Jove

The README opens with

Jove turns scanned documents into structured data and a knowledge graph. Give it scanned pages (PDFs or images) and a data schema (a JSON Schema). It tags every piece of text with the schema field that text gives a value for, and it emits a JSONL graph for downstream systems, with exports to Obsidian, FalkorDB and TypeDB. Each tag carries a hash that points back to the exact page and bytes it came from.

It is a general tool: the schema is yours, and nothing in the pipeline is tied to one kind of document or one jurisdiction. The repo ships example schemas for legal intake, court filings (with e-signing and production-stamp data), flyers and posted notices, and document structure.