.csvTabular observations
Event windows, liquidity measures, and portfolio changes in a row-oriented structure.
Financial & AI research datasets
Financial data.
Built to be tested.
Market events, financial documents, and model evaluation corpora. Twelve focused collections with inspectable fields and explicit research uses.
Start with the schema. Build the right research request.
Schema preview. Field placeholders are shown, not dataset records.
12Purpose-built
research collections
research collections
03Open, familiar
file formats
file formats
03Research
categories
categories
Dataset directory
Find the data for your next test.
Inspect the specification, choose your collections, and put together a research request.
Files, not friction
Familiar formats.
Inspectable structure.
Explore proposed schemas before integrating a dataset into a notebook, a document pipeline, or an evaluation harness.
Fit your existing workflowPythonpandasDuckDBREvaluation harnesses
Before you begin
A few practical
questions.
What is included in a schema download?
The proposed field names, supported format options, source context, and intended research use. Schema files do not contain underlying dataset records.
How do I request several datasets?
Open any collection and add it to your research list. Complete the project brief below and download the combined request as a JSON file.
Do all collections share the same data source?
No. Source coverage is defined separately for each collection. Filings, credit events, document annotations, and AI training corpora have their own source requirements.
Is dataset delivery available instantly?
This catalog presents dataset concepts. Availability, date coverage, quality checks, source entitlements, and delivery terms must be confirmed for each research request.
Your research list
Start with a
clear question.
Select the collections you need. Add your intended use and download a brief to scope coverage, licensing, and delivery.