The world is entering a new era.
So is the way you trade.
Backtest your idea,
not your gut.
Describe a trading strategy in plain English. We turn it into Python, run a real vectorbt backtest with Monte Carlo, stress-test it for robustness, and export to TradingView Pine Script. Try it now, free, no card.
Your idea, put to the market's test.
We looked for the setting that ruins you.
We found it.
Example
A product that sells you a curve will never show you that one.
You are not alone facing this.
Your idea is not the problem — your 3 % stop is. It takes you out on noise: of the 25 variants, every one that dies runs a tight stop. Worst case goes down to −44 %.
“Widen it then, and size down.”
Done. Stop at 5 %, position size at 50 %. I am re-running the sweep over the same 25 combinations — same protocol, otherwise the comparison is worthless.
Fair warning: you will make less. That is what robustness costs, and I would rather tell you now than let you find out.
The strategy · live
Built to last,
not to impress.
−8.7 points of return. A strategy that holds up has a price. We would rather show you the bill than sell you the opposite.
When we don't know,
we say nothing.
A dash, where others would have written 0% in green.
The strategy has not taken a single position yet. So there is no real drawdown to compare against the backtest's. Showing “18 points of margin” would be arithmetically exact, and factually false.
This is a bug we had, that we found, and that we fixed.
You write a sentence.
You walk away with code.
« RSI oversold sur AAPL avec 3 % stop loss »
def calculate_signals(df): rsi_14 = ta.momentum.RSIIndicator( df['Close'], window=14).rsi() entries = (rsi_14 < 30) exits = (rsi_14 > 70) return {'entries': entries, 'exits': exits}
Ask the question that stings.
Is this another AI hype tool that gives me a fake equity curve?
No. We use vectorbt to run real backtests on real 5-year OHLCV data (yfinance), with slippage and commission modeled. Monte Carlo bootstrap gives you a P(profit) confidence band so you can see if the win was luck or structure. Every strategy gets a ROBUST or FRAGILE verdict via parameter sweeps to flag overfitting.
Do I need to know Python to use it?
Not at all. You describe your idea in plain language (EN or FR) and we generate the Python for you. If you want to read or edit the code, it's downloadable. If you just want to trade, we export to TradingView Pine Script in one click.
What about my data and my strategies, are they private?
Your saved strategies, chat history, watchlist, and personal documents are all isolated per account, another user cannot see them. Public reference data (SEC filings, market prices) is shared across all users for cost reasons. We're in Pre-Alpha, we're not selling your data.
When will paid tiers launch?
When we have ~20+ active users on the free tier and demand for Pro features (unlimited strategies, Copilot, Robustness Test). Join the waitlist on the pricing page, with no commitment.
Can I cancel? Will I get charged?
Nothing to cancel · Pre-Alpha is fully free, no card required. When paid tiers launch, you automatically keep the free plan (3 strategies/month). You only move to Pro, the first paid tier, if you want more.
Will my strategies work in live trading?
Exported Pine Script runs identically on TradingView. That said, no backtest is a perfect predictor of live results, slippage in live markets is harder to model, and past data doesn't guarantee the future. Always paper-trade a new strategy before risking capital.
Bringyours.
10 days unlimited, no card · one minute to sign up