Loading data...
The full intelligence layer runs on Bitcoin: cascade probability, whale alpha, and on-chain conviction, all measured at the source. Raw derivatives data spans 24 exchanges. You see structure before price responds to it.
Direct WebSocket connections, no aggregator middleware. Every venue runs on its own dedicated pipeline.
Nexus Glass produces the data, never resells it. The same primary-source feed runs every endpoint at every tier.
Aggregator APIs lag 30–60 seconds and break on cascade days. You burn cycles cleaning data instead of training models.
Direct WebSocket from 24 exchanges. ML-ready feature store with 200+ REST endpoints, sub-50ms API P95. Backfilled history from 2024.
You read price action but the cascade already triggered. By the time the heatmap updates, the move is over.
Sub-second event delivery. Cascade warnings 30+ minutes ahead, whale entries 4–12 hours before retail sees the position.
You need vendor independence with audit-grade provenance. Most providers resell aggregated data they cannot reproduce.
Quantitative surfaces (Gatheral SVI, Feynman-Kac PDE) recalibrated every 5 minutes. Cross-layer derivative-onchain correlations. Direct engineering access.
Three signals pulled directly from the production API. Same endpoints, same auth, same payload your integration will consume.
No wrappers around third-party SDKs. No outsourced compute. Every engine in the platform runs on Nexus Glass infrastructure, end-to-end.
Nexus Glass computes liquidation cascade probability from orderbook structure and returns a risk score, a trigger zone, and a time window.
The pipeline runs end-to-end in under one second. No human interpretation. No dashboard refresh. Your models receive the signal, the direction, and the window.
Derivatives, on-chain, options, and whale data, collected from the source and served through a single integration.
Every metric in the platform is reachable via REST and streamable over WebSocket. One authentication, one key namespace, four official SDKs.
Edge-terminated in Frankfurt, Tokyo, and Virginia. Measured against real traffic, not lab tests.
Push delivery the moment events fire. P0 / P1 / P2 priority channels, each independently throttled.
Every endpoint returns the same shape every time. Versioned semver. No silent schema drift.
Python, TypeScript, Rust, Go, generated from the same OpenAPI document your models will consume.
# Liquidation heatmap · ThermalMap
GET /v1/heatmap?symbol=BTCUSDT
{
"symbol": "BTCUSDT",
"as_of": "2026-05-16T14:22:07Z",
"buckets": [
{
"price": 98500,
"side": "SHORT_LIQ",
"intensity_usd": 284210.55,
"cascade_risk_prob": 0.73,
"magnetism_index": 67,
"breakability_score": 84,
"absorption_state": "BUILDING"
}
]
}Twenty-five endpoints no other platform produces. Direct mathematical output from our compute layer, purpose-built for hedge funds, quant desks, and prop firms.
Gatheral SVI calibration with 5 free parameters, constrained by Fukasawa no-arbitrage bounds. Feynman-Kac PDE solver produces a 3D liquidation-probability surface across price, time, and leverage. Recalibrated every 300 seconds from the full Deribit options chain.
125 whales qualified by 6-month Sharpe-based alpha persistence. Each signal includes entry price, position size, and real-time P/L attribution. Backtested across 2,450+ historical signals.
10 derivative-onchain correlation metrics. Funding rate vs NUPL divergence detection. OI vs realized cap alignment scoring. SOPR vs liquidation-cluster density mapping.
Bidirectional liquidation-chain modeling with Hawkes-process excitation. Input a trigger price and receive cascade depth, cumulative volume, and terminal price estimate.
From an independent desk to an institutional quant team. Monthly or annual; upgrade at any time.
Nexus Glass produces the signals. Your models. Your edge.