Cross-exchange liquidation analytics, calibrated directional synthesis, and validation rigor that names its own confidence. Built for traders who want their tooling to tell them when it's wrong, not just when it's right.
Private beta · 5–10 invited users · provisional analytics with documented hit rates · no "validated edge" claims
Six capabilities, one operating principle: discipline machinery, not strategy delivery. We don't tell you what to trade. We give you the instrumentation to know when your strategy is working and when it's drifted.
Real-time event feed across Bybit, Binance, and OKX. Time-decayed cluster heatmap. Per-band leverage breakdown. Magnet detection that surfaces gravitational targets near current mark.
1-hour-horizon directional prediction with documented hit rate. Five layers (liquidity, HTF trend, news, funding, OI) blended into one signal. Auto-disables symbols where calibration falls below random + 10pp.
Walk the book to estimate fill price for any proposed trade size. Per-level visualization. Aggregate depth across exchanges. WSS-streamed order book intelligence on hot symbols.
Live trades dashboard with realized P&L (lifetime), unrealized P&L, funding accumulation, age-in-trade, distance-to-liq. Limit Trim 25/50/75% and Limit Add follow-on at top-of-book.
Structured journal with tags, ratings, mistakes. Stats grouped by setup, regime, time-of-day. Every prediction surface ships with its calibration metadata in the footer.
Funding rate divergence, OI velocity, and L/S extreme readings across the active universe. Highlights regimes where your strategy may face headwinds.
We're a tools company, not a signals service. The reason that matters:
You bring it. We don't see it, copy it, or aggregate it. Per-tenant data isolation by construction — your trade journal, your wallet lookups, your API keys live in a row scoped to you and only you.
We charge you a flat fee. We don't take a cut of your P&L. We don't sell your trade flow. If our tooling doesn't make you more disciplined, you cancel. That's the loop.
Our directional synthesis ships with a hit rate displayed in the footer. When a symbol's calibration falls below random + 10pp, we auto-disable it. We tell you when our tools are wrong before they cost you.
Most analytics platforms show you the prediction and hide the prediction's track record. We show both — and we audit ours every six hours.
Engine Read predictions are scored against actual price moves at the prediction horizon. The global hit rate must beat random + 10pp (3-class baseline 0.33). When it doesn't, the synthesis is flagged provisional in the dashboard footer.
Each symbol's hit rate is tracked separately. If it falls below 0.25 (well below 3-class random), that symbol is automatically muted from the directional read. The auto-disable circuit fires every 6 hours.
For every symbol, we compute "would removing this layer change accuracy?" — for liquidity, HTF trend, news, funding, OI. When a layer is hurting (not helping), its weight is reduced or zeroed. We document why in the changelog.
Hit rates are computed over the last 7 days of predictions, refreshed every 6 hours. New regimes are picked up within ~7 days; old patterns roll out of the window naturally. Calibration metadata appears in the dashboard footer in real-time.
Our calibration log is part of the product. Calibration metadata — hit rates, auto-disable status, model version, weight rebalance history — appears in the dashboard footer in real-time, on every prediction surface. We don't get to claim accuracy that we haven't earned.
The crypto-analytics market is full of dashboards. Most show you the data. None of the ones we surveyed publish their own calibration record.
| Hunter Killer | Coinglass | Hyblock | CryptoQuant | Velo / Laevitas | |
|---|---|---|---|---|---|
| Cross-exchange liquidation feed | ✓ | ✓ | ✓ | partial | ✓ |
| Calibrated directional predictions | ✓ | — | — | — | — |
| Per-layer ablation transparency | ✓ | — | — | — | — |
| Position management desk | ✓ | — | — | — | — |
| Trade journaling + calibration metadata | ✓ | — | — | — | — |
| Federated tenant isolation | ✓ | basic | basic | basic | ✓ |
| Pro tier price (typical) | $50/mo | ~$70/mo | ~$100/mo | ~$50–200/mo | institutional only |
| Institutional tier price (typical) | $200/mo | ~$300+/mo | ~$500+/mo | ~$700+/mo | $1k+/mo |
Pricing intervals are approximate ranges from public sources as of 2026-05; competitor pricing changes frequently. Our differentiation isn't price alone — it's the calibration record. We're the only platform in this list that publishes its own hit rate and auto-disables itself when it's wrong.
Three tiers. No hidden trial-to-paid surprises. Cancel any time, keep your data, export everything.
Beta-cohort pricing: Pro features are free during the private beta period for invited participants. We'll give 30 days' notice before the first charge. You can cancel before that without entering payment details.
We're inviting 5–10 traders to the first cohort. Calibration corpus, UX feedback, and product roadmap shape from your hands-on use.
Engineer note: replace [BETA_INTAKE_EMAIL] with the actual intake address when owner email sender is provisioned.