/
WanoTradingComing Soon

Algorithmic trading,
at machine speed

WanoTrading ingests MBO/L3 market data at 1M+ events/sec via Rithmic, stores everything in ClickHouse, and serves AI-driven signals with sub-100ms latency. From tick to decision — zero compromise.

1M+ Events/sec
ClickHouse Storage
<100ms Latency
AI Signals

Two apps. One mission. Zero latency.

WanoTrading splits ingestion and intelligence into two independent systems connected by ClickHouse. No coupling, no compromise.

High-Performance Ingestion

Go workers connect directly to Rithmic with one worker per symbol. Parse MBO/L3 data and write to ClickHouse at 200k–1M events/sec with sub-5ms tick-to-DB latency.

Time Series at Scale

ClickHouse columnar storage compresses billions of rows and returns queries in under 1 second. The single bridge between ingestion and intelligence — no other dependency.

AI Decision Engine

PyTorch models trained on local GPU — zero cloud cost. Feature engineering on volume imbalance, order flow toxicity, and iceberg detection. FastAPI serves signals in under 100ms.

Built for speed, not convenience

Every component chosen for raw performance. No GIL, no single thread, no cloud dependency.

MBO/L3 Ingestion

Full market-by-order data from Rithmic. Every bid, ask, and trade captured at wire speed with zero data loss.

ClickHouse Storage

Columnar time series optimized for billions of ticks. Sub-second queries on historical data with extreme compression.

Go Workers

One goroutine per symbol, no GIL, no shared connections. Designed for 200k–1M events/sec across 20+ futures.

Feature Engineering

Volume imbalance, order flow toxicity, iceberg ratio — all computed from raw L3 data for maximum signal extraction.

Real-Time Signals

AI signals overlaid on live charts. Control confidence thresholds per symbol, activate or deactivate models on the fly.

Local GPU Training

Train models on your own hardware. No cloud GPU costs, no data leaving your infrastructure. Full control over your edge.

Ops Dashboard

Monitor workers, events/sec, latency, errors, and disk usage. Start, stop, and restart ingestion per symbol from the UI.

From tick to signal

Five stages, two independent apps, one ClickHouse bridge.

Ingest

Rithmic feeds MBO/L3 data to Go workers — one per symbol, zero contention.

Store

ClickHouse persists every tick with columnar compression. Sub-5ms write latency.

Engineer

Python reads ClickHouse, computes features: volume imbalance, toxicity, iceberg ratio.

Train

PyTorch trains models on local GPU. Export to ONNX for production inference.

Signal

FastAPI serves predictions in <100ms. Trading UI overlays signals on live charts.

Stack & instruments

Purpose-built tech stack for futures trading. Each component chosen for raw performance.

GO
Go
RI
Rithmic
CL
ClickHouse
PY
Python
PY
PyTorch
ON
ONNX
FA
FastAPI
NE
Next.js
DO
Docker
ES
ES
NQ
NQ
CL
CL
GC
GC
YM
YM
RT
RTY
ZB
ZB

20+ futures instruments supported at launch

Built for serious traders

Whether you trade discretionary or systematic, WanoTrading gives you the infrastructure edge.

Discretionary Traders

Overlay AI signals on your existing charts. Set confidence thresholds per symbol, keep full control of execution.

Example use cases:
"Show me ES order flow imbalance"
"Alert when toxicity spikes on NQ"
"Backtest this signal on last 30 days"

Quant Engineers

Access raw MBO/L3 data in ClickHouse, build custom features, train models on local GPU, deploy via FastAPI.

Example use cases:
"Query 3 months of CL tick data"
"Train iceberg detection model on GC"
"Deploy new signal with <100ms SLA"

Trading Desks

Monitor multiple workers, track ingestion health, and get real-time PnL across all active futures positions.

Example use cases:
"Show all worker statuses"
"Events/sec across 20 futures"
"PnL breakdown by instrument"
Coming Soon

Get the infrastructure edge

WanoTrading is currently in development. Join the waitlist to get early access and trade futures with machine-grade infrastructure.

Back to Home