Public docket sites list cases. A few will sell you the PDF. None of them have read it. The difference between an index and a reader is the whole product.
I
The harvest
Cases, dockets and documents are gathered continuously from the public record and kept permanently. Cheap structured data lands first and is searchable immediately; dockets, page images and readings fill in behind it, within rate budgets, forever.
II
The reading layer
Every document is read by AI in a carefully architected pipeline that extracts citable outcomes: findings of fact, custody awards, support figures, parenting schedules, fee shifting. Each extraction carries its quote, its page citation and a confidence figure, and can be verified against the page by a human.
III
Honest answers
Where the corpus is thin, results say so, tier by tier: docket, documents, readings, outcome. A question that reaches unswept territory never returns nothing — it queues the harvest and shows its progress, and one click moves your question to the front of the line.