Bulk and SQL access

Query the whole corpus as a database: columnar files in an object store, readable in place over HTTPS.

INSTALL httpfs; LOAD httpfs;
SELECT citation, section_title, word_count
FROM read_parquet('https://data.lexbharat.example/v2026.09.1/us_ak_constitutions.parquet')
LIMIT 5;

Parquet files#

The corpus is split into one file per jurisdiction and kind of law, for example us_tx_statutes.parquet.

One combined archive#

A single large archive contains all files for those who prefer to download everything.

A manifest with checksums#

An index.json lists every file with a SHA-256 so you can verify downloads.

Dated snapshots#

Files are published under a dated version folder so results are reproducible.