For futures & prop traders

Build bots. Test thousands.
Find real edge.

Describe a setup in plain English → TapeScript builds it into Pine & Python, backtests it with zero lookahead, sweeps thousands of parameters, mutates and filters it, runs the gauntlet that kills curve-fits, and ships it as a monitored bot. The edge is proven before it ever risks a funded account.

No lookahead, no repaintProp-firm rule checksPine · Python export
30+
research tracks per strategy
10
prop firms rule-checked
8
AI models you can run it on
0
lookahead. ever.
See it work
Sound familiar?

You don't have an edge problem.
You have a proof problem.

Your backtest prints money. Your live account doesn't.

Lookahead, repainting and curve-fit make a flat strategy look like a rocket. TapeScript backtests with next-bar fills, real costs and zero peeking — the number you see is the number you'd have traded.

You've blown evals on “edges” that were really noise.

A strategy that won once out of a hundred tries isn't an edge — it's luck wearing a costume. The Deflated-Sharpe and overfitting gates flag it before it costs you another reset fee.

You burn weekends re-running the same research by hand.

Sweeps, sessions, regimes, MAE/MFE, walk-forward, Monte-Carlo — TapeScript runs the whole battery in parallel while you do literally anything else.

You can't tell if last month was skill or variance.

An untouched holdout the optimizer never saw is replayed once, at the end. Pass it and you have evidence. Fail it and you just dodged a funded account.

Backtest you can trust

If it works here,
it works live.

Real history. Next-bar fills. Costs modeled. Zero lookahead, zero repaint. The equity curve you see is the one you'd have actually traded — graded A through F so you know exactly where you stand.

Profit factorSharpe / SortinoR-distributionMax drawdown
Equity · MNQ 1mA
+2.1R expectancy1.9 profit factor−8R max DD
Test thousands · find the plateau

Sweep 10,000 parameter combos before your coffee's cold.

Every stop, target, cap and threshold — backtested in parallel. TapeScript doesn't hand you the single luckiest setting; it finds the stable plateau and rejects the isolated peak that dies the moment the market shifts.

Parameter heatmapsStability clustersWalk-forward
Parameter sweep · 2,000 combosno isolated peak
stop × target × capstable plateau selected
Mutate · filter · evolve

Don't tweak strategies.
Evolve them.

Spin off parameter candidates, feature filters, regime gates, instrument and timeframe variants — each a real, re-validated version. See every branch side by side with its delta vs baseline, and promote the one that earns it.

MutationsFeature engineeringVersion tree
Research tree · every variationvs baseline
v1.0Baseline+0.04R
v1.1Parameterstop 18 / 4R+0.21R
v1.2Feature+ close-inside filter+0.27R
v1.3RegimeNY-open only+0.33R
v1.4MutationMSS confirm+0.18R
We try to kill your edge

2,000 alternate histories.
Does it survive?

Monte-Carlo reshuffles your trades thousands of ways — regime-aware, block-bootstrapped — and reports the odds of ruin. Plus an untouched holdout the optimizer never saw, replayed once. Pass that, and the edge is real.

Monte-CarloUntouched holdoutOverfit gates
Monte-Carlo · 2,000 paths97% survive
block-bootstrap, regime-awareruin risk 1.2%
Your second brain for the markets

Every book, note and chart —
one connected mind.

Drop in trading books, PDFs and your own notes. The Knowledge Brain links them into a living concept graph the AI reads from on every prompt — and writes back to. It gets sharper the more you feed it.

Linked conceptsBacklinksAuto-surfaced context
Knowledge Brain · 142 conceptsfiring
liquidity sweep ↔ ORB ↔ regimedenser every upload
From idea to live bot

Build a bot.
Not a liability.

Promote a validated version to a monitored bot. A Risk Governor checks every single order against hard caps — contracts, daily loss, trailing drawdown, session — with a kill switch always armed. Paper today, your broker next. Never an order-cannon.

Risk GovernorKill switchProp-rule awareTrade journal
Automation · Guarded paper bot● live
1 contract3 trades/day$500 max loss🛑 kill armed
filledlong NQ @ 18,420 · +2R target
rejecteddaily-loss limit — order blocked
filledexit · target hit · +$420
Risk Governor on every orderno order-cannon
Institutional-grade validation, for independent traders

The whole arsenal — in one Command Center.

Plain English → executable

Say the setup out loud. Get clean Pine and Python you can read, trust and ship — not a black box.

Backtests you can actually trust

Equity curve, profit factor, Sharpe, drawdown, R-distribution — on real history, costs modeled, no lookahead, no repaint.

The Gauntlet kills curve-fits

Parameter sweep, walk-forward, thousands of Monte-Carlo paths, untouched holdout, Deflated-Sharpe + PBO overfit gates. We try to kill your edge so the market can't.

Prop-firm rule checks

Topstep, Apex, FundedNext and more: pass probability, trailing-drawdown risk, daily-loss and blow-up odds — before you wire the eval fee.

MAE / MFE truth

See exactly where your stops bleed and your targets leave money on the table — then tighten the geometry to the real excursion data.

Regime & session edge

Find out when your setup actually works — which sessions, which volatility regimes — and gate out the hours that quietly drain the account.

Every variation, side by side

Sweeps, features, mutations, instruments, timeframes — each variant with its own stats and delta vs baseline, in one comparison view.

Run it as a monitored bot

Promote a validated version to a paper bot today (broker connectors next): a Risk Governor checks every order, with hard caps and a kill switch. No order-cannon.

Your AI, your choice

Run the research brain on Claude, GPT, Gemini, Grok, DeepSeek and more — your key, your model, switchable in Settings.

Stop trading on hope. Start trading on proof.

Build your idea, put it through the gauntlet, and know — in plain numbers — whether the edge is real. Your first backtest takes minutes.

Find your edge — free →