies, not vibes

Every trade comes from a backtested, statistically-edged strategy with a documented account-size fit.

Retail-first guardrails

Hard daily loss limits, paper-first defaults, and your money stays at your broker. No custody, no leverage tricks.

Quick questions

What is an AI agent here?

An autonomous program that picks a strategy, screens the market, sizes positions, places orders, and monitors risk — all within the limits you set. It is not a chatbot; it is a trader that runs on a schedule.

What strategies do the agents run?

Classic systematic quant: mean reversion, cross-sectional momentum, trend-following, volatility carry, and a few income strategies. Each is backtested, sharpe-scored, and matched to account size.

How is this different from a robo-advisor?

Robo-advisors rebalance passive index funds. We run active, signal-driven strategies — the kind quant funds run — with paper-trade graduation and per-strategy kill switches.

Is my money safe?

Your money stays at your broker. The agents never hold funds. Every strategy has hard-coded loss limits, and you can pause everything in one click.

How do I get access?

Drop your email on the access page. We onboard in small batches to keep agent compute and execution quality high.

Put the agents to work.

We onboard in small batches. Drop your email and we'll send an invite when there's a slot.

Request access