Find out if your strategy really works.

Kwaant is an AI quantitative researcher for retail traders. Describe the strategy you already trade. It tests it honestly on 16+ years of historical CME Group data, searches for real improvements, throws out the ones that are only luck, and shows you all the evidence.

See how it works

Takes about a minute. No payment needed to apply. What's included in early access?

The AI never computes a numberEvery figure comes from code.

Every variation is tested for luckThe more you try, the higher the bar.

A sealed final test, opened onceOn history nobody has looked at.

The real question

You trade it every day. Do you know if it actually works?

Most strategies run on feel. A good month feels like proof. A bad week feels like the end. Neither tells you what's real, and finding out the hard way costs real money.

Be honest. Which of these sound like you?Tap any that apply

Tap the ones that sound familiar.

See how it works

What's included in early access?

How it works

From your words to honest evidence.

01

Describe it

Type your strategy the way you'd explain it to a friend. Kwaant turns it into exact rules. When a term like "fair value gap" can mean different things, it asks you, with a small chart of each version. Then it shows you example trades so you can confirm it understood.

Buy the first return to a fair value gap after the London open. Stop under the gap.
Which one do you mean?
Gap between wicks
Gap between bodies
02

It's tested honestly

Your strategy runs on 16+ years of historical CME Group data, with commission, slippage and margin included. The AI proposes ideas worth trying. Each is tested across hundreds or thousands of variations, checked on data the search never saw, and run through statistical tests built to catch luck. Survivors face a sealed final test that can be opened only once.

03

You get the evidence

A clear verdict, everything that was tried, what was thrown out as luck and why. Every number traces back to the test that produced it.

Improved
Everything tried
Thrown out as luck
Survived every check
Where it loses
04

Take it further with Pine Script

Export any strategy Kwaant has tested as Pine Script, and automate it yourself if you choose. Kwaant writes the code; you stay in control of every trade.

//@version=6 strategy("London gap retest", overlay = true) // rules exported from Kwaant inGap = low <= gapTop and close >= gapBottom if inLondon and inGap and firstTouch strategy.entry("Long", strategy.long)
Export as Pine Script

Quant trading, minus the quant degree

Research like a quant. No code. No statistics degree.

Quant traders test every idea with clean data, strict statistics and a sealed test they never peek at. Doing that yourself usually means months of learning and building. Kwaant does the heavy lifting. You bring the idea.

The usual wayWith Kwaant
  • Learn to code
  • Source years of clean data
  • Build a backtester that's fair
  • Learn the statistics of luck
  • Keep a test you never peek at
One step

Describe your strategy in plain English.

  • 16+ years of CME Group data
  • Luck tests built in
  • A sealed final test, opened once
  • Pine Script when you're done
Five hard stepsor one sentence.

The tests

Eight tests. One honest answer.

The AI decides what to test. Deterministic code decides what is true. Every number below comes from code that gives the same answer every time, never from the AI.

What your strategy has to survive

Out-of-sample test

Built and tuned on one stretch of history, then judged on a later stretch it has never seen.

Walk-forward test

Tested again and again on later windows, like trading it forward through time. Does the edge last?

Parameter stability

Every setting is nudged up and down: a 20-bar average becomes 18 or 22. A real edge survives small changes.

Cost stress

The whole backtest re-run with trading costs at 1.25×, 1.5× and 2×. Does the profit survive worse fills?

Monte Carlo simulation

Trades shuffled into thousands of alternative histories, to show how bad a drawdown could realistically get.

Market conditions

Results split by volatility, trend and time of day. Does it only work in one kind of market?

Overfitting checks

The more variations tried, the higher the bar: the deflated Sharpe ratio and the probability of backtest overfitting.

The sealed final test

A final slice of history, locked away and opened once for one final version. Software opens it, never the AI.

Built-in honesty: the strategy language can't see the next bar, and known defects are put back into the engine on every check to prove the tests still catch them.

