Build a strategy in minutes, or wire every leg by hand. Zetta tests it, rehearses it, then trades it while you sleep.
Most platforms pick a side. They are either a template gallery you cannot change, or a scripting environment you have to learn first. Zetta is one engine with two front doors, and you can walk between them.
New to this
Twelve named strategies ship ready to go. Choose one, set how far out of the money you want to be and when to get out, and the platform fills in lot sizes, costs and margin for you. Run it on paper for as long as you like before any money is involved. Nothing to install, no code to write.
Been doing this a while
Six legs, your own strikes, a different distance for every expiry, re-entry rules, per-deployment risk caps. Sweep thousands of parameter combinations over your own tick data and rank what comes back. Every default the template chose for you is a field you can override.
It is holding the plan at 2:40pm when the position is red and the exit is ten minutes away. An algorithm does not widen a stop because it feels unlucky, or skip an entry because yesterday hurt, or stay up watching a screen. It does the thing you decided when you were calm, every time, including overnight.
Entry, exit, stop, target, size, re-entry. Written down as rules, tested against two years of real candles, not remembered as intentions.
The same engine and the same risk limits as live trading. The only thing that changes is the adapter underneath. Run it for a week or a quarter.
It places the orders, watches the stops and squares off on schedule. You get a tradebook in the morning instead of a screen to sit in front of.
A short strangle and a buyback build-up have nothing in common except the account they settle into. Each module runs its own engine and its own execution rules. Every position lands in the same book.
Short and long straddles and strangles, iron condors and butterflies, all four vertical spreads, naked calls and puts, plus a leg editor for anything else. Backtested candle by candle against your own historical data, not a vendor's summary.
Every NSE corporate action, ranked by what the trade is actually worth: dividend yield on the market price, days left to take it, and a score you can open up and argue with.
Index and stock futures on the same spine: the same backtest, the same paper mode, the same broker connections and the same position guards. Nothing about running a futures book is a separate code path.
“Final Dividend 425% @ Rs. 8.50 per share.”
The 425% is a percentage of face value, a number with no bearing on any trade you could place. The share costs ₹2,215. The actual yield is 0.38%.
What the incumbents show
| Stock | LTP | Dividend |
|---|---|---|
| APL Apollo | 2,215.00 | 8.50 |
| KDDL | 3,991.90 | 8.00 |
| Polyplex | 1,180.00 | 1.00 |
| Skipper | 604.40 | 0.10 |
Four rows of a few hundred. You divide, by hand, to find the two that matter.
What Zetta shows
| Stock | Yield | To ex | Score |
|---|---|---|---|
| APL Apollo | 0.38% | 0d | 4 |
| KDDL | 0.20% | 0d | 2 |
| Polyplex | 0.08% | 2d | 1 |
| Skipper | 0.02% | 0d | 1 |
All four are untradeable, and the table says so in one glance. On a real day's feed, 5 of 201 upcoming actions cleared the bar.
Timing is weighted alongside size, deliberately. A 6% yield two days from the ex-date is mostly spent; a 4% yield fifteen days out is the trade. Rank on size alone and the table confidently recommends positions you can no longer establish.
Two years of NSE corporate actions, August 2024 to August 2026, across a universe screened to ₹500 cr market cap and ₹5 cr average daily turnover.
| Rule | Window | n | Win | Avg |
|---|---|---|---|---|
| Buyback build-uptender-offer grind into the record date | ex-20 to ex-3 | 39 | 77% | +8.1% |
| High-yield dividendyield of 5% or better at intimation | intimation to ex-1 | 11 | 91% | +12.4% |
| Split ridehalf size; the edge is real but thin | ex-15 to ex-1 | 67 | 69% | +4.3% |
| Rights issuesnegative in every strategy and every phase | not traded | - | - | avoid |
These are backtested results. Two of the strongest rules rest on small samples, n=39 and n=11, which is exactly why they appear with their sample sizes rather than as headline percentages.
The study spans one broad market regime, covering both a correction and a rally. Rights issues were negative in every strategy and every phase, so the platform refuses to execute them however well they would score.
Handing execution to an algorithm only works if you know exactly what it is not allowed to do. These constraints are enforced in code, with tests that fail if anyone removes them.
Your broker account holds positions this platform did not put there: stock bought manually, options taken on a phone app. Every exit is derived from a position the platform itself recorded. Broker position lists are read, never used as a list of things to close.
If you square something off yourself, the platform's record is stale. Sending a buy to close a short that is already gone does not fail harmlessly. It opens a new position. So it checks, and refuses.
A paper deployment is barred from a real broker connection at creation and again on every tick, because tomorrow's bad database row is a likelier enemy than today's mistake.
Not a stored constant. The platform asks your broker what the basket actually costs, hedge benefits included, and re-checks at the moment of entry against the strikes it is about to trade.
A kill switch that squares off every position the platform opened, across options, futures and event deployments, paper and live, and touches nothing else in your account.
Every score traces back to inspectable inputs, and says in words why it scored that way. A ranking you cannot interrogate is a ranking you should not trade.
No. The twelve named structures are forms, not scripts: pick one, set the distance from the money and the exit rules, and deploy. The leg editor adds control when you want it, and it is still a form. There is nothing to install and no language to learn.
No. You trade through your own broker account: Zerodha, Dhan, Groww or AngelOne. Credentials are encrypted at rest and masked on every read. The platform places orders; it never custodies funds.
Neither. It gives you no advice and makes no recommendations about your finances. It is a tool for building, testing and running your own strategies, and every number it shows traces back to inputs you can inspect.
The strategies run on the server, not in your browser, so closing the tab changes nothing. Orders are placed through your broker's API and live in your broker account, where they stay valid whether or not the platform is reachable.
Corporate actions and filings come from NSE directly, including the PDFs attached to board-meeting outcomes. Historical option data is your own: the platform reads local files and never uploads them. Live quotes come through your connected broker.
Backtesting, the strategy builder, the event screener and paper trading are free to use. Brokerage and exchange charges are your broker's and are modelled in every backtest, so the numbers you see already account for them.
Start with a backtest, move it to paper, connect a broker when the numbers have earned it. The engine never changes underneath you.