The platform
One place to build it, test it,
and run it
Four surfaces, one engine underneath. What you approve in the builder is what the backtester runs and what goes live — there is no translation step in between, and no second set of rules for manual orders.
- 4
- surfaces over one engine
- One object
- what you approve is what backtests and goes live
- 1.8 ms
- median signal-to-order in internal testing
The surfaces
Build it, test it, run it — in one place.
Each surface does one job and hands off without reinterpreting anything. The strategy you approved is the object that trades.
Strategy Builder
Plain conditions you can audit line by line. Describe a rule in words, start from a template, or hand-edit each one — the same strategy underneath either way.
No black box
Backtester
Costs, slippage and financing are charged, not assumed. A result you cannot reproduce with the same inputs is a bug, and we treat it as one.
Costs charged
Risk Centre
Daily loss, position cap, drawdown and cluster ceilings live here. The engine will not start a route until they exist.
Limits first
Trading Terminal
Manual orders pass the same pre-trade checks as automated ones. If your rulebook requires a stop, the ticket will not submit without one.
Same rules
One object, four views
What you approve is what trades.
There is no export step and no second representation. The strategy you build is the object the backtester runs and the object that goes live.
Underneath all four
The risk layer sits in front of every one of them.
The four surfaces are different views of the same engine, and every order any of them produces passes the same pre-trade check. A manual ticket from the terminal is checked exactly as an automated route is.
That is why the limits are set once per account rather than once per surface — there is one book, and one set of ceilings over it.
- Limits per strategy, and a master guard per account
- Manual and automated orders share one exposure figure
- A breach is refused before the order leaves, on every surface
End to end
Idea to live route, without a translation step.
Each stage hands the same object to the next. Nothing is re-authored on the way.
Build it
Plain conditions you can audit line by line — described in words, from a template, or hand-edited. Contradictions surface while you write.
Test it honestly
The same rule set against history with spread, commission and slippage charged, and a count of the days your own caps would have intervened.
Set the limits
Per-strategy caps and the account ceiling above them. This is the screen before deployment, not a setting you find later.
Run it on demo, then live
Promote the strategy when it has behaved. The object moves across and its limits go with it.
Build one, test it, then watch it run.
Run the strategy, breach a limit on purpose, and watch what the engine does. That is the only demo worth having.