What you get back

Returns

  • Net profit
  • Total return
  • Yearly growth rate (CAGR)

Risk-adjusted

  • Sharpe ratio
  • Sortino ratio like Sharpe, but only counts downside swings
  • Calmar ratio yearly return against the worst drawdown

Risk

  • Maximum drawdown in dollars and %
  • Ulcer index how deep and how long the losing stretches are
  • Time in the market

Trades

  • Number of trades
  • Win rate
  • Profit factor
  • Average win and average loss
  • Win/loss ratio
  • Average profit per trade
  • How long trades last

Breakdowns

  • What made the worst drawdown
  • Longs versus shorts
  • Which time of day the edge comes from

The verdict

Three honest answers. Decided by the software, explained by the AI.

"No honest improvement found" is a common result. That's the point.

Improved

A variation passed every check, including the sealed final test. On history. It is evidence, not a prediction.

No honest improvement found

Nothing tried beat your strategy honestly. Not the answer anyone hopes for, but the one you want before you size up. The report still shows where your strategy loses.

Can't tell yet

Not enough evidence either way. The report gives the exact reason and what would change it: more history, more trades, or a bigger research session.

Ready to see what your strategy is made of?

Who it's for

Built for retail traders, at every stage.

Whether you've traded the same setup for years or you're just getting serious, Kwaant gives you the kind of research quants rely on, in plain English.

You already trade a strategy

Find out if it really works, and whether it can honestly be made better.

You're testing a new idea

Bring a setup you've seen or sketched, and find out if it holds up before you risk real money.

You want to get into quant trading

Research like a quant without learning to code or earning a statistics degree.

You've been burned by a backtest

Get the luck checks your last backtest skipped, before you trust it again.

Any styleICT and smart moneyPrice actionSupply and demandIndicatorsBreakoutsOpening rangesSystematic

Prop firm mode

Trading a prop firm challenge? Test it under the firm's rules.

A strategy that makes money over the years can still fail an evaluation in its first bad week. The firm won't care that it would have recovered by Friday. Prop firm mode replays your strategy under the challenge's rules, and the AI looks for variations built to pass.

  • Profit targetReach it and you pass.
  • Maximum lossFixed or trailing your high. Touch it and you're out.
  • Daily loss limitHit it and trading stops for the day.
  • Minimum trading daysThe evaluation's clock counts too.

Illustrative replay, not real data or results. Passing on history doesn't mean you'll pass live.

What it is not

Clear lines, so you know exactly what you're getting.

  • Kwaant never places a trade.If you export to Pine Script, what happens next is your call.
  • No signals.It won't tell you what to buy today, or text you at 3am.
  • It doesn't invent strategies.It tests and improves the one you bring.
  • It promises no returns.Every result is on historical data and says so.

Early access

What is early access?

Your place in line for Kwaant's private beta, plus a few extras for getting here first. Being early has its perks.

Your place in line

When the private beta opens, we'll email you as soon as it's your turn.

A free strategy test, for a few

We'll pick a few applicants and test their strategy with Kwaant for free, before anyone else gets in.

Our Discord community

An optional invite to our free Discord, to talk with the team and other traders while we build.

An early-access price, for life

Join the beta when it launches and keep a special early-access price for as long as you stay. The details come before launch.

Takes about a minute. No payment needed to apply.
The beta itself will be paid. You'll see the price before you're asked to pay anything.

Questions

Good questions. Straight answers.

Still have a question?Ask us anything at hello@kwaant.com. No question is too basic, "what's a backtest?" included.

Will Kwaant make me profitable?

No software can promise that, and you should be wary of any that does. Kwaant tells you what your strategy's history really says, including when the answer is no.

Is it a trading bot?

No. Kwaant researches and tests; it never places a trade. If you want to automate a strategy, you export it as Pine Script and set that up yourself.

Do I need to code?

No. You describe your strategy in plain language, and Kwaant does the rest.

Which markets?

Futures, tested on 16+ years of historical CME Group data across a wide range of contracts, including FX futures. Options on futures are planned after launch.

Can I automate a strategy?

Yes, outside Kwaant. Export any tested strategy as Pine Script and automate it yourself if you choose. Kwaant itself never places a trade.

How much does it cost?

Pricing isn't public yet. You'll see it before you're asked to pay anything.

When can I start?

Apply for early access, and we'll email you when it's your turn.

Find out if your strategy really works.

Apply for early access. We'll email you when it's your turn.

Apply for early access

Takes about a minute. No payment needed to apply. What's included in early access?