Skip to main content

​
Historical Data

Status: planned public product, backed by live internal datasets. The historical layer should package the R2 archive into clean customer-facing resources rather than exposing raw internal storage paths directly.

Markets

Reference history, lifecycle metadata, and canonical identifiers.

Trades

The core event stream for execution and backtesting work.

Candles

A simpler time-series product for charting and lightweight analytics.

​
Product goal

Give customers a stable way to retrieve market reference history, trade history, candle history, and later full order-book history.

​
Underlying Dataset Families

The current lake and registry already support dataset families such as:
  • kalshi_trade
  • kalshi_candle
  • spot_trade
  • spot_candle
  • spot_quote
  • book_delta

​
Delivery modes

The eventual public product should support at least two modes:
  • Query API for filtered slices
  • Bulk Parquet export for larger historical pulls

​
First recommendation

Start the historical docs around data families and query patterns first. Lock public endpoint names after you decide whether this launches as REST-only, bulk-only, or both.