<- Support projects

Support Project · Python

DocumentIntelligenceLab

FastAPI document-intelligence proof for intake, classification, extraction, confidence review, human approval, batch processing, and export.

Repository readme / adapted

Document Intelligence Lab

SOURCEREADME.mdAdapted for portfolio readability; claims remain source-derived.

A Python/FastAPI proof for review-first document classification, candidate field extraction, confidence, provenance, batch processing, and evidence export.

**Status:** native-validated public release (`v1.0.0`) **Evidence:** 3 screenshot groups / 12 native browser captures

01

Core proof

  • validated document intake;
  • candidate invoice / purchase-order classification;
  • confidence and reason evidence;
  • field extraction with provenance;
  • Approved / PartiallyApproved / Rejected / NeedsReview states;
  • batch review routing;
  • evidence export;
  • deterministic tests and native HTTP smoke.
02

Boundary

All documents, OCR text, classifications, extracted fields, confidence values, and review decisions are synthetic. The app consumes supplied OCR text, does not claim production OCR accuracy, and never auto-approves extracted fields.