One platform for the entire life of a systematic strategy.
Built for private investment offices and hedge funds: engineered for speed, scale and AI-native research — identical from the first backtest to the live book.
The platform we wanted did not exist
One system that carries a trading idea from raw market data through research, validation and optimization into live execution. Fast enough to iterate hundreds of experiments a day, scalable enough to watch entire exchanges tick by tick.
One engine and one set of strategy definitions: what is researched is exactly what trades. The platform is asset-class agnostic, and has been built and validated on exchange-traded futures and multi-leg spreads.
to replay a full trading day — whole book, fills, margin and accounting
of market history on tap, down to quote level
of instruments followed in real time, quote by quote
deterministic: repeated runs are byte-identical, by test
From raw data to the live book
- 01
Ingest
Exchange data at quote level, broker feeds, fundamentals and news.
- 02
Normalize
Every provider mapped to one identity per instrument.
- 03
Backtest
Years of history in seconds, with realistic fills and costs.
- 04
Validate
Statistical guardrails against overfitting, enforced by the tooling.
- 05
Optimize
Systematic parameter search, every trial archived.
- 06
Paper → Live
Scheduled trading day, crash recovery and daily monitoring.
AI-native by construction
Every part of the platform is controllable through plain text: strategies, data pipelines, risk policy and schedules are human-readable configuration, and every result is a plain, diffable file.
An AI agent can drive the whole lifecycle — import data, write a strategy, run backtest batteries, read the statistical verdict, optimize, promote — under human review. Deterministic runs make agent experiments trustworthy; research cycles that conventionally take weeks compress into days.
What the platform covers
All the data in one place
Bars, quotes and trades alongside fundamentals, positioning, weather and news — one consistent reference model.
Execution you can trust
Fill models calibrated against measured live slippage, quote sanity guards, multi-leg spreads handled natively.
Real margin & accounting
Exchange-grade portfolio margin in research and before every live order; per-strategy P&L that reconciles exactly.
Independent strategy sleeves
Dozens of strategies share the same markets, net internally and keep fully separate risk attribution.
Research discipline, built in
Every experiment is archived and judged with multiple-testing corrections. Negative results are kept, too.
Hands-off operations
The trading day runs itself on a schedule, with monitoring, alerting, a full web dashboard and an API.
For quants, traders and risk managers who want to challenge it
We built gtrade for our own desk. We are glad to walk investment offices and funds through the platform, under NDA.
gtrade is proprietary technology. Nothing on this page is an offer, a solicitation or investment advice of any kind, and no performance claims are made. Figures describe the software in its current working setup.