Use point-in-time data without silent restatements or look-ahead leaks. Keep the dataset version with each result so you can rerun and explain the work later.
Built for research that must be rerun, reviewed and explained.
Retained versions carry content hashes. A correction creates a new version rather than silently changing the prior one.
Every response carries row count, a completeness flag and a cursor. A partial page is labelled partial.
The as-of boundary is enforced before any row is returned. Unsupported requests are explained instead of silently falling through to present-day data.
Notebooks, APIs and agent tools read the same governed dataset, with its version and provenance attached to every response.
The surfaces that matter most for this job, in the order you would meet them.
Query versioned market data through an API.
Explore →TRADEPOLARIS MCPBring TradePolaris into your AI workspace.
Explore →DATAMarket data with dates and coverage shown.
Explore →AI RESEARCH ANALYSTResearch with the chart, query and sources attached.
Explore →Tell us what you run and we will show you the parts that matter, on one of your own books.