You describe the strategy. Jo builds and tests it.
Tell Jo the setup in plain English. The tearsheet here is a real backtest, not a mockup: one sentence in, 74 trades out, worst month included. Then the gauntlet:
Free: 300K AI tokens a week · 3 active projects · no card. First backtest tonight; the optimizer and paper desk are Pro.
"Go long the NIFTY 50 when the 20-hour SMA crosses above the 50. Flip short when it crosses back down."
Idea to live paper trade without leaving the project. No notebook, no spreadsheet, no copy-pasting between five tools.
Plain English. Entry, exit, universe. Jo asks when something is ambiguous.
Real, auditable code you own — open it or ignore it. Every version is saved.
Next-bar-open fills, transaction costs included, intrabar stops. The trade log exports to CSV.
Grid search on a train/test split, scored by lower confidence bound. Overfit settings get flagged; the ones that survive out-of-sample rank to the top.
Real market bars, simulated money. Positions and P&L tick live; the run keeps going unattended, paper only.
Most backtesting flatters the strategy. QuantJo is built to disappoint you on paper, before the market does it with your money.
A signal on bar N fills at bar N+1's open. Your backtest can't trade on a close it hasn't seen yet — the most common way a backtest flatters itself.
Platform exit rules (stop loss, take profit) execute intrabar against the high and low, on the touch. A gap through your stop fills at the open, not at your price. No end-of-bar mercy.
Parameter search trains on one window and is judged on the next, scored by lower confidence bound. Overfit ideas get flagged with the reason, and the ones that don't survive out-of-sample sink to the bottom.
Explore your data directly: rank the Nifty 50 by a signal, chart a pair spread, keep every analysis alongside the strategy.
Better to find that out from a backtest than from the market. Put a hunch through it, free →
Connect a provider one time. Jo checks what that provider offers (historical, live feed) and uses that data everywhere: backtests, research, paper trading.
| Provider | Markets | Bars | Auth |
|---|---|---|---|
| ZerodhaIndia | NSE & BSE equities, futures, options | Daily + intraday, 1m to 240m | OAuth |
| DatabentoGlobal | CME, NYSE, NASDAQ, crypto | Tick data, any interval reconstructed | API key |
| MassiveUS | US equities, ETFs & indices | Daily, hourly, intraday | API key |
Every trading tool oversells. Here is our own risk disclosure.
Built in the open, with real ship dates and status.
One strategy, every name in the bucket. Sweep the whole Nifty 50 in a single run. Pro.
Native 1-second bars on CME futures via Databento, historic and live. Pro.
NSE options data is already connected. Strategy support lands in the SDK next.
Screen on earnings and balance-sheet data alongside price. More exchanges beyond NSE, BSE, and US venues.
Start free. Pro is ₹3,250 a month at launch, half its list price.
Describe an idea, get the code, backtest it against real data.
The full loop: optimization, live paper trading, bucket sweeps.
Subscribe at launch and keep this price for as long as you stay subscribed.
What the product does and doesn't do, in plain terms.
Yes. You describe the strategy in plain English and Jo writes the code. You can read and edit the file if you want to; you never have to.
Yes. NSE and BSE data come through Zerodha, intraday and daily. US and global markets come through Databento and Massive.
The Free plan costs nothing and needs no card: 300K AI tokens a week, 3 active projects, backtesting and research included. Pro is ₹3,250 a month at launch, half the ₹6,500 list price, and subscribers who join at launch keep that rate for as long as they stay subscribed. The optimizer, paper trading, bucket runs, and 1-second data are Pro.
Signals generated on one bar fill at the next bar's open, so a backtest can never trade on a close it hasn't seen. Stops and targets execute intrabar against the high and low, and a gap through your stop fills at the open.
No. QuantJo paper trades: live market data, simulated money. It does not connect to a brokerage and it cannot place a real order.
It's either an edge or it isn't. Under a minute to a project, your first backtest tonight.
Free plan: 300K AI tokens a week · 3 active projects · no card.