Pathnovo Logo

Extraction schemas

7 classes. 22 document types. 1,100+ typed fields.

TL;DR
Pathnovo extraction schemas are typed JSON shapes, one per document type. Every uploaded document is routed to one of 7 classes (P&IDs, isometrics, datasheets, HAZOP, mill certificates, and more) and returned in that class's schema. All 22 schemas embed the shared 28-field title_block header so cross-document joins work out of the box.