A team of AI agents, each with one job
Each agent does one job, and a server referee with no AI in it checks their work: costs, grade caps, size and every price the verdict quotes.
Is there a trade right now?
The same zero-cost pre-check the engine runs before any paid analysis — BTC, ETH and SOL on scalp and intraday, live. Most of the time the honest answer is no; the skill is catching when it flips.
scanning…
Scalp · 5m
Intraday · 15m
Intraday · 30m
BTC
…
…
…
ETH
…
…
…
SOL
…
…
…
BTC/USDT · 15m
loading snapshot…
Entry, stop and target for this cell are computed in the app — the free chart shows the market and its reference levels only.
Deterministic pre-check, no AI call, no account needed. Not a signal.
AI agents read the chart, the derivatives flow and the news, test the key levels on the candles and hand you a graded plan.
Educational product. All trading here runs on a virtual balance: no real money, no exchange accounts, no orders on any exchange. Not investment advice.
Each agent does one job, and a server referee with no AI in it checks their work: costs, grade caps, size and every price the verdict quotes.
Entry trigger, stop, targets, net reward-to-risk after fees and slippage, a time horizon, and an A, B or C grade — or an honest skip.
Virtual money only: no real funds, no exchange, no orders anywhere. With auto mode on and the app open, a simulated position opens when a scenario triggers, sized by its grade.
Fresh headlines feed the analysis, spam filtered out. A CPI, FOMC or NFP release 8–24h out caps the grade at B; in the 4h before one, virtual trade size shrinks to the expected move.
In thin Asian hours, at the US open and at weekends, scalp and intraday stops get extra room, and a signal conflict in a thin session counts one notch softer.
A free pre-check reads market readiness with no AI call. Or queue a report: it runs on its own once conditions are clean, and credits are charged only when it runs.
The app's chart draws levels, figures and indicators on exchange candles, then lays your active scenarios and the liquidity map on top. Every layer is one click away.
Funding, open interest, taker CVD, spot vs perp: who is driving the move.
Pick one, or let the system suggest one from the market regime. Each has its own timeframes, stops and horizon.
In every live scenario card: your risk per trade in; size and funding cost out.
Brier score and a calibration curve: how well the predicted odds matched what really happened.
Every resolved scenario, losses included, checked bar by bar on exchange candles.
See every outcome →Free tier: 1 credit every month, no card required.
No trade, most of the time
Most analyses come back “no trade” — when no setup clears the cost-adjusted reward-to-risk bar, the engine declines instead of forcing one.
Net of real costs
A setup only becomes tradeable once its reward-to-risk clears a strict minimum — after realistic round-trip trading costs, not before.
Grades capped by context
The grade is capped by context, not taken on the model's word: if the main setup fights the higher-timeframe trend, the call tops out at C; a high-impact release 8–24h out or conflict across every scenario caps it at B. Until win rates are proven, size stays a small probe.
A full market report costs 1 credit and a follow-up answer 0.5 — on any model, Claude Opus 5.5 included. Start free — 1 credit every month.
10 credits
$15
one-time$1.50 per credit
25 credits
$35
one-time$1.40 per credit
60 credits
$79
one-time≈ $1.32 per credit
No. The app can auto-trade your scenarios only on a virtual balance: practice trades with no exchange connection, no real orders and no funds held. It is an educational tool; what you do on a real exchange is your own decision.
Credits pay for analysis: a full market report costs 1 credit and a follow-up question 0.5 — on any model, Claude Opus 5.5 included. The free tier includes 1 credit a month to try it — no card required.
It means the analysis ran and no setup cleared the bar. That happens more often than a graded call. A refusal is a result of the analysis, not a failed request, so it consumes the report's credits the same way a scenario does.
No — it's an educational and informational tool. Cryptocurrency markets are highly volatile; always do your own research.
It's logged the same as a win. Losses are part of the same public journal, not hidden or deleted.