The AI-native quant research desk

Describe your edge. QBX proves it.

Turn a plain-language idea into a typed strategy graph, a fee-aware backtest, a portfolio, and a reproducible validation certificate. The product is built around the checks that matter: type safety, no look-ahead, selection awareness, and traceable research artifacts.

Open QBX Book a demo
No look-ahead at compile time Typed node graph Certified validation
QBX strategy engine live
>
typed graph generated from the strategy registry
source kline.close
source kline.volume
series.zscore@1 ret_z3
series.percent_rank@1 vol_rank3
logic.all@1 entry_condition
intent enter_long
valid graph phi = 0 fee 5bps Sharpe 1.52
01 / 04

It starts with a sentence.

QBX turns research intent into a structured graph you can inspect.

QBX copilot

>

compiling against QBX/v0.2 block registry...
US Stock Z-Score Mean Reversion 16 nodes
market_data kline
price.returns@1 ret1
series.zscore@1 ret_z3
series.percent_rank@1 vol_rank3
trigger.true_for_n_bars@1 oversold_confirmed
intent enter_long
PLTR - 1h NAV 1.529
Total Return52.93%
Sharpe1.52
Max DD-14.67%
Trades114
Portfolio Picker
SOXL Bollinger Volume40% weight
PLTR Z-Score Mean Reversion30% weight
SMCI Bollinger Volume20% weight
SOXL SMA Trend10% weight
Matrix & Diagnostics Pairwise correlation across selected runs.
Low High
Score2.816
Avg Corr0.136
Max Pair0.337
Members4
SOXL-BB
PLTR-Z
SMCI-BB
SOXL-SMA
SOXL-BB
1.00
0.08
0.14
0.34
PLTR-Z
0.08
1.00
0.05
0.12
SMCI-BB
0.14
0.05
1.00
0.18
SOXL-SMA
0.34
0.12
0.18
1.00
validation certificate passed annualized 40.36% / max DD -8.96%
What the product enforces

The story is the product.

The hero and workflow above are not generic illustrations. The nodes, strategy names, execution settings, portfolio weights, and validation metrics are pulled from QBX examples and target portfolio captures.

Ready when you are

Bring your edge. QBX will do the proof work.

Open the product app or send a note to schedule a private walkthrough.

Open app.qbxwin.com Email the QBX team