Signal engine
Models watch momentum, funding rates, order-book depth and realised volatility across major venues, scoring opportunities continuously rather than on a fixed schedule.
Our technology
The trading stack behind every daily credit
Discretionary trading is inconsistent by nature. Our stack replaces hesitation, revenge trades and missed entries with rules that execute identically at 3am and 3pm — and a ledger that lets you audit the result.
Risk controls run first
Position sizing, exposure caps and automatic drawdown limits are evaluated ahead of every entry. Capital preservation gates each trade before a profit target is ever considered.
Low-latency execution
Orders are routed and split across venues to reduce slippage, with fills reconciled continuously against the strategy book so the position on paper matches the position held.
Transparent reporting
Every credit to your balance is a dated, double-entry ledger row you can open, trace back to its plan and export from your dashboard. Nothing is summarised away.
The pipeline
What happens between a signal and your balance
Four stages, each with its own failure mode. Naming them is more useful than a vague claim about artificial intelligence, so here they are in order.
-
01
Observe
Price, depth, funding and volatility are sampled continuously across venues. Stale or contradictory data is discarded rather than traded on — a bad feed is the fastest way to lose money systematically.
-
02
Score
Candidate positions are ranked against the current regime. A signal that is strong in a trending market is not automatically acted on in a choppy one.
-
03
Size and gate
Exposure caps, correlation limits and drawdown rules are applied before any order exists. A position that would breach a limit is reduced or skipped, never waived through.
-
04
Execute and reconcile
Orders are split to reduce market impact, then fills are reconciled against the strategy book so the position on paper always matches the position held.
What automation does not fix
The limits, stated plainly
Systematic execution removes a category of human error. It does not remove market risk, and any platform telling you otherwise is selling something. These are the constraints that remain no matter how good the models are.
Read the full risk guide-
Correlation spikes in a crisis
Assets that normally diversify each other tend to move together precisely when it matters. Position limits reduce the damage; they do not eliminate it.
-
Liquidity disappears first
The spread on a thin market widens exactly when you most want out. Execution logic accounts for this, but a forced exit in a dislocated market still costs more than a planned one.
-
Past relationships can stop holding
Every model is fitted to history. A regime that has not occurred before is, by definition, one no backtest covered.
-
Venue and counterparty risk is real
Capital held at an exchange depends on that exchange. Spreading exposure helps; it does not make the risk zero.
Reporting
Every credit is a row you can open
Balances are not a single number that moves. Each change is a dated, double-entry ledger row written inside a database transaction with the balance locked, so two simultaneous operations cannot produce a figure that fails to reconcile. You can open any entry, see which plan produced it and export the history.
How we protect accountsIllustrative ledger
Example| Date | Entry | Amount |
|---|---|---|
| Day 7 | Daily earning · Growth | +3.58 |
| Day 6 | Daily earning · Growth | +3.57 |
| Day 5 | Daily earning · Growth | +3.57 |
| Day 4 | Daily earning · Growth | +3.57 |
Note day 7: a cent higher than the others. That is the rounding correction that makes the week total exactly 25.00 rather than 24.99.
See the ledger from the inside
Create an account to explore the dashboard, then fund a plan when you are ready.