For Quants & PMs
Backtest infrastructure that doesn't lie.
Real prices, real splits, real dividends, point-in-time. We started with the dataset we wished we had as analysts — survivorship-free, fully versioned, accessible from Python and TypeScript clients. Composing factor models, running seasonality studies, and rolling correlation matrices used to mean a weekend in a notebook. Now it's a button.
All use casesBacktests / month
Up to 100
API calls / day
25K
Bulk export
Parquet
Composing factor models, running seasonality studies, and rolling correlation matrices used to mean a weekend in a notebook. Now it's a button.
Day in the life
What you actually do
Five steps from raw data to a decision.
- Step 01
Define a signal
~5m
Value, momentum, quality, size — or compose your own from primitives.
- Step 02
Run a backtest
~5m
1-line expression, 30 years of history. Quintile spread, IS decay, regime breakdown.
- Step 03
Push to production
~10m
REST, streaming, Parquet — same data the backtest ran on.
- Step 04
Cross-check with news
~5m
Filter transcripts by signal trigger. See whether narrative supports the backtest.
What you unlock
The features this audience uses most.
- 01Point-in-time price + fundamentals with full audit trail
- 02Survivorship-free universe — delisted tickers stay in history
- 03Pre-built factor models + composable alpha sandbox
- 04Python + TypeScript clients with type-safety end-to-end
Start free, scale when it matters.
Same data, three audiences, free forever for historical